diff --git a/README.md b/README.md
index f8ee28c..40911f3 100644
--- a/README.md
+++ b/README.md
@@ -10,45 +10,72 @@ The environment is described in [this paper](https://www.researchgate.net/public
- `/examples` contains prototype code for the interaction of RL algorithms with an emulator building model from BOPTEST.
- `/testing` contains code for unit testing of this software.
-## Quick-Start
+## Quick-Start (using BOPTEST-Service)
+BOPTEST-Service allows to directly access BOPTEST test cases in the cloud, without the need to run it locally. Interacting with BOPTEST-Service requires less configuration effort but is considerably slower because of the communication overhead between the agent and the test case running in the cloud. Use this approach when you want to quickly check out the functionality of this repository.
+
1) Create a conda environment from the `environment.yml` file provided (instructions [here](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html#creating-an-environment-from-an-environment-yml-file)).
-2) Run a BOPTEST case with the building emulator model to be controlled (instructions [here](https://github.com/ibpsa/project1-boptest/blob/master/README.md)).
-3) Develop and test your own RL algorithms. See example below using the [Bestest hydronic case with a heat-pump](https://github.com/ibpsa/project1-boptest/tree/master/testcases/bestest_hydronic_heat_pump) and the [A2C algorithm](https://stable-baselines.readthedocs.io/en/master/modules/a2c.html) from Stable-Baselines:
+2) Check out the `boptest-gym-service` branch and run the example below that uses the [Bestest hydronic case with a heat-pump](https://github.com/ibpsa/project1-boptest/tree/master/testcases/bestest_hydronic_heat_pump) and the [DQN algorithm](https://stable-baselines3.readthedocs.io/en/master/modules/dqn.html) from Stable-Baselines:
```python
-from boptestGymEnv import BoptestGymEnv, NormalizedActionWrapper, NormalizedObservationWrapper
-from stable_baselines3 import A2C
-from examples.test_and_plot import test_agent
-
-# BOPTEST case address
-url = 'http://127.0.0.1:5000'
-
-# Instantite environment
-env = BoptestGymEnv(url = url,
- actions = ['oveHeaPumY_u'],
- observations = {'reaTZon_y':(280.,310.)},
- random_start_time = True,
- max_episode_length = 24*3600,
- warmup_period = 24*3600,
- step_period = 900)
-
-# Add wrappers to normalize state and action spaces (Optional)
+from boptestGymEnv import BoptestGymEnv, NormalizedObservationWrapper, DiscretizedActionWrapper
+from stable_baselines3 import DQN
+
+# url for the BOPTEST service.
+url = 'https://api.boptest.net'
+
+# Decide the state-action space of your test case
+env = BoptestGymEnv(
+ url = url,
+ testcase = 'bestest_hydronic_heat_pump',
+ actions = ['oveHeaPumY_u'],
+ observations = {'time':(0,604800),
+ 'reaTZon_y':(280.,310.),
+ 'TDryBul':(265,303),
+ 'HDirNor':(0,862),
+ 'InternalGainsRad[1]':(0,219),
+ 'PriceElectricPowerHighlyDynamic':(-0.4,0.4),
+ 'LowerSetp[1]':(280.,310.),
+ 'UpperSetp[1]':(280.,310.)},
+ predictive_period = 24*3600,
+ regressive_period = 6*3600,
+ random_start_time = True,
+ max_episode_length = 24*3600,
+ warmup_period = 24*3600,
+ step_period = 3600)
+
+# Normalize observations and discretize action space
env = NormalizedObservationWrapper(env)
-env = NormalizedActionWrapper(env)
+env = DiscretizedActionWrapper(env,n_bins_act=10)
+
+# Instantiate an RL agent
+model = DQN('MlpPolicy', env, verbose=1, gamma=0.99,
+ learning_rate=5e-4, batch_size=24,
+ buffer_size=365*24, learning_starts=24, train_freq=1)
-# Instantiate and train an RL algorithm
-model = A2C('MlpPolicy', env)
-model.learn(total_timesteps=int(1e5))
+# Main training loop
+model.learn(total_timesteps=10)
-# Test trained agent
-observations, actions, rewards, kpis = test_agent(env, model,
- start_time=0,
- episode_length=14*24*3600,
- warmup_period=24*3600,
- plot=True)
+# Loop for one episode of experience (one day)
+done = False
+obs, _ = env.reset()
+while not done:
+ action, _ = model.predict(obs, deterministic=True)
+ obs,reward,terminated,truncated,info = env.step(action)
+ done = (terminated or truncated)
+
+# Obtain KPIs for evaluation
+env.get_kpis()
```
+## Quick-Start (running BOPTEST locally)
+Running BOPTEST locally is substantially faster
+
+1) Create a conda environment from the `environment.yml` file provided (instructions [here](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html#creating-an-environment-from-an-environment-yml-file)).
+2) Run a BOPTEST case with the building emulator model to be controlled (instructions [here](https://github.com/ibpsa/project1-boptest/blob/master/README.md)).
+3) Check out the `master` branch of this repository and run the example above replacing the url to be `url = 'http://127.0.0.1:5000'` and avoiding the `testcase` argument to the `BoptestGymEnv` class.
+
+
## Citing the project
Please use the following reference if you used this repository for your research.
diff --git a/__init__.py b/__init__.py
new file mode 100644
index 0000000..d312e28
--- /dev/null
+++ b/__init__.py
@@ -0,0 +1,5 @@
+'''
+Created on Dec 20, 2020
+
+@author: Javier Arroyo
+'''
diff --git a/docs/tutorials/CCAI Summer School 2022/Tutorial_2_Building_Control_with_RL_using_BOPTEST.ipynb b/docs/tutorials/CCAI Summer School 2022/Tutorial_2_Building_Control_with_RL_using_BOPTEST.ipynb
deleted file mode 100644
index 48b60d4..0000000
--- a/docs/tutorials/CCAI Summer School 2022/Tutorial_2_Building_Control_with_RL_using_BOPTEST.ipynb
+++ /dev/null
@@ -1,3090 +0,0 @@
-{
- "cells": [
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "view-in-github",
- "colab_type": "text"
- },
- "source": [
- ""
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "V1CcDG8FanTw"
- },
- "source": [
- "#**Key Learning Objectives** 🎯"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "oT2QjTu24zwV"
- },
- "source": [
- "\n",
- "This is an introductory, hands-on tutorial to guide you through the main concepts of Reinforcement Learning (RL) for controlling Heating, Ventilation and Air Conditioning (HVAC) systems for buildings. \n",
- "We are going to apply RL to a building emulator from the Building Optimization Testing (BOPTEST) framework **[1]** using the BOPTEST-Gym interface **[2]**. \n",
- "BOPTEST is a framework for performance benchmarking of control algorithms.\n",
- "Further information and documentation can be found here: \n",
- "\n",
- "[https://ibpsa.github.io/project1-boptest/](https://ibpsa.github.io/project1-boptest/)\n",
- "\n",
- "You will learn:\n",
- "\n",
- "- What RL is, how it works and how it can be used in the application of building energy management. \n",
- "- The most popular standard for representing general RL problems: OpenAI-Gym.\n",
- "- The BOPTEST API and its Gym interface. \n",
- "\n",
- "📌 **Note**: This tutorial was prepared for use with BOPTEST v0.3.0. \n",
- "and will make usage of a web-based version of BOPTEST (called \"BOPTEST-Service\") so as not to require installation of any BOPTEST software on a user's own device. It is also possible to use BOPTEST on a user's own (local) device. Both the web-based and local versions have the same functionality, and will produce the same results, with only small changes in the API (changing the BOPTEST-service url to your localhost url, that is, to: `http://127.0.0.1:5000/`). \n",
- "\n",
- "**EDIT**: This tutorial was originally developed with BOPTEST v0.2.0. and has been updated to version 0.3.0. on *Oct 25, 2022*. There are just small changes required for this update, basically retrieving the `'payload'` after each request. That is the origin of the differences between the notebook explained in the recording and this updated notebook. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "TSTpxm2GrjhR"
- },
- "source": [
- "# **Outline** ⏰\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "VUbaQ5GqrvIl"
- },
- "source": [
- "[Part 1: Background](#background)\n",
- " 1. [Introduction to Reinforcement Learning](#introRL)\n",
- " 1. [Application of Reinforcement Learning in buildings](#applicationRL)\n",
- "\n",
- "[Part 2: The OpenAI-Gym standard](#openAIGym)\n",
- " 1. [What is OpenAI-Gym?](#whatIsOpenAIGym)\n",
- " 1. [Example using an OpenAI-Gym environment](#exampleOpenAIGym)\n",
- "\n",
- "[Part 3: The Building Optimization Testing (BOPTEST) Framework](#boptest)\n",
- " 1. [What is BOPTEST?](#whatIsBoptest)\n",
- " 1. [Selecting a building test case](#selectBuilding)\n",
- " 1. [Obtaining general information about the building](#obtainInfo)\n",
- " 1. [Getting control input and measurement points](#gettingIOs)\n",
- "\n",
- "[Part 4: Implementing RL for a building with BOPTEST-Gym](#implementingRL)\n",
- " 1. [What is BOPTEST-Gym?](#whatIsBoptestGym)\n",
- " 1. [Starting up a BOPTEST-Gym environment](#startingUpBoptestGym)\n",
- " 1. [Interacting with a BOPTEST-Gym environment](#interactingWithBoptestGym)\n",
- " 1. [Developing a basic RL algorithm](#developingRlAlgo)\n",
- " 1. [Testing our RL algorithm in BOPTEST-Gym](#testingRlAlgo)\n",
- "\n",
- "[Gearing up](#gearingUp)\n",
- "\n",
- "[Further resources](#furtherResources)\n",
- "\n",
- "[Feedback](#feedbackForm)\n",
- "\n",
- "[Annex I: Formal RL theory](#theoryRlFormal)\n",
- "\n",
- "[References](#references)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "oEzP9ZW4MXPv"
- },
- "source": [
- "#**Part 1: Background** 📖 "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "Fas232CyMX6_"
- },
- "source": [
- "##**Introduction to Reinforcement Learning** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "TAy9fRjUTSdb"
- },
- "source": [
- "Could you imagine a magic oracle able to decide on the best actions to optimize any process? Could you imagine this oracle not needing any prior information of the process but just learning from interacting with it? Powerful, right? Well, that is exactly what RL is meant for. \n",
- "\n",
- "Reinforcement Learning (RL) is one of the categories of machine learning, along with unsupervised learning and supervised learning. The main difference from the other categories is that RL learns from dynamic data, that is, data that are obtained while learning. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "e853vYumSx08"
- },
- "source": [
- "\n",
- "\n",
- "*Figure: The categories of machine learning. Source: [Mathworks](https://www.mathworks.com/discovery/reinforcement-learning.html)*\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "2_qdE6Ab4aE9"
- },
- "source": [
- "In RL the goal is to learn the actions to be taken to achieve a predefined objective. RL relies on the principle of *repetitive experimentation*, that is, an approach where we roll out several **episodes of experience** where an agent 🤖 (the RL algorithm) interacts with its environment 🌎 (the process to be optimized) to learn based on a **reward** signal that is returned for every **action** taken from a specific **state** of the environment. \n",
- "\n",
- "Let's take the example of teaching a dog to grab a stick. In this case, the dog is the agent and all its surroundings conform the environment. Whenever the dog observes that there is a person throwing a stick it will perform an action. In case it grabs the stick and brings it back, the person will provide a cookie as a reward to encourage that behavior. In case the dog does not go for the stick but just runs around or goes chasing other dogs, the person will not provide the cookie. Eventually, the dog will associate the actions that bring the most rewards to specific observations and will be taking those actions accordingly. \n",
- "\n",
- "A more formal introduction to RL and its associated terminology can be found at the end of this tutorial, in [Annex I: Formal RL theory](#theoryRlFormal). "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "mMppgppKX4Fy"
- },
- "source": [
- "\n",
- "\n",
- "*Figure: RL notation when teaching a dog. Source: [Mathworks](https://www.mathworks.com/discovery/reinforcement-learning.html)*\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "pNy7uRzo8foI"
- },
- "source": [
- "⚠️ **Important note:** ⚠️ It is common to find in the RL literature that the same term indistinctly designates the\n",
- "state and the observation. This is not strictly correct for partially observable environments (most of the cases) where the observation only conveys part of the information that defines the state. For example, the state of the Tic-Tac-Toe game can be fully observed because there is a finite number of possibilities that define the state of a game. On the contrary, the thermal state of a building is only partially oservable. We can observe the indoor air temperature, but we cannot measure all temperatures from walls, ground, furniture... which also influence the building's thermal state. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "pNC0UnC2WYyE"
- },
- "source": [
- "\n",
- "\n",
- "What is particularly extraordinary of RL algorithms is that the same algorithm can be successfully used for a variety of tasks, from [robotic motion control](https://www.technologyreview.com/2021/04/08/1022176/boston-dynamics-cassie-robot-walk-reinforcement-learning-ai/) to [defeat the human world champion in the game of Go](\n",
- "https://www.youtube.com/watch?v=WXuK6gekU1Y&ab_channel=DeepMind).\n",
- "The latter is an astonishing achievement. It is true that the IBM supercomputer Deep Blue could previously [defeat Garry Casparov in chess](https://en.wikipedia.org/wiki/Deep_Blue_versus_Garry_Kasparov), but Go is to chess what chess is to Tic-Tac-Toe ([*Chris Wiltz*](https://www.designnews.com/design-hardware-software/googles-ai-beat-go-champion-mimicking-human-intuition)). And what is more important, professionals of Go state that this game has so many possible combinations that mastering it requires certain intuition and creativity, qualities that have only been attributed to humans so far... if AlphaGo defeated the best human player of Go, could machines resemble these qualities? Well, that is more a philosophical question. This tutorial is limited to investigate whether machines can efficiently control buildings, which you will see is already an enormous challenge!\n",
- "\n",
- "\n",
- "\n",
- " "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "XOcvaJpkNbho"
- },
- "source": [
- "\n",
- "\n",
- "*Figure: Netflix documentary that explains how AlphaGo, a RL algorithm developed by [DeepMind](https://www.deepmind.com/), could defeat Lee Sedol (4-1) and Fan Hui (5-0), the human world champions in the game of Go.*\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "o1jhKBOeOciu"
- },
- "source": [
- "📌 **QUICK FACTS:**\n",
- "- RL is a **category of machine learning** algorithms, together with supervised and unsupervised learning. \n",
- "- Contrarily to other machine learning techniques, RL learns from **dynamic data**, that is, data that are obtained from interactions with the environment. \n",
- "- Particularly, it learns from **state-action-reward** samples, so there is no need of domain knowledge to model the environment.\n",
- "\n",
- "\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "6G1nWECgbmuW"
- },
- "source": [
- "##**Application of Reinforcement Learning in buildings** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "xoTh8XvAM_OR"
- },
- "source": [
- "During the last decade, there has been a clear interest growth in using optimal control for HVAC systems **[3]**. The figure below underlines this increased interest by showing the number of yearly peer-reviewed scientific publications related to optimal control in buildings. \n",
- "RL algorithms have\n",
- "gained particular popularity for their application in a **demand response** setting.\n",
- "An extensive review for this application was written by Vázquez-Canteli et al. **[4]** This review is\n",
- "not limited to HVAC systems but also demand response for charging electric\n",
- "vehicles or thermal energy storage."
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "7NywyXo6hD5n"
- },
- "source": [
- "\n",
- "\n",
- "*Figure: Evolution of the number of scientific publications about optimal control in buildings during the last decades. Data obtained from the Clarivate Web of Science.*"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "8KR-sTeJiJG2"
- },
- "source": [
- "RL has already attracted the attention of the building control community for\n",
- "many years. The figure below is obtained from the popular paper of Chen et al. **[5]** who graphically summarized the application of RL in buildings indicating the amount of data required by each research work to train the implemented RL algorithm. \n",
- "\n",
- "\n",
- "\n",
- "\n",
- "\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "nCH0DBBbPEC7"
- },
- "source": [
- "\n",
- "\n",
- "\n",
- "*Figure: Summary of the data required in the history of RL applications to buildings. Chen et al.* **[5]** ."
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "Ae18iXNKWV5I"
- },
- "source": [
- "You can see from the figure that the feasibility and potential of applying RL for HVAC control\n",
- "was first investigated by Liu and Henze back in 2006. Then, the interest was lost for a period, probably because Model Predictive Control (MPC) has been typically preferred for optimal control in buildings because it is much more data-efficient (it does not need as much data to be implemented). A comprehensive and complete review on the application of MPC for building energy management is provided by Drgona et al. **[6]**. \n",
- "The reasons why RL is gaining momentum again are clear: \n",
- "\n",
- "- Evolution in deep learning\n",
- "- We have much more data than before\n",
- "- We have much more computational power than before\n",
- "\n",
- "In fact, there exist very recent developments for the application of RL in buildings, most of them using the OpenAI-Gym standard that is introduced in the next section. It is worth mentioning:\n",
- "\n",
- "- [CityLearn](https://github.com/intelligent-environments-lab/CityLearn) ➡️ Gym environment for providing demand response scenarios at an urban scale. That is, the goal of the RL agent is to flatten the energy demand of a district. It considers static\n",
- "building heating and cooling load data and simplified models for energy storage. \n",
- "- [Gym-Eplus](https://github.com/zhangzhizza/Gym-Eplus) ➡️ Gym environment wrapper around EnergyPlus simulation models. \n",
- "- [Sinergym](https://github.com/ugr-sail/sinergym) ➡️ Extension of Gym-Eplus. \n",
- "- [Energym](https://github.com/bsl546/energym) ➡️ Gym wrapper around building simulation models to assess controller performance. \n",
- "- [Beobench](https://github.com/rdnfn/beobench) ➡️ A Toolkit for Unified Access to BuildingSimulations for Reinforcement Learning.\n",
- "- 👉🏻[BOPTEST-Gym](https://github.com/ibpsa/project1-boptest-gym) ➡️ Gym environment for the BOPTEST Framework. The goal of the RL agent in this environment is to efficienty control an individual building. It allows testing against high-fidelity building models. \n",
- "\n",
- "The last of which is the focus of this tutorial. \n",
- "\n",
- "These RL frameworks for HVAC control bring hope\n",
- "to the adoption of this technology in buildings. However, there is still a clear\n",
- "need to different techniques and understand the best practices of RL \n",
- "for this particular application. Let's investigate how!"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "7YnuNAQdM_L2"
- },
- "source": [
- "#**Part 2: The OpenAI-Gym standard** 🤖 "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "Sv728rc3M_Ir"
- },
- "source": [
- "##**What is OpenAI-Gym?** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "AhQfyrBCUigq"
- },
- "source": [
- "[OpenAI-Gym](https://www.gymlibrary.ml/) is a software package that gathers a **collection of tasks** called environments with a **unique Python interface** to control all of them. This unique interface is a key feature in the software package, and has given rise to a standard for the format in which RL agents are developed and treated, independently of\n",
- "their internal functioning. The tasks defined in the Gym environments involve\n",
- "a wide variety of fields like video games, classic control theory problems, or high dimensional robotic locomotive tasks. You can find a list of available environment [here](https://www.gymlibrary.ml/environments/classic_control/).\n",
- "\n",
- "\n",
- "\n",
- "The OpenAI-Gym philosophy heavily relies on the episodic aspect of RL, i.e.\n",
- "the agent’s history is broken down into a series of experiences called **episodes** that may be of\n",
- "variable length. The agent interacts with the environment until it reaches a\n",
- "terminal state when the episode is finished. The goal is to maximize the total\n",
- "cumulative reward per episode.\n",
- "\n",
- "The main methods of the OpenAI-Gym interface are the following:\n",
- "\n",
- "- `obs = env.reset()` ➡️ The `reset` method is the one called first to initialize the environment `env` (whatever it is). The environment returns the first observation `obs` (state). \n",
- "- `next_obs,reward,done,info = env.step(action)` ➡️ The `step` method is used iteratively to interact with the environment. The RL agent computes an `action`, and the environment returns the next observation, associated reward, whether the episode is done (=terminated), and some other optional information. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "ntNSOBzJPJuF"
- },
- "source": [
- "##**Example using an OpenAI-Gym environment** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "mQ2879zIPOJ0"
- },
- "source": [
- "Now that we understand the main concepts of OpenAI-Gym we are going to illustrate its typical usage with a quick example. We're going to use the [CartPole environment](https://www.gymlibrary.ml/environments/classic_control/cart_pole/), which is one of the classic control problems available in the OpenAI-Gym framework.\n",
- "Let's start by installing the dependencies that we require: \n",
- "\n",
- ""
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 1,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "E0sfte45O8iN",
- "outputId": "de423891-b2c5-4901-f197-592ce81e3233"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
- "Requirement already satisfied: gym==0.21.0 in /usr/local/lib/python3.7/dist-packages (0.21.0)\n",
- "Requirement already satisfied: numpy>=1.18.0 in /usr/local/lib/python3.7/dist-packages (from gym==0.21.0) (1.21.6)\n",
- "Requirement already satisfied: importlib-metadata>=4.8.1 in /usr/local/lib/python3.7/dist-packages (from gym==0.21.0) (4.13.0)\n",
- "Requirement already satisfied: cloudpickle>=1.2.0 in /usr/local/lib/python3.7/dist-packages (from gym==0.21.0) (1.5.0)\n",
- "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata>=4.8.1->gym==0.21.0) (3.9.0)\n",
- "Requirement already satisfied: typing-extensions>=3.6.4 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata>=4.8.1->gym==0.21.0) (4.1.1)\n"
- ]
- }
- ],
- "source": [
- "!pip install gym==0.21.0"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "RnmDsSAnM_F-"
- },
- "source": [
- "**Cartpole environment description:**\n",
- "\n",
- "\"*A pole is attached by an un-actuated joint to a cart, which moves along a frictionless track. The pendulum is placed upright on the cart and the goal is to balance the pole by applying forces in the left (-1) and right (+1) direction on the cart. A reward of +1 is provided for every timestep that the pole remains upright.*\"\n",
- "\n",
- "\n",
- "You can also check out the physics of the environment in the [GitHub repository of OpenAI-Gym](https://github.com/openai/gym).\n",
- "See below an example of the evolution of an episode of the Cartpole environment. Note that most of the Gym envronments can be rendered to show how the RL agent is performing. \n",
- "\n",
- "\n",
- "\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "eMa2F3q_pV-C"
- },
- "source": [
- "First, we are going to import `gym` and then `make` our Cartpole environment (version 1). Note how it is possible to instantiate the registered Gym environments by referring to their names with a string. \n",
- "After that, we are going to interact with the environment for a maximum number of episodes of experience that we are going to indicate with `max_num_episodes`. In each episode, the environment is initialized with the `reset` method, and then we interact with the environment until the episode is `done`, which happens when the pole is down. \n",
- "\n",
- "It is important to note that in this example we are not implementing RL just yet. Instead, we are using the `sample` method from the action space of the environment to compute a random control action. This is useful when we want to quickly check how an environment behaves, but we should aim to replace that line by some intelligent RL agent able to compute a control action that optimizes the performance of the environment. "
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 2,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "LBxXhZc5nGb3",
- "outputId": "7ca9fb87-6573-4e69-ec2d-c9f5ac0025ad"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "\n",
- " Episode #1 had 17 steps and total_reward=17.0\n",
- "\n",
- " Episode #2 had 10 steps and total_reward=10.0\n",
- "\n",
- " Episode #3 had 19 steps and total_reward=19.0\n",
- "\n",
- " Episode #4 had 49 steps and total_reward=49.0\n",
- "\n",
- " Episode #5 had 15 steps and total_reward=15.0\n",
- "\n",
- " Episode #6 had 38 steps and total_reward=38.0\n",
- "\n",
- " Episode #7 had 28 steps and total_reward=28.0\n",
- "\n",
- " Episode #8 had 10 steps and total_reward=10.0\n",
- "\n",
- " Episode #9 had 38 steps and total_reward=38.0\n",
- "\n",
- " Episode #10 had 25 steps and total_reward=25.0\n"
- ]
- }
- ],
- "source": [
- "import gym\n",
- "\n",
- "env = gym.make('CartPole-v1')\n",
- "max_num_episodes = 10 # maximum number of episodes\n",
- "\n",
- "for episode in range(max_num_episodes):\n",
- " done = False\n",
- " obs = env.reset()\n",
- " total_reward = 0.0\n",
- " step = 0\n",
- " while not done:\n",
- " action = env.action_space.sample() # Compute random action. This is to be replaced by a RL algo\n",
- " obs,reward,done,info = env.step(action) # send the action to the environment\n",
- " total_reward += reward\n",
- " step += 1\n",
- "\n",
- " print('\\n Episode #{} had {} steps and total_reward={}'.format(episode+1,step,total_reward))\n",
- "\n",
- "env.close()"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "ai9dHGWksZeu"
- },
- "source": [
- "Notice how every episode lasts for a different number of steps because we are applying random forces to the cart. Also, notice how the total reward of each episode is equal to the number of steps, because the Cartpole environment gives a reward of +1 every timestep that we get to maintain the pole upright.\n",
- "\n",
- "\n",
- "\n",
- ""
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "tpeeOA8BM-5L"
- },
- "source": [
- "#**Part 3: The Building Optimization Testing (BOPTEST) framework** 🏠 "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "z0ry1NQwuMXa"
- },
- "source": [
- "Now that we understand how RL and OpenAI-Gym work, let's use that knowledge for the particular application of buildings. \n",
- "In this tutorial we are going to connect with a BOPTEST building emulator that we will use as our environment to control through RL.\n",
- "This emulator is a simulation model that was configured based on detailed physics and that has been peer-reviewed to ensure that it represents the behavior of an actual building as realistically as possible. Hence, although it is a simulation model, we are going to consider this emulator as the real building for control, testing and benchmarking. \n",
- "\n",
- "In this section we explain what BOPTEST is and how it can be generally used. Next section will exclusively focus on BOPTEST-Gym, the OpenAI-Gym interface of BOPTEST, to learn how we can use it to implement and assess RL algorithms for building control. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "OcPk7llkJP4m"
- },
- "source": [
- "##**What is BOPTEST?** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "HtDrzTFuJU0e"
- },
- "source": [
- "BOPTEST is a software framework enables the performance evaluation and benchmarking of advanced control algorithms for building HVAC control through simulations. The software is developed and is available on the BOPTEST GitHub respository at [https://github.com/ibpsa/project1-boptest](https://github.com/ibpsa/project1-boptest) \n",
- "\n",
- "and general information about BOPTEST can be found through the following link:\n",
- "\n",
- "[https://ibpsa.github.io/project1-boptest/](https://ibpsa.github.io/project1-boptest/)\n",
- "\n",
- "In the link below you can also find information about the overarching project that gave birth to BOPTEST, IBPSA Project 1:\n",
- "\n",
- "[https://ibpsa.github.io/project1/](https://ibpsa.github.io/project1/)\n",
- "\n",
- "\n",
- "\n",
- "\n",
- ""
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "sNiHr2w0IFYI"
- },
- "source": [
- "\n",
- "\n",
- "*Figure: The BOPTEST concept.*\n",
- "\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "Jj7kbbF8JXEG"
- },
- "source": [
- "The main use case of the BOPTEST framework is the one where a control developer wants to evaluate the performance of his/her building control strategy. Testing in a real building may be very expensive, or just not possible. BOPTEST offers a menu of emulator building models so that the control developer can select one of them, interact in co-simulation, and eventually assess the performance of his/her controller with a set of Key Performance Indicators (KPIs) that are calculated by the BOPTEST framework. \n",
- "\n",
- "Note that using a standardized set of building emulators, testing scenarios, and KPIs enables benchmarking, allows to compare across different controllers, and throws light on what are the best practices for building control. In addition, making these emulators easily and rapidly available to use allows for control developers without expertise in building modeling to utilize them for controls testing and evaluation. \n",
- "\n",
- "In this section we are going to explain the basic BOPTEST functionality to connect to a building test case and obtain available control inputs and measurement points. For a more complete description on how to use BOPTEST please visit this [BOPTEST Colab tutorial](https://github.com/ibpsa/project1-boptest/blob/master/docs/workshops/BS21Workshop_20210831/Introduction_to_the_BOPTEST_framework.ipynb). "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "l5Fgw7eJHEjy"
- },
- "source": [
- "##**Selecting a building test case** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "owb2Z2rqHEjz"
- },
- "source": [
- "BOPTEST test cases are developed as [Functional Mock-up Units (FMU's)](https://fmi-standard.org/) and deployed within a containerized environment using the [Docker](https://www.docker.com/) software with:\n",
- "\n",
- "* A detailed emulator **building model**.\n",
- "* Yearly **boundary condition data** for weather, schedules, pricing, and emission factors. These data are representative of the building location. \n",
- "* An **API** that allows for, among other things, initializing a simulation or testing scenario, advancing a simulation with a control input, receiving forecast data, receiving emulator data, and receiving computed KPIs. The full API is described [here](https://github.com/ibpsa/project1-boptest/tree/boptest-service#test-case-restful-api).\n",
- "\n",
- "The basic workflow to test a controller is:\n",
- "\n",
- "1. Select a **test case** from the menu of those available. \n",
- "2. Select one of the **testing scenarios** defined for the given test case. Testing scenarios are standardized for each emulator. \n",
- "3. Set **parameters** for the interaction with your test controller, such as the control step or forecast horizon, if required. \n",
- "4. Run the test case scenario in a **co-simulation** loop with your controller. \n",
- "5. Request the KPIs and **evaluate** your controller's performance. \n",
- "\n",
- "We start by selecting and launching a BOPTEST building test case from the [repository of currently available test cases](https://ibpsa.github.io/project1-boptest/testcases/index.html). In this example, we are going to work with the test case called `bestest_hydronic_heat_pump`, which is a single-zone residential building with radiant floor heating and a heat pump. This is a high-fidelity, yet, relatively simple test case that allows us to focus on fundamental aspects. You may want to note the other test cases available in the repository as well as the fact that there are more under development. \n",
- "\n",
- "We can launch our chosen test case as follows. First, import the Python `requests` library so that we can make HTTP requests to the BOPTEST API at the address indicated by the `url`. Then, use the `POST /testcases//select` BOPTEST API endpoint to launch the test case and receive a corresponding `testid`. While the `url` is the common gateway for everyone to access the BOPTEST web-service, the `testid` is a unique identifier for you to address the test case that you have selected and launched. "
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 3,
- "metadata": {
- "id": "V_qU6ukZghTb"
- },
- "outputs": [],
- "source": [
- "import requests\n",
- "\n",
- "# url for the BOPTEST service\n",
- "url = 'https://api.boptest.net' \n",
- "\n",
- "# Select test case and get identifier\n",
- "testcase = 'bestest_hydronic_heat_pump'\n",
- "\n",
- "# Check if already started a test case and stop it if so before starting another\n",
- "try:\n",
- " requests.put('{0}/stop/{1}'.format(url, testid))\n",
- "except:\n",
- " pass\n",
- "\n",
- "# Select and start a new test case\n",
- "testid = \\\n",
- "requests.post('{0}/testcases/{1}/select'.format(url,testcase)).json()['testid']\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "eRZGKWDlHEj2"
- },
- "source": [
- "Please do not get distracted by the `try-except` statement. We are using that one to stop already created test cases if we are revisiting this cell. This prevents from having several dangling test cases that can overwhelm our server. \n",
- "\n",
- "Once you have successfully obtained the `testid`, it is possible to start interacting with your selected test case using the rest of the BOPTEST API. You will need this `testid` for all further interactions with this test case. For example, use the `GET /name` BOPTEST API endpoint, along with your `testid`, to request the name of your test case and check that it matches the one we want."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 4,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "8mdK5JtNI-e_",
- "outputId": "d2d023e0-eb35-4122-cbd6-11420f694521"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "{'name': 'bestest_hydronic_heat_pump'}\n"
- ]
- }
- ],
- "source": [
- "# Get test case name\n",
- "name = requests.get('{0}/name/{1}'.format(url, testid)).json()['payload']\n",
- "print(name)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "gUOQXYjlHEj3"
- },
- "source": [
- "With our unique `testid` in-hand and having some practice using the BOPTEST API, we are ready to move on to start using our building emulator. For this tutorial, we are going to explain only how to obtain information about the building using the BOPTEST API before moving to learn BOPTEST-Gym. \n",
- "Note that the test case will timeout after 15 minutes of no requests. If the test case times out, you can simply select and start a new one by repeating the steps described above.\n",
- "\n",
- "\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "jmglIZGFHEj3"
- },
- "source": [
- "##**Obtaining general information about the building** \n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "mJ6leLGvRJya"
- },
- "source": [
- "The first thing we want to do is learn about the building and system that we want to control. All building information can be found under documentation provided for each specific test case on the [Test Cases tab](https://ibpsa.github.io/project1-boptest/testcases/index.html) of the BOPTEST website. \n",
- "\n",
- "The building information includes a description of the building envelope, the HVAC system design, the functioning of the baseline controller, available control inputs and measurement outputs, and available testing scenarios. Understanding how the system works is an important practice for control design, so take as much time as needed to understand the equipment, the points that can be measured, and the points that can be overwritten by your controller. \n",
- "We briefly summarize the `bestest_hydronic_heat_pump` case here for completeness, but it is strongly recommended to have a deeper look into the [documentation](https://ibpsa.github.io/project1-boptest/testcases/ibpsa/testcases_ibpsa_bestest_hydronic_heat_pump/). \n",
- "\n",
- "The building represents a residential dwelling of 192 $m^2$ for a family of 5 members. \n",
- "An air-to-water modulating heat pump of 15 $kW$ nominal heating capacity extracts energy from the ambient air to heat up the floor heating emission system, as shown in the figure below. \n",
- "An evaporator fan blows ambient air through the heat pump evaporator when the heat pump is operating. \n",
- "The floor heating system injects heat into the floor using water as the working fluid."
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "SMQcNDl1HEj4"
- },
- "source": [
- "\n",
- "\n",
- "\n",
- "*Figure: Schematic of HVAC system and control for the `bestest_hydronic_heat_pump` test case.*"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "lNLRXp2eHEj4"
- },
- "source": [
- "A baseline controller is embedded in every test case emulator that is meant to be representative of a typical controller for that type of building. The baseline controller includes local loop control such that supervisory set points may be the focus of a test controller, although many of those local loop control signals are also available for overwriting if a user chooses. The baseline controller can also be considered an initial benchmark for control performance. \n",
- "\n",
- "In our selected test case, the baseline controller consists of a PI controller with the zone operative temperature as the controlled variable and the heat pump modulation signal for compressor frequency as the control variable, as depicted as C1 in the figure above and shown in the figure below. \n",
- "The control variable is limited between 0 and 1, and it is computed to drive the zone operative temperature towards its set point, which is defined as a function of the occupancy schedule. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "RqVtoDgTHEj4"
- },
- "source": [
- "\n",
- "\n",
- "\n",
- "*Figure: Primary PI controller C1.*\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "ryOicW_KHEj5"
- },
- "source": [
- "All other equipment (fan for the heat pump evaporator circuit and floor heating emission system pump) are switched on when the heat pump is working (modulating signal higher than 0) and switched off otherwise. This is depicted in the figure of the HVAC schematic as controller C2."
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "-J5j60bRHEj5"
- },
- "source": [
- "##**Getting control input and measurement points** \n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "AGG7G6VeR4WB"
- },
- "source": [
- "While control input and measurement points are described in the documentation, they are also available to retreive from the BOPTEST API. This is especially useful to store for later when requesting data for a specific point.\n",
- "\n",
- "Retrieve the control input and measurement outputs using the `GET /inputs` and `GET /measurements` BOPTEST API endpoints."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 5,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "0IKRxBykJY6u",
- "outputId": "cdf6b6dc-0bf4-426a-c0c5-a55f0951f599"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "TEST CASE INPUTS ---------------------------------------------\n",
- "dict_keys(['oveTSet_activate', 'ovePum_activate', 'ovePum_u', 'oveHeaPumY_u', 'oveTSet_u', 'oveHeaPumY_activate', 'oveFan_activate', 'oveFan_u'])\n",
- "TEST CASE MEASUREMENTS ---------------------------------------\n",
- "dict_keys(['weaSta_reaWeaPAtm_y', 'reaPFan_y', 'reaQHeaPumCon_y', 'reaTRet_y', 'weaSta_reaWeaNOpa_y', 'weaSta_reaWeaTBlaSky_y', 'reaQHeaPumEva_y', 'weaSta_reaWeaNTot_y', 'weaSta_reaWeaSolAlt_y', 'reaTZon_y', 'weaSta_reaWeaHHorIR_y', 'weaSta_reaWeaLon_y', 'weaSta_reaWeaSolTim_y', 'weaSta_reaWeaCloTim_y', 'reaPPumEmi_y', 'weaSta_reaWeaHGloHor_y', 'weaSta_reaWeaHDifHor_y', 'weaSta_reaWeaRelHum_y', 'reaTSetHea_y', 'reaCO2RooAir_y', 'weaSta_reaWeaSolDec_y', 'reaPHeaPum_y', 'weaSta_reaWeaHDirNor_y', 'reaTSetCoo_y', 'weaSta_reaWeaWinDir_y', 'reaTSup_y', 'weaSta_reaWeaSolZen_y', 'reaQFloHea_y', 'reaCOP_y', 'weaSta_reaWeaTDryBul_y', 'weaSta_reaWeaTWetBul_y', 'weaSta_reaWeaTDewPoi_y', 'weaSta_reaWeaWinSpe_y', 'weaSta_reaWeaLat_y', 'weaSta_reaWeaCeiHei_y', 'weaSta_reaWeaSolHouAng_y'])\n"
- ]
- }
- ],
- "source": [
- "# Get inputs available\n",
- "inputs = requests.get('{0}/inputs/{1}'.format(url, testid)).json()['payload']\n",
- "print('TEST CASE INPUTS ---------------------------------------------')\n",
- "print(inputs.keys())\n",
- "# Get measurements available\n",
- "print('TEST CASE MEASUREMENTS ---------------------------------------')\n",
- "measurements = requests.get('{0}/measurements/{1}'.format(url, testid)).json()['payload']\n",
- "print(measurements.keys())"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "A4L6Gw6YJU5L"
- },
- "source": [
- "The naming convention is such that the extension `_y` indicates a measurement point, `_u` indicates the value of an input which can be overwritten by a test controller, and `_activate` indicates the enabling (with value 0 or 1) of a test controller to overwrite the corresponding input value. \n",
- "Hence, `_u` is enabled for overwriting by the test controller when `_activate=1`.\n",
- "`weaSta_` indicates a measurement for a weather point, so that historical weather data can be easily retrieved.\n",
- "\n",
- "Notice that the jsons returned from the `GET /inputs` and `GET /measurements` BOPTEST API endpoints also include a description and unit of each variable, as well as the minimum and maximum value for inputs variables:"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "U7guJ_I10QOF"
- },
- "source": [
- "Now let's stop the test case since we are not going to use it for a while. We do this to not overwhelm the server."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 6,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "v5_1Q_H80Z5k",
- "outputId": "c2e714bb-ef91-4769-c8c4-76465733548d"
- },
- "outputs": [
- {
- "output_type": "execute_result",
- "data": {
- "text/plain": [
- ""
- ]
- },
- "metadata": {},
- "execution_count": 6
- }
- ],
- "source": [
- "requests.put('{0}/stop/{1}'.format(url, testid))"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "UHYHM9MjSz_C"
- },
- "source": [
- "# **Part 4: Implementing RL for a building with BOPTEST-Gym** 🤖 🏠 "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "BEC76h9HT7gL"
- },
- "source": [
- "##**What is BOPTEST-Gym?** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "z7A9k7GFUBsK"
- },
- "source": [
- "BOPTEST-Gym is the OpenAI-Gym interface of BOPTEST that helps to train RL agents for the application of building climate control.\n",
- "The BOPTEST-Gym interface accomodates the BOPTEST API to have BOPTEST building emulators as environments that follow the OpenAI-Gym standard. \n",
- "Therefore, the BOPTEST-Gym interface facilitates the development of RL agents as it allows interacting with the BOPTEST building emulators with a standard that is very well known by the machine learning community. Or even better, it allows us to directly use existing RL agents that have been developed following this standard, like those from the [Stable Baselines 3](https://stable-baselines3.readthedocs.io/en/master/) repository.\n",
- "\n",
- "You can find more information about BOPTEST-Gym in [this paper](https://publications.ibpsa.org/conference/paper/?id=bs2021_30380), but here we summarize the main points you should know:\n",
- "- BOPTEST-Gym enables the interaction of RL agents with a set of physics-based and highly **detailed building models** to assess RL for the application of building climate control. \n",
- "- All **hyperparameters** of the environment are initialized when the environment is instantiated. A particularly relevant hyperparameter is `testcase`, a string specifying the BOPTEST emulator of choice. This string selects the building model from the [menu of BOPTEST building emulators](https://ibpsa.github.io/project1-boptest/testcases/index.html). \n",
- "- The **state** of any building emulator environment can have a *time* component e.g. a weekly schedule, a *measurement* component with a subset (or all) measurements available in the building, and an *exogenous* component including disturbances of any kind of boundary condition data to the building such as electricity prices, ambient temperature, or temperature set-points. \n",
- "- The **action** space is defined based on any subset (or all) inputs available to the emulator. These can be either building set-points, like zone\n",
- "operative temperature set-points, or lower level actuator signals, such as heat\n",
- "pump modulating signal or a pump stage.\n",
- "- The **`reset()`** method is called at the beginning of every episode to return the environment to a logical initial state. \n",
- "- The **`step()`** method is called every time step to take the action computed by the RL agent, overwrite the building inputs with the vector of action values and advance the building simulation model during one time step period. BOPTEST-Gym also has wrappers for discretization of the state and action spaces. This functionality comes in handy when training RL agents. \n",
- "- A default **reward** function is implemented in the `compute_reward` method of the BOPTEST-Gym environment that can be overwritten. It is convenient to use the BOPTEST `/kpis` API to obtain the KPI values at the present time for defining custom reward functions. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "kvMgiRhLX2i8"
- },
- "source": [
- "##**Starting up a BOPTEST-Gym environment** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "WcrOX1Z_UvTY"
- },
- "source": [
- "BOPTEST-Gym uses RL algorithms from the [Stable Baselines 3](https://stable-baselines3.readthedocs.io/en/master/) repository to exemplify and test its functionality. Therefore, we need to install stable-baselines3.\n",
- " "
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 7,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "jpZk6qJKTuYl",
- "outputId": "8b5166ec-7c96-480a-84bc-732bc7521f3a"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
- "Requirement already satisfied: stable-baselines3==0.8.0 in /usr/local/lib/python3.7/dist-packages (0.8.0)\n",
- "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from stable-baselines3==0.8.0) (3.2.2)\n",
- "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from stable-baselines3==0.8.0) (1.21.6)\n",
- "Requirement already satisfied: cloudpickle in /usr/local/lib/python3.7/dist-packages (from stable-baselines3==0.8.0) (1.5.0)\n",
- "Requirement already satisfied: gym>=0.17 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3==0.8.0) (0.21.0)\n",
- "Requirement already satisfied: pandas in /usr/local/lib/python3.7/dist-packages (from stable-baselines3==0.8.0) (1.3.5)\n",
- "Requirement already satisfied: torch>=1.4.0 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3==0.8.0) (1.12.1+cu113)\n",
- "Requirement already satisfied: importlib-metadata>=4.8.1 in /usr/local/lib/python3.7/dist-packages (from gym>=0.17->stable-baselines3==0.8.0) (4.13.0)\n",
- "Requirement already satisfied: typing-extensions>=3.6.4 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata>=4.8.1->gym>=0.17->stable-baselines3==0.8.0) (4.1.1)\n",
- "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata>=4.8.1->gym>=0.17->stable-baselines3==0.8.0) (3.9.0)\n",
- "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3==0.8.0) (1.4.4)\n",
- "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3==0.8.0) (2.8.2)\n",
- "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3==0.8.0) (3.0.9)\n",
- "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3==0.8.0) (0.11.0)\n",
- "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.1->matplotlib->stable-baselines3==0.8.0) (1.15.0)\n",
- "Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.7/dist-packages (from pandas->stable-baselines3==0.8.0) (2022.4)\n"
- ]
- }
- ],
- "source": [
- "!pip install stable-baselines3==0.8.0 "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "4ljXzd7W4R0H"
- },
- "source": [
- "Now that we have all package dependencies, let's clone the BOPTEST-Gym repository. We are going to clone the `boptest-gym-service` branch which works in the same way as the `master` branch but allows us to directly use the web-based version of BOPTEST that is readily available such that we do not have to deploy the building test case Docker containers locally. \n"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 8,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "McqIwvAz5ZuD",
- "outputId": "e734b733-5092-4ed6-bb0d-c97d61e3df40"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "Cloning into 'project1-boptest-gym'...\n",
- "remote: Enumerating objects: 2896, done.\u001b[K\n",
- "remote: Counting objects: 100% (442/442), done.\u001b[K\n",
- "remote: Compressing objects: 100% (268/268), done.\u001b[K\n",
- "remote: Total 2896 (delta 213), reused 345 (delta 153), pack-reused 2454\u001b[K\n",
- "Receiving objects: 100% (2896/2896), 46.80 MiB | 22.24 MiB/s, done.\n",
- "Resolving deltas: 100% (1490/1490), done.\n"
- ]
- }
- ],
- "source": [
- "try:\n",
- " !rm -rf project1-boptest-gym\n",
- "except:\n",
- " pass\n",
- "!git clone -b boptest-gym-service https://github.com/ibpsa/project1-boptest-gym.git"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "cs9guwYo5w50"
- },
- "source": [
- "Now we move our working directory to our recently cloned repository, import the `BoptestGymEnv` class, and instantiate our first BOPTEST-Gym environment! "
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 9,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "mZsXUZIQ5iIj",
- "outputId": "76d5b60a-9b92-4f98-82e2-b4618b5dfb7e"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stderr",
- "text": [
- "/usr/local/lib/python3.7/dist-packages/gym/spaces/box.py:74: UserWarning: \u001b[33mWARN: Box bound precision lowered by casting to float32\u001b[0m\n",
- " \"Box bound precision lowered by casting to {}\".format(self.dtype)\n"
- ]
- }
- ],
- "source": [
- "import os\n",
- "os.chdir('/content/project1-boptest-gym')\n",
- "from boptestGymEnv import BoptestGymEnv\n",
- "\n",
- "# Instantite environment\n",
- "env = BoptestGymEnv(url = url,\n",
- " testcase = 'bestest_hydronic_heat_pump',\n",
- " actions = ['oveHeaPumY_u'],\n",
- " observations = {'reaTZon_y':(280.,310.)}, \n",
- " random_start_time = False,\n",
- " start_time = 31*24*3600,\n",
- " max_episode_length = 24*3600,\n",
- " warmup_period = 24*3600,\n",
- " step_period = 3600)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "8XVI61rnU4QZ"
- },
- "source": [
- "You have connected to a BOPTEST building emulator and wrapped it around a Gym environment. Let's examine more in detail the arguments that you have used:\n",
- "- `url`: the domain where your test case lives. In this case it is the url to BOPTEST-service, but it could be your localhost if you decide to spin a test case in your machine using Docker. \n",
- "- `testcase`: The string identifier of the testcase.\n",
- "- `actions`: List of strings indicating the action space. \n",
- "- `observations`: Dictionary mapping observation keys to a tuple with the lower and upper bound of each observation. These bounds define the typical operational range for discretization and normalization purposes. Observation keys must belong either to the set of measurements or to the set of forecasting variables of the BOPTEST test case.\n",
- "- `max_episode_lenght`: Maximum duration of each episode in seconds.\n",
- "- `random_start_time`: Set to True if desired to use a random start time for each episode. That is typically usefull when training an RL agent to run several episodes with different boundary condition data. In our case, we set it to False and specify the start time of the episode.\n",
- "- `start_time`: start time of the episode. It is specified in seconds from the beginning of the year. To be used in combination with `random_start_time=False`. \n",
- "- `warmup_period`: Desired simulation period to initialize each episode, in seconds. In our case, we simulate the testcase for one day right before the beginning of the episode. \n",
- "- `step_period`: The period of each control step, in seconds. In this case is set to one hour. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "-ZcNOH0SYEiR"
- },
- "source": [
- "Now you can interact with the building emulator following the Gym standard. Everytime you use one of the methods of your environment, BOPTEST-Gym will send the associated commands through the BOPTEST API that you have learned above as to provide the desired functionality. A schematic of this process is shown in the figure below. This figure illustrates the typical steps that take place when training an agent and the mapping between the BOPTEST-Gym interface and the BOPTEST API. It is important to note that a state can be returned not only with current measurements, but also with boundary condition forecast or regressive values. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "rcnaIJvhYDa5"
- },
- "source": [
- "\n",
- "\n",
- "*Figure: Sequence diagram for training an agent withthe BOPTEST-Gym environment.*"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "AZVIz69qXyCZ"
- },
- "source": [
- "##**Interacting with a BOPTEST-Gym environment** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "dA9_wVo8bMxr"
- },
- "source": [
- "Let's see what we can do with our building Gym environment. Recall that the first step is using the `reset` method to simulate the building right before the episode start time a time period specified in `warmup_period`. This will bring the building to a reasonable initial state and the environment will return an observation `obs` which, in our case, it is comprised of only the zone operative temperature (`reaTZon_y`). This temperature is in Kelvins, so we convert it to degrees Celsius."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 10,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "Z4n5GsjXV08x",
- "outputId": "8c37e26f-e97d-4b4a-d41d-792fc143dbbf"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "Zone temperature: 21.37 degC\n",
- "Episode starting day: 31.0 (from beginning of the year)\n"
- ]
- }
- ],
- "source": [
- "obs = env.reset()\n",
- "print('Zone temperature: {:.2f} degC'.format(obs[0]-273.15))\n",
- "print('Episode starting day: {:.1f} (from beginning of the year)'.format(env.start_time/24/3600)) "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "IeJyBvLYqC11"
- },
- "source": [
- "📌 **Note: About initialization** \n",
- "\n",
- "The initial state in the emulator consists of all states after simulation during the warmup period without any external input from an external controller. This particular emulator has 63 continuous time states comprising temperatures of walls, floor, roof, water, etc. During the warmup period, the baseline controller embedded in the emulator is used. After initialization the baseline controller will also work at any time unless some of the control variables are intentionally overwritten by an external controller. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "9i9VfDrdYJ0e"
- },
- "source": [
- "We can inspect the observation and action space of any environment as follows:"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 11,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "a_PC0YAEYR5U",
- "outputId": "dc3c02a9-e656-4adf-a4d9-884f58e8d587"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "Observation space of the building environment:\n",
- "Box([280.], [310.], (1,), float32)\n",
- "Action space of the building environment:\n",
- "Box([0.], [1.], (1,), float32)\n"
- ]
- }
- ],
- "source": [
- "print('Observation space of the building environment:')\n",
- "print(env.observation_space)\n",
- "print('Action space of the building environment:')\n",
- "print(env.action_space)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "SBVCnncbePIQ"
- },
- "source": [
- "So this environment has a Box (continuous and bounded) observation space which is the indoor building temperature. The operational range of this variable goes from $280$ $K$ to $310$ $K$. That is, from ~$7$ $°C$ to $37$ $°C$. On the other hand, the action space is a continuous variable that goes from $0$ to $1$. The latter variable represents the heat pump compressor frequency with $0$ meaning no heating, and $1$ meaning the heat pump working at full capacity. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "pkx5Os6-Yltb"
- },
- "source": [
- "But actually, the BOPTEST-Gym environment can be directly printed to show a lot of useful information to control the building:"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 12,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "FGzL_ZskfoyO",
- "outputId": "470e1fa7-b197-487d-ea2b-2bae9b9b2191"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "\n",
- "========================\n",
- "BOPTEST CASE INFORMATION\n",
- "========================\n",
- "\n",
- "Test case name\n",
- "--------------\n",
- "{'name': 'bestest_hydronic_heat_pump'}\n",
- "\n",
- "All measurement variables\n",
- "-------------------------\n",
- "{'reaCO2RooAir_y': {'Description': 'CO2 concentration in the zone',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'ppm'},\n",
- " 'reaCOP_y': {'Description': 'Heat pump COP',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': '1'},\n",
- " 'reaPFan_y': {'Description': 'Electrical power of the heat pump evaporator '\n",
- " 'fan',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'W'},\n",
- " 'reaPHeaPum_y': {'Description': 'Heat pump electrical power',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'W'},\n",
- " 'reaPPumEmi_y': {'Description': 'Emission circuit pump electrical power',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'W'},\n",
- " 'reaQFloHea_y': {'Description': 'Floor heating thermal power released to the '\n",
- " 'zone',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'W'},\n",
- " 'reaQHeaPumCon_y': {'Description': 'Heat pump thermal power exchanged in the '\n",
- " 'condenser',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'W'},\n",
- " 'reaQHeaPumEva_y': {'Description': 'Heat pump thermal power exchanged in the '\n",
- " 'evaporator',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'W'},\n",
- " 'reaTRet_y': {'Description': 'Return water temperature from radiant floor',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'K'},\n",
- " 'reaTSetCoo_y': {'Description': 'Zone operative temperature setpoint for '\n",
- " 'cooling',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'K'},\n",
- " 'reaTSetHea_y': {'Description': 'Zone operative temperature setpoint for '\n",
- " 'heating',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'K'},\n",
- " 'reaTSup_y': {'Description': 'Supply water temperature to radiant floor',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'K'},\n",
- " 'reaTZon_y': {'Description': 'Zone operative temperature',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'K'},\n",
- " 'weaSta_reaWeaCeiHei_y': {'Description': 'Cloud cover ceiling height '\n",
- " 'measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'm'},\n",
- " 'weaSta_reaWeaCloTim_y': {'Description': 'Day number with units of seconds',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 's'},\n",
- " 'weaSta_reaWeaHDifHor_y': {'Description': 'Horizontal diffuse solar radiation '\n",
- " 'measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'W/m2'},\n",
- " 'weaSta_reaWeaHDirNor_y': {'Description': 'Direct normal radiation '\n",
- " 'measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'W/m2'},\n",
- " 'weaSta_reaWeaHGloHor_y': {'Description': 'Global horizontal solar '\n",
- " 'irradiation measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'W/m2'},\n",
- " 'weaSta_reaWeaHHorIR_y': {'Description': 'Horizontal infrared irradiation '\n",
- " 'measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'W/m2'},\n",
- " 'weaSta_reaWeaLat_y': {'Description': 'Latitude of the location',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'rad'},\n",
- " 'weaSta_reaWeaLon_y': {'Description': 'Longitude of the location',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'rad'},\n",
- " 'weaSta_reaWeaNOpa_y': {'Description': 'Opaque sky cover measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': '1'},\n",
- " 'weaSta_reaWeaNTot_y': {'Description': 'Sky cover measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': '1'},\n",
- " 'weaSta_reaWeaPAtm_y': {'Description': 'Atmospheric pressure measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'Pa'},\n",
- " 'weaSta_reaWeaRelHum_y': {'Description': 'Outside relative humidity '\n",
- " 'measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': '1'},\n",
- " 'weaSta_reaWeaSolAlt_y': {'Description': 'Solar altitude angle measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'rad'},\n",
- " 'weaSta_reaWeaSolDec_y': {'Description': 'Solar declination angle measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'rad'},\n",
- " 'weaSta_reaWeaSolHouAng_y': {'Description': 'Solar hour angle measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'rad'},\n",
- " 'weaSta_reaWeaSolTim_y': {'Description': 'Solar time',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 's'},\n",
- " 'weaSta_reaWeaSolZen_y': {'Description': 'Solar zenith angle measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'rad'},\n",
- " 'weaSta_reaWeaTBlaSky_y': {'Description': 'Black-body sky temperature '\n",
- " 'measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'K'},\n",
- " 'weaSta_reaWeaTDewPoi_y': {'Description': 'Dew point temperature measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'K'},\n",
- " 'weaSta_reaWeaTDryBul_y': {'Description': 'Outside drybulb temperature '\n",
- " 'measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'K'},\n",
- " 'weaSta_reaWeaTWetBul_y': {'Description': 'Wet bulb temperature measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'K'},\n",
- " 'weaSta_reaWeaWinDir_y': {'Description': 'Wind direction measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'rad'},\n",
- " 'weaSta_reaWeaWinSpe_y': {'Description': 'Wind speed measurement',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': 'm/s'}}\n",
- "\n",
- "All forecasting variables\n",
- "-------------------------\n",
- "['winDir',\n",
- " 'TDewPoi',\n",
- " 'LowerSetp[1]',\n",
- " 'PriceElectricPowerConstant',\n",
- " 'UpperSetp[1]',\n",
- " 'PriceElectricPowerHighlyDynamic',\n",
- " 'solTim',\n",
- " 'solHouAng',\n",
- " 'nOpa',\n",
- " 'InternalGainsRad[1]',\n",
- " 'nTot',\n",
- " 'HGloHor',\n",
- " 'winSpe',\n",
- " 'TBlaSky',\n",
- " 'solDec',\n",
- " 'lon',\n",
- " 'PriceElectricPowerDynamic',\n",
- " 'HDifHor',\n",
- " 'InternalGainsCon[1]',\n",
- " 'solZen',\n",
- " 'HHorIR',\n",
- " 'relHum',\n",
- " 'pAtm',\n",
- " 'Occupancy[1]',\n",
- " 'ceiHei',\n",
- " 'lat',\n",
- " 'InternalGainsLat[1]',\n",
- " 'TWetBul',\n",
- " 'TDryBul',\n",
- " 'HDirNor',\n",
- " 'EmissionsElectricPower',\n",
- " 'cloTim',\n",
- " 'solAlt',\n",
- " 'time',\n",
- " 'UpperCO2[1]']\n",
- "\n",
- "All input variables\n",
- "-------------------\n",
- "{'oveFan_activate': {'Description': 'Activation for Integer signal to control '\n",
- " 'the heat pump evaporator fan either on or '\n",
- " 'off',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': None},\n",
- " 'oveFan_u': {'Description': 'Integer signal to control the heat pump '\n",
- " 'evaporator fan either on or off',\n",
- " 'Maximum': 1,\n",
- " 'Minimum': 0,\n",
- " 'Unit': '1'},\n",
- " 'oveHeaPumY_activate': {'Description': 'Activation for Heat pump modulating '\n",
- " 'signal for compressor speed between 0 '\n",
- " '(not working) and 1 (working at '\n",
- " 'maximum capacity)',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': None},\n",
- " 'oveHeaPumY_u': {'Description': 'Heat pump modulating signal for compressor '\n",
- " 'speed between 0 (not working) and 1 (working '\n",
- " 'at maximum capacity)',\n",
- " 'Maximum': 1,\n",
- " 'Minimum': 0,\n",
- " 'Unit': '1'},\n",
- " 'ovePum_activate': {'Description': 'Activation for Integer signal to control '\n",
- " 'the emission circuit pump either on or '\n",
- " 'off',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': None},\n",
- " 'ovePum_u': {'Description': 'Integer signal to control the emission circuit '\n",
- " 'pump either on or off',\n",
- " 'Maximum': 1,\n",
- " 'Minimum': 0,\n",
- " 'Unit': '1'},\n",
- " 'oveTSet_activate': {'Description': 'Activation for Zone operative '\n",
- " 'temperature setpoint',\n",
- " 'Maximum': None,\n",
- " 'Minimum': None,\n",
- " 'Unit': None},\n",
- " 'oveTSet_u': {'Description': 'Zone operative temperature setpoint',\n",
- " 'Maximum': 308.15,\n",
- " 'Minimum': 278.15,\n",
- " 'Unit': 'K'}}\n",
- "\n",
- "Default simulation step (seconds)\n",
- "---------------------------------\n",
- "3600\n",
- "\n",
- "Default forecasting parameters (seconds)\n",
- "----------------------------------------\n",
- "{'horizon': 86400, 'interval': 3600}\n",
- "\n",
- "Default scenario\n",
- "----------------\n",
- "{'electricity_price': 'constant'}\n",
- "\n",
- "Test case scenario\n",
- "------------------\n",
- "{'electricity_price': 'constant'}\n",
- "\n",
- "===========================\n",
- "GYM ENVIRONMENT INFORMATION\n",
- "===========================\n",
- "\n",
- "Observation space\n",
- "-----------------\n",
- "Box([280.], [310.], (1,), float32)\n",
- "\n",
- "Action space\n",
- "------------\n",
- "Box([0.], [1.], (1,), float32)\n",
- "\n",
- "Is a regressive environment\n",
- "---------------------------\n",
- "False\n",
- "\n",
- "Is a predictive environment\n",
- "---------------------------\n",
- "False\n",
- "\n",
- "Regressive period (seconds)\n",
- "---------------------------\n",
- "None\n",
- "\n",
- "Predictive period (seconds)\n",
- "---------------------------\n",
- "None\n",
- "\n",
- "Measurement variables used in observation space\n",
- "-----------------------------------------------\n",
- "['reaTZon_y']\n",
- "\n",
- "Predictive variables used in observation space\n",
- "----------------------------------------------\n",
- "[]\n",
- "\n",
- "Sampling time (seconds)\n",
- "-----------------------\n",
- "3600\n",
- "\n",
- "Random start time\n",
- "-----------------\n",
- "False\n",
- "\n",
- "Excluding periods (seconds from the beginning of the year)\n",
- "----------------------------------------------------------\n",
- "None\n",
- "\n",
- "Warmup period for each episode (seconds)\n",
- "----------------------------------------\n",
- "86400\n",
- "\n",
- "Maximum episode length (seconds)\n",
- "--------------------------------\n",
- "86400\n",
- "\n",
- "Environment reward function (source code)\n",
- "-----------------------------------------\n",
- "(' def compute_reward(self):\\n'\n",
- " \" '''\\n\"\n",
- " \" Compute the reward of last state-action-state' tuple. The \\n\"\n",
- " ' reward is implemented as the negated increase in the objective\\n'\n",
- " ' integrand function. In turn, this objective integrand function \\n'\n",
- " ' is calculated as the sum of the total operational cost plus\\n'\n",
- " ' the weighted discomfort. \\n'\n",
- " ' \\n'\n",
- " ' Returns\\n'\n",
- " ' -------\\n'\n",
- " ' Reward: float\\n'\n",
- " \" Reward of last state-action-state' tuple\\n\"\n",
- " ' \\n'\n",
- " ' Notes\\n'\n",
- " ' -----\\n'\n",
- " ' This method is just a default method to compute reward. It can be \\n'\n",
- " ' overridden by defining a child from this class with\\n'\n",
- " ' this same method name, i.e. `compute_reward`. If a custom reward \\n'\n",
- " ' is defined, it is strongly recommended to derive it using the KPIs\\n'\n",
- " ' as returned from the BOPTEST framework, as it is done in this \\n'\n",
- " ' default `compute_reward` method. This ensures that all variables \\n'\n",
- " ' that may contribute to any KPI are properly accounted and \\n'\n",
- " ' integrated. \\n'\n",
- " ' \\n'\n",
- " \" '''\\n\"\n",
- " ' \\n'\n",
- " ' # Define a relative weight for the discomfort \\n'\n",
- " ' w = 1\\n'\n",
- " ' \\n'\n",
- " ' # Compute BOPTEST core kpis\\n'\n",
- " ' kpis = '\n",
- " \"requests.get('{0}/kpi/{1}'.format(self.url,self.testid)).json()['payload']\\n\"\n",
- " ' \\n'\n",
- " ' # Calculate objective integrand function at this point\\n'\n",
- " \" objective_integrand = kpis['cost_tot'] + w*kpis['tdis_tot']\\n\"\n",
- " ' \\n'\n",
- " ' # Compute reward\\n'\n",
- " ' reward = -(objective_integrand - self.objective_integrand)\\n'\n",
- " ' \\n'\n",
- " ' self.objective_integrand = objective_integrand\\n'\n",
- " ' \\n'\n",
- " ' return reward\\n')\n",
- "\n",
- "Environment hierarchy\n",
- "---------------------\n",
- "(,\n",
- " ,\n",
- " )\n",
- "\n",
- "\n"
- ]
- }
- ],
- "source": [
- "print(env)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "-KYf1BksgfQj"
- },
- "source": [
- "Note that this descriptive summary provides information not only about the Gym environment but also all information about the original BOPTEST test case. This may be useful, for example, if we want to extend our observation space or if we want to change our control action. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "zQ_y22lrg1cM"
- },
- "source": [
- "BOPTEST-Gym comes along with other functionality that may be useful when training RL agents, like the capacity to discretize and normalize observation and action spaces. For instance, we are dealing now with continuous action environment meaning that the agent could decide to take any action between 0 and 1. However, it is probably helpful to the agent to decide on just whether the heating needs to be turned on (action=1) or off (action=0). For that, we can wrap our environment around a discretization wrapper with only one action bin (one bin has two extremes). The concept of wrappers is very powerful in Gym environments. With them, we are capable to customize observation, action, step function, etc. of an environment. No matter how many wrappers are applied, `env.unwrapped` always gives back the internal original environment object. Let's see how it works with BOPTEST-Gym:"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 13,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "zIqfeNwgh9VK",
- "outputId": "d58259b5-ccd1-4fcb-98ea-171224f8141a"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "Action space of the wrapped agent:\n",
- "Discrete(2)\n",
- "Action space of the original agent:\n",
- "Box([0.], [1.], (1,), float32)\n"
- ]
- }
- ],
- "source": [
- "from boptestGymEnv import DiscretizedActionWrapper\n",
- "env = DiscretizedActionWrapper(env,n_bins_act=1)\n",
- "print('Action space of the wrapped agent:')\n",
- "print(env.action_space)\n",
- "print('Action space of the original agent:')\n",
- "print(env.unwrapped.action_space)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "Ghlx_zaf282q"
- },
- "source": [
- "Another thing that we can do is to interact with the building environment for one episode of experience (one day). This is similar to what we did with the Cartpole example, but this time we are going to run just one episode and use a hysteresis controller that will turn on the heating the temperature is below a predefined temperature setpoint, and turn it off when the temperature goes above the setpoint. We first configure such controller:"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 14,
- "metadata": {
- "id": "MrO0o7hNf5pB"
- },
- "outputs": [],
- "source": [
- "import numpy as np\n",
- "np.set_printoptions(precision=3)\n",
- "\n",
- "class SimpleController(object):\n",
- " '''Simple controller for this emulator. \n",
- " \n",
- " '''\n",
- " def __init__(self, TSet=22+273.15):\n",
- " self.TSet = TSet\n",
- " \n",
- " def predict(self, obs):\n",
- " # Compute control\n",
- " if obs[0]"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "JAIt_IfivAHN"
- },
- "source": [
- "In this section we are going to develop a very simple RL agent based on the very well known *q-learning* algorithm. Although simple, this exercise will help us understand the main concepts of RL and how this machine learning technique can be helpful to mitigate climate change by enhancing building's operational efficiency. Recall that our objective is to develop an RL agent that can decide on the best action to take in each situation (each state) just from interactions with the environment (the building). Imagine we are at time $k$ in a certain state $\\pmb{s}$ and take an action $\\pmb{a}$. In return, we obtain a reward $r'$ the next time step and end up in a state $\\pmb{s}'$ :"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "nxO3gVx3vqrL"
- },
- "source": [
- "\n",
- "\n",
- "\n",
- "\n",
- "*Figure: The backup diagram. Edited version from the book of Richard S. Sutton and Andrew G. Barto* **[6]**."
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "XflFYx7lylyw"
- },
- "source": [
- "In *q-learning* we aim to derive an *action-value function*, the q-function. The q-function indicates what is the **long-term** value of taking an action $a$ from a certain state $s$. With this information we not only have an estimation of the value of each state, but we can also decide to take the next action $\\pmb{a}'$ that leads to the highest value from the next state $\\pmb{s}'$. This principle relies on the so-called *Bellman optimality equation* that is presented below:\n",
- "\n",
- "\\begin{align}\n",
- " q(\\pmb{s},\\pmb{a}) = r' + \\gamma \\max_{\\pmb{a}'} q(\\pmb{s}',\\pmb{a}')\n",
- "\\end{align}\n",
- "\n",
- "This equation states that the total expected cummulative return of taking action $\\pmb{a}$ from state $\\pmb{s}$ equals the immediate reward $r'$ plus the maximum achievable reward that we can obtain from the following state $\\pmb{s}'$. Note that the q-function estimates the **TOTAL EXPECTED CUMULATIVE RETURN** of taking action $\\pmb{a}$ from state $\\pmb{s}$ (not just the immediate reward). So given the q-function we can know straight-away what is the best action to take for each state $\\pmb{s}$. You can imagine a q-function with one-dimensional state and action spaces as follows:"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "HBpa3qjuysK-"
- },
- "source": [
- "\n",
- "\n",
- "\n",
- "\n",
- "*Figure: Example of how a q-function may look like for the case with one-dimensional state and action spaces. Note that, given the q-function, we can pick the action $a$ that leads to the highest expected cumulative reward $q_*$ from state $s$.*\n",
- "\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "db8AVf3GoCz9"
- },
- "source": [
- "Powerful, right? now the question remains how to derive the q-function 😅.\n",
- "\n",
- "The q-function is inferred iteratively using the reward received by the agent each control step and bootstrapping with the Bellman optimality equation presented above. The sum of the immediate reward and the next-state q-function estimate is called the target. We use this target to recursively update the q-function at a learning rate $\\alpha$. The difference between the target and our current q-function estimate is called *temporal difference*. In summary, the q-learning method consists of recursively updating the q-function using the following formula:\n",
- "\n",
- "\\begin{align}\n",
- " q(\\pmb{s},\\pmb{a}) = q(\\pmb{s},\\pmb{a}) + \\alpha [ \\underbrace{\\underbrace{r' + \\gamma \\max_{\\pmb{a}'} q(\\pmb{s}',\\pmb{a}')}_\\text{target} - q(\\pmb{s},\\pmb{a})}_\\text{temporal difference}]\n",
- "\\end{align}\n",
- "\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "qjBqNfXd_pY2"
- },
- "source": [
- "So in summary, the agent observes the reward once it has taken an action from a state. It has to explore the rewards from different state-action pairs and update its q-function as it goes.\n",
- "\n",
- "In our example we are going to use tabular state and action spaces to expedite learning and to easily store and visualize the q-function. Note, however, that we could use general function approximators like neural networks to configure the q-function. \n",
- "\n",
- "📌 **Note: The exploration-exploitation dilema** ⚖️\n",
- "\n",
- "RL always faces the so-called exploration-exploitation dilema. That is, how much of what we have learned we should exploit and how much we should explore to find even better solutions? In our case, we implement an *Epsilon-greedy* approach to balance exploration and exploitation of the RL agent. That is, the agent sometimes picks a random action (exploration), and sometimes picks an \"intelligent\" action (exploitation). The frequency at which the agent picks a random action is determined by *Epsilon* (`eps`) and it follows a linearly decaying schedule. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "ZQ7Um2UtLHk4"
- },
- "source": [
- "\n",
- "\n",
- "*Figure: The epsilon-greedy strategy for balancing exploration and exploitation.*\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "6aOHW96nLqOZ"
- },
- "source": [
- "Our `Q_Learning_Agent` consists of only three methods: \n",
- "\n",
- "- `__init__` ➡️ The constructor.\n",
- "- `predict` ➡️ Method to decide on an action given an observation. \n",
- "- `learn` ➡️ Method for learning with the q-learning method explained above. \n"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 16,
- "metadata": {
- "id": "9U81QUVcUfoW"
- },
- "outputs": [],
- "source": [
- "class Q_Learning_Agent(object):\n",
- "\n",
- " def __init__(self, env, eps_min=0.01, eps_decay=0.01, alpha=0.05, gamma=0.9):\n",
- " '''Constructor of q-learning agent. Assumes discrete state and action spaces.\n",
- "\n",
- " '''\n",
- " self.env = env\n",
- " self.eps_min = eps_min\n",
- " self.eps_decay = eps_decay\n",
- " self.alpha = alpha\n",
- " self.gamma = gamma\n",
- "\n",
- " # Initialize epsilon \n",
- " self.eps = 1.0\n",
- "\n",
- " # Initialize q-function as a null function\n",
- " self.q = np.zeros((env.observation_space.n,\n",
- " env.action_space.n))\n",
- " \n",
- " def predict(self, obs, deterministic=True):\n",
- " '''Method to select an action with an epsilon-greedy policy. \n",
- "\n",
- " '''\n",
- " if deterministic:\n",
- " # Use q-function to decide action\n",
- " return np.argmax(self.q[obs])\n",
- " else:\n",
- " if self.eps > self.eps_min:\n",
- " # Linearly decreasing schedule\n",
- " self.eps -= self.eps_decay\n",
- " if np.random.random() < self.eps:\n",
- " # Explore with random action\n",
- " return np.random.choice([a for a in range(env.action_space.n)]) \n",
- " else:\n",
- " # Exploit the information of our q-function\n",
- " return np.argmax(self.q[obs])\n",
- "\n",
- " def learn(self, total_episodes=10):\n",
- " '''Learn from a number of interactions with the environment.\n",
- "\n",
- " '''\n",
- " for i in range(total_episodes):\n",
- " # Initialize enviornment\n",
- " done = False\n",
- " obs = env.reset()\n",
- " # Print episode number and starting day from beginning of the year:\n",
- " print('-------------------------------------------------------------------')\n",
- " print('Episode number: {0}, starting day: {1:.1f} ' \\\n",
- " '(from beginning of the year)'.format(i+1, env.unwrapped.start_time/24/3600))\n",
- "\n",
- " while not done:\n",
- " # Get action with epsilon-greedy policy and simulate\n",
- " act = self.predict(obs, deterministic=False)\n",
- " nxt_obs, rew, done, _ = env.step(act)\n",
- " # Compute temporal difference target and error to udpate q-function\n",
- " td_target = rew + self.gamma*np.max(self.q[nxt_obs])\n",
- " td_error = td_target - self.q[obs][act]\n",
- " self.q[obs][act] += self.alpha*td_error\n",
- " # Make our next observation the current observation\n",
- " obs = nxt_obs\n",
- " # Print the q-function after every episode to show progress\n",
- " print('q(s,a) = ')\n",
- " print(self.q)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "RWPbW8WKaQET"
- },
- "source": [
- "##**Testing our RL algorithm in BOPTEST-Gym** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "SV8bk8x75C_0"
- },
- "source": [
- "Now that we have a RL agent ready, let's test it in BOPTEST-Gym! We are going to exploit the features of BOPTEST-Gym to: \n",
- "\n",
- "- Define a custom reward function of the enviornment.\n",
- "- Instantiate the environment and define its state and action spaces. \n",
- "- Train our RL agent.\n",
- "\n",
- ""
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "wy1TpSGEPxYr"
- },
- "source": [
- "### Define a custom reward function of the environment"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "J9jwn5BQQCyj"
- },
- "source": [
- "The definition of the reward function is **KEY**🗝 since it is what drives the learning of an agent. \n",
- "The `BoptestGymEnv` Class allows to override its `compute_reward` method that is called every control step as to freely define any reward function of choice. \n",
- "\n",
- "In our example, the goal is to implement a RL agent to identify the actions that keep comfort inside the building, and we should encode our reward function accordingly. We could implement this function by integrating the temperature deviations out of the comfort range. However, this approach is error-prone. We typically want to directly use signals from the environment to define the reward, preferrably those that are directly related to the function we want to optimize so that we make sure we strive for the ground truth optimum. In BOPTEST we use the `GET /kpis` API to obtain the so-called core KPIs at the present time, which are:\n",
- "\n",
- "\n",
- "* **Thermal discomfort**: reported with units of [$K \\, h/zone$], defines the cumulative deviation of zone temperatures from upper and lower comfort limits that are predefined within the test case FMU for each zone, averaged over all zones. Air temperature is used for air-based systems and operative temperature is used for radiant systems.\n",
- "* **Indoor Air Quality (IAQ) Discomfort**: reported with units of [$ppm \\, h/zone$], defines the extent that the CO$_2$ concentration levels in zones exceed bounds of the acceptable concentration level, which are predefined within the test case FMU for each zone, averaged over all zones.\n",
- "* **Energy Use**: reported with units of [$kWh/m^2$], defines the HVAC energy usage. \n",
- "* **Cost**: reported with units of [USD/$m^2$] or [EUR/$m^2$], defines the operational cost associated with the HVAC energy usage.\n",
- "* **Emissions**: reported with units of [$kg \\, CO_2/m^2$], defines the CO$_2$ emissions from the HVAC energy usage.\n",
- "* **Computational time ratio**: defines the average ratio between the controller computation time and the test simulation control step. The controller computation time is measured as the time between two emulator advances.\n",
- "\n",
- "The time series graph below shows how thermal discomfort and energy use are computed by the BOPTEST `GET /kpis` API call. \n",
- "\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "ZQagxsvtr3ow"
- },
- "source": [
- "\n",
- "\n",
- "*Figure: Integration of thermal discomfort (top) and energy use (bottom). In BOPTEST, the `GET /kpis` API can directly return these values every control step. Note that the integration step is significantly smaller than the control step.*"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "esAUAwHdr8y1"
- },
- "source": [
- "The core KPIs are normally calculated at the end of the simulation to assess the controller performance, although they can be computed at any time. The warmup period is not taken into account for the calculation of the KPIs. See below how we define the `compute_reward` method using the `GET /kpi`. Every control step we check whether there has been a discomfort increment. If there is not discomfort increment, we reward our agent with $1$, otherwise we return a $0$ (no reward). Clipping the reward is a good practice to accelerate learning. "
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 17,
- "metadata": {
- "id": "hTcc3XiVP-A6"
- },
- "outputs": [],
- "source": [
- "# Redefine reward function\n",
- "class BoptestGymEnvCustomReward(BoptestGymEnv):\n",
- " '''Define a custom reward for this building\n",
- " \n",
- " '''\n",
- " def compute_reward(self):\n",
- " '''Custom reward function. To expedite learning, we use a clipped reward \n",
- " function that has a value of 1 when there is no increase in discomfort \n",
- " and 0 otherwise. We use the BOPTEST `GET /kpis` API call to compute the \n",
- " total cummulative discomfort from the beginning of the episode. Note \n",
- " that this is the true value that BOPTEST uses when evaluating \n",
- " controllers. \n",
- " \n",
- " '''\n",
- " # Compute BOPTEST core kpis\n",
- " kpis = requests.get('{0}/kpi/{1}'.format(self.url, self.testid)).json()['payload']\n",
- " # Calculate objective integrand function as the total discomfort\n",
- " objective_integrand = kpis['tdis_tot']\n",
- " # Give reward if there is not immediate increment in discomfort\n",
- " if objective_integrand == self.objective_integrand:\n",
- " reward=1\n",
- " else:\n",
- " reward=0\n",
- " # Record current objective integrand for next evaluation\n",
- " self.objective_integrand = objective_integrand\n",
- " return reward"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "2hpd_svcOhDy"
- },
- "source": [
- "### Instantiate the environment and define its state and action spaces"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "xszlVIQtOkiz"
- },
- "source": [
- "Similarly to our `SimpleController` example, now we are going to use an agent that observes only the current indoor temperature and decides whether to turn heating on or off. However, instead of hard-coding such logic, we are going to use our very own implementation of the `Q_Learning_Agent` to see if it can learn how to do that. \n",
- "For this, we are going to let our RL agent interact with the building for some episodes of experience. \n",
- "Since we are now going to run several episodes for training, we want to stop our previous environment and start one that randomly initializes our building emulator throughout the year. \n",
- "This allows to train our agent when using different boundary condition data in our building environment. We are also going to exclude the Spring, Summer, and Fall periods for training since we are only focused on learning the heating behavior. \n",
- "\n"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 18,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/",
- "height": 35
- },
- "id": "24fsDMTv8tSF",
- "outputId": "7ce7d024-c3ff-49f5-f2cb-af0a46af88ab"
- },
- "outputs": [
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ]
- },
- "metadata": {}
- }
- ],
- "source": [
- "env.stop()\n",
- "import random\n",
- "\n",
- "# Seed for random starting times of episodes\n",
- "seed = 123456\n",
- "random.seed(seed)\n",
- "# Seed for random exploration and epsilon-greedy schedule\n",
- "np.random.seed(seed)\n",
- "\n",
- "# Winter period goes from December 21 (day 355) to March 20 (day 79)\n",
- "excluding_periods = [(79*24*3600, 355*24*3600)]\n",
- "# Temperature setpoints\n",
- "lower_setp = 21 + 273.15\n",
- "upper_setp = 24 + 273.15\n",
- "# Instantite environment\n",
- "env = BoptestGymEnvCustomReward(url = url,\n",
- " testcase = 'bestest_hydronic_heat_pump',\n",
- " actions = ['oveHeaPumY_u'],\n",
- " observations = {'reaTZon_y':(lower_setp,upper_setp)}, \n",
- " random_start_time = True,\n",
- " excluding_periods = excluding_periods,\n",
- " max_episode_length = 2*24*3600,\n",
- " warmup_period = 24*3600,\n",
- " step_period = 3600,\n",
- " render_episodes = True)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "NU8aoMvV9AdE"
- },
- "source": [
- "We have set the zone temperature as the only observation of the environment state. We have also set the lower and upper bounds of this variable to be $21$ and $24 °C$, respectively, which are the bounds of the comfort range during occupied periods. These bounds can be used by the environment for normalization or discretization purposes. In fact, we are going to discretize both the action and observation spaces to expedite learning. We decide to set only one bin for the action space (two possible actions: heating on or off). We split the observation space in three bins with the outer bounds of the comfort range as bins of the observation space (`outs_are_bins=True`). That is, the observation space is defined by $[-∞,21,24,+∞]$ as shown on the left hand side of the figure below. Note that only the middle bin is always comfortable whereas the other bins may lead to discomfort. If we had set `outs_are_bins=False` we would have had all our bins within the comfort range. The latter would give the agent a notion of what is the temperature within the comfort range (close to the lower bound, middle, or close to the upper bound), but it would raise an error if the temperature is out of the range. "
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 19,
- "metadata": {
- "id": "uCUZKrOMOIEN"
- },
- "outputs": [],
- "source": [
- "from boptestGymEnv import DiscretizedObservationWrapper\n",
- "env = DiscretizedActionWrapper(env, n_bins_act=1)\n",
- "env = DiscretizedObservationWrapper(env, n_bins_obs=3, outs_are_bins=True)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "Ab6WP3zLEvnb"
- },
- "source": [
- "\n",
- "\n",
- "*Figure: Possibilities for the discretization of the state space.*\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "GTvGxERwOOI6"
- },
- "source": [
- "### Train our RL agent"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "Sc3XqYSDOuGq"
- },
- "source": [
- "The only missing step is to let our RL agent learn by rolling out episodes of experience with the environment. We use the previously defined `learn` method for this. Note that, since we set `render_episodes=True`, we will be seeing a plot with relevant variables after each episode is finished. This is helpful to check if the agent is learning as expected from early stages. If the agent is not showing any sign of life we can prematurely stop the learning process to use new learning settings while saving some valuable time and computational cost. "
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 20,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/",
- "height": 1000
- },
- "id": "jtOpX5y_RTsV",
- "outputId": "83940660-f744-4d0f-a440-79ec2d5c45f9"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "-------------------------------------------------------------------\n",
- "Episode number: 1, starting day: 11.4 (from beginning of the year)\n"
- ]
- },
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- },
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "q(s,a) = \n",
- "[[0. 0. ]\n",
- " [1.936 1.398]\n",
- " [0.594 0. ]]\n",
- "-------------------------------------------------------------------\n",
- "Episode number: 2, starting day: 67.8 (from beginning of the year)\n"
- ]
- },
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- },
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "q(s,a) = \n",
- "[[0. 0.17 ]\n",
- " [2.414 2.116]\n",
- " [1.491 0.594]]\n",
- "-------------------------------------------------------------------\n",
- "Episode number: 3, starting day: 0.9 (from beginning of the year)\n"
- ]
- },
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- },
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "q(s,a) = \n",
- "[[1.594 1.141]\n",
- " [2.411 2.221]\n",
- " [1.491 0.594]]\n",
- "-------------------------------------------------------------------\n",
- "Episode number: 4, starting day: 29.9 (from beginning of the year)\n"
- ]
- },
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- },
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "q(s,a) = \n",
- "[[2.411 2.219]\n",
- " [2.382 2.409]\n",
- " [1.491 0.594]]\n",
- "-------------------------------------------------------------------\n",
- "Episode number: 5, starting day: 19.8 (from beginning of the year)\n"
- ]
- },
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- },
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "q(s,a) = \n",
- "[[2.411 2.219]\n",
- " [2.833 2.967]\n",
- " [2.287 1.918]]\n",
- "-------------------------------------------------------------------\n",
- "Episode number: 6, starting day: 11.0 (from beginning of the year)\n"
- ]
- },
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- },
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "q(s,a) = \n",
- "[[2.411 2.219]\n",
- " [4.677 4.24 ]\n",
- " [2.367 1.918]]\n",
- "-------------------------------------------------------------------\n",
- "Episode number: 7, starting day: 45.2 (from beginning of the year)\n"
- ]
- },
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- },
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "q(s,a) = \n",
- "[[2.452 3.17 ]\n",
- " [4.847 4.609]\n",
- " [2.432 1.997]]\n",
- "-------------------------------------------------------------------\n",
- "Episode number: 8, starting day: 362.2 (from beginning of the year)\n"
- ]
- },
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- },
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "q(s,a) = \n",
- "[[2.452 3.271]\n",
- " [5.66 5.162]\n",
- " [2.737 2.014]]\n",
- "-------------------------------------------------------------------\n",
- "Episode number: 9, starting day: 72.3 (from beginning of the year)\n"
- ]
- },
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- },
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "q(s,a) = \n",
- "[[2.452 3.917]\n",
- " [6.03 6.114]\n",
- " [2.737 2.014]]\n",
- "-------------------------------------------------------------------\n",
- "Episode number: 10, starting day: 357.8 (from beginning of the year)\n"
- ]
- },
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- },
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "q(s,a) = \n",
- "[[2.56 4.458]\n",
- " [6.448 6.346]\n",
- " [3.256 2.165]]\n"
- ]
- }
- ],
- "source": [
- "model = Q_Learning_Agent(env, eps_min=0.01, eps_decay=0.001, alpha=0.1, gamma=0.9)\n",
- "model.learn(total_episodes=10)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "Qblijq9WVobQ"
- },
- "source": [
- "Since our environment has been defined with one-dimensional state and action spaces, we can plot the q-function after training as follows:\n"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 21,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/",
- "height": 248
- },
- "id": "z_u32nxmzSDm",
- "outputId": "59c90ac3-7659-4b5f-91a4-73c66f7d837e"
- },
- "outputs": [
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "iVBORw0KGgoAAAANSUhEUgAAAV0AAADnCAYAAAC9roUQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOy9eZQcZ3k++nxV1fvs0kgazUiaVastyZZlWbHBEI6DA4F7HEzYcgnYHJyAc+3rQMg9yYWfDbENJBCHnHByEwJxwHA5uSEsCYuDrZAQ25IjW7a8aPZ933rvru27f1R/NVW9VndXb6N6DsPINbV1d/VTb73v8z4voZTCgQMHDhxUB1ytT8CBAwcOriY4pOvAgQMHVYRDug4cOHBQRTik68CBAwdVhEO6Dhw4cFBFCAX+7kgbHDhw4KB4kFx/cCJdBw4cOKgiHNJ14MCBgyrCIV0HDhw4qCIc0nXgwIGDKsIhXQcOHDioIhzSdeDAgYMqwiFdBw4cOKgiHNJ14MCBgyrCIV0HDhw4qCIc0nXgwIGDKsIhXQcOHDioIhzSdeDAgYMqwiFdBw4cOKgiCrmMOXCQE5RSqKqKZDIJWZYhCAI4jgPP8+A4DhzHgZCcZksOHFyVIAUGUzrWjg4yQCmFoiiQZdn0b/Y3I9EyEmY/Dhk7uEqQ8wJ3SNeBZaSTLSEEhBDIsgxZlsFxXMb6xh+HjB1cRch5ITvpBQcFQSmFLMuYnZ1FS0sLAoFABsFmAyPlbPsDAFmWIUmS6W8OGTvY7nBI10FOMLJlqYNgMAiv14umpqay9ssINJ1I08nYGB0TQsDzvJ43ZuTskLGDRoNDug4yoKqqKU/LIlaO46Cqasb6dhFfITJOT21QSvNGxg4hO6hHOKTrQIeqqpBlGYqiAMhMDzCiqzasknH6NhzHQRAEh4wd1BUc0r3KwYpckiTpUWwuYuI4riakmwuFyJjJ2QDgypUrOHTokE7G6akKh4wdVAsO6V6lYBpbWZYLki0DISRreqHekI2Mo9EoeJ7XX7eiKBBF0bSdMU3BomOHjB3YDYd0rzKkky0jFSvEki29oCgKZmZmEI/H0dTUhEAgAL/fb0ndUAvkU1QwMk6Xt2XLGTuKCgelwiHdqwS5NLbFEIcxvSDLMmZmZjA/P4/du3ejubkZsVgMq6uriMViAACv14tAIKD/bDcyduRtDkqBQ7rbHNnItlTiI4RAkiSMjo5iaWkJ3d3duOmmmwAAkiSZ9kspRTweRzQaRTQa1cmYUgqfz7ctyJjlwkVRdMjYgWU4pLtNwTS2xgitHGJLJpNYXl5GNBrFwMAAzp49q++PqR2MIITA7/fD7/ejs7PTdF7xeByxWAyRSASrq6uIx+NQVVUnY7/fj6amJvh8PvA8X/I5VxJO44eDUuGQ7jZDekNDuWQbj8cxMTGBzc1NtLS0YNeuXdi3b59pnWJIw0jGO3fuNJ13IpHQI+P19XXEYjGdjP1+vykyLpaMq6W6sNr4YQQjX6fx4+qAQ7rbBLkaGkpFNBrFxMQEwuEw+vr6cOTIEczPz2cQhl0ghMDn88Hn81km42w543qOjI2/GZzGj6sPDuk2OAo1NBSLcDiM8fFxJBIJ9Pf349ixYybCqLZO1y4yrleScho/rj44LmMNiGIaGqwiGAxibGwMqqqiv78fHR0dGessLi4iFouhv7/ftFxVVYiiWBfFsHQyjkajev64o6OjYSLjXGDfV3YNGOE0ftQVHJex7YBSGhoKYX19HePj4+A4DgMDA2htbc25biM0R+SKjM+fP4+hoaFtm6YAzPI2p/GjfuGQbgOgnIaGXPtbXV3FxMQEPB4PDh06hObm5oLb1cp7wQ5YSVPEYrGsZMzUFPVMxkBxWmP222n8qD4c0q1j2NHQkL6/5eVlTExMoKmpCceOHUMgELC8fSOTbi4YydgIRsaxWAzRaBSzs7OIRqM6GRvVFIFAYNuQMYMjb6scHNKtQ9jZ0ABoOdfFxUVMTk6ira0NJ06cyCAZK8hl7QjYZ+9YLzCS8Y4dO/TllFIkk0k9TTE3N6eTscfjMRFxI5OxsfHjypUrOHjwoEnO5pBx6XBIt45gd0ODqqqYm5vD9PQ0du7ciVOnTsHj8ZS8v+0Y6RYLQgi8Xi+8Xu9VQ8axWAyCsEUVTuNHeXBItw5gd0MDM6GZm5vD7t27cfr0abjd7rLPM5+1Y/rjab2h0jcLO8jY7/ebyK1ekP7Zltr44ZCxhvr7hK8iqKqKeDyOjY0N7Nixo+xqsizLmJ6exsLCAvbu3YszZ87Y+iXOp164mr9E+VAMGcdiMSiK0jBkzFBO44dR2na1KCrq95PcxjA2NMRiMczPz5v8CYqFKIqYmprC8vIyenp6cNNNN1Xk8TVXeqERviT1do7lkLFRTbEdyDh9m+3e+FG/n9g2Q66GBpfLldUwxgoSiQQmJyexvr6O/fv3m0xoKoFs6YVYLIaxsTFsbm7qEiujr2495y7rEfnIWBRFRCIRxGKxDDJOV1M0MhkbJ34Yt9kujR/1+8lsExRqaOB5vmjSZSY0wWAQvb29+hiaSsOYXohGoxgbG0M8Hkd/fz+GhoYgimJOwxpGBsw9rB661xoJhBB4PB54PJ6sZMze94WFBUSjUSiKArfbnVHAK5aMmS68GrCj8YMFNS0tLVU551LgkG6FYLWhIZ8MKx0Xpy/i1YlXsYPswKnDp3DkyJGq3uWZn+6lS5eQTCYxMDCgtwuLopgzQsvmqwtAJ2MWGft8voaLWmoNIxkbW7ftImNVVeviacWK1hgAfvrTn+LSpUt4+OGHq32KluGQrs0otqHBSqQbCoUwPj6OH079EP+68q8QBAFNs03ob+tHf3s/+tv6MdA+gIH2AbR4KnOHD4fDGBkZQSgUwsmTJ9HR0aG/pnzKgFy+uqyIGI1GEQ6Hsbi4iHg8DkKIiQyamprg8XgcMi4SdpGxoih1Qbq5kP7dCgaDeVvZ6wEO6dqEUhsa8pHJ5uYmxsbGAAD9/f044j6Cn278FAAQESN4afklvLT8kmmbHb4dGGgf0Al5oH0AfW198Arekl5XOBzG6OgoZFnG/v37oaqqKZItFRzH6V/sXbt26ctZcTEajSIYDGJ+fh6JREJfn0XFgUAAbrfbIeMiUSwZS5IESZIwMjLSEDnjYDCItra2Wp9GXtTnO9dAsLuhgVKqm9AIgoDBwUH9zu1ZLtzYsBZfw1p8Defnz+vLOMKhq6lLj4j72vsw0D6AAy0HwHPZo5hQKISxsTEoioKBgQG0t7dDFEVMT0+X/NqsgOd5NDc3Z3hByLKsu4Wtra1henoaoihCEISMyNjlclX0HLcjcpFxOBzG5OQkduzYUTAy9vv9NX/vg8EgDhw4UNNzKASHdEuE3Q0NlFKsrKxgYmICPp8PR44cQVNTk2kdj1BaN5lKVcyF5zAXnsN/zvynvtzFubC/dT/62vp0Qt4l7EJkIQIAGBgYMEUN+ZojKg1BENDS0pJRIJEkSY/O2PsnyzJcLpcpKq53d7R6haqqcLvd6OjoyIiMJUlCJBJBNBrF4uIiIpEIFEWBy+Uy3QSrScZOpLsNYfeEBnbxPvvss2hubsa1114Lv9+fdV03X35XmRGSKmFsYwxjG2OQFRnJRBKEELQF2jC4YxD9CXO+uNXdWnfk5XK50NbWlvFFY4/KkUgECwsLiMfjOH/+vKnxoBGcw2qNXDldQkheMmY3wsXFRUSjUf1GmJ4ztpuMQ6GQQ7rbBXZPaFBVFQsLC5iamoKiKDh58mRBExoPX7pvQi7IsoxkMqnrQ3meh0hFvLr6Kl5dfdW0bpunDQExgLPkrE7G/W398Luy3yRqCbfbDbfbjfb2dgDaY/INN9xgajyYmZkx2TgaI+N6nlBcTaiqWtT7wMjY+N4DmWS8tLRUETIOhUJOIa2RUYkJDYqiYG5uDjMzM+js7MSpU6fwwgsvWPJGKDW9kA3ZyLYQNpObmI3MYu61OdPy3YHdONB8AP3t/ehr7UN/ez8OtBywPTIvF/kaD5inLptQ7MjaNNilXshFxgBMBbxyyTgYDGbsv97gkG4WVGJCgyzLmJmZwfz8PPbs2YMbb7xRv4CYbKzQxW1HpFsK2RbCUnQJC+EFPLfwnL6MIxx6mnvQ19qn54z72/rR3dwNjlQ3giyUh85lcG6UtUUiEZOsjbXiMjLwer3bkoyrIRmzk4ydnG6Dgcm+FEWxZUIDoBV6pqensbi4mNOEhud5S7nScki3EmSbDypVMR2axnRoGv8x8x/6cjfvxoGWA+hr08h4oE1LUXT6S/eeqBSMsjYjVFXVlRTZZG3GyLjRZW3FphfsRDFkfOXKFfzN3/wNIpEIvva1r+Gaa67BsWPHss76s4rNzU185CMfweXLl0EIwd///d/j7Nmz5b4sh3QB+yc0ANqFMTk5iZWVFezbty+vCQ3HcZZagUt5XK822RaCqIgY2RjByMaIaXmTqwm9rb1bKYpUZFypZo9ywHEcmpqaMtQliqKY2qCZrI3neVNU3EiyNqZGqCdkI+Prr78et956K+68804AwHe/+13s378fn/zkJ0s+zn333Yfbb78d//RP/wRRFPWUU7m4qknX7gkNgGZCMzExgY2NDRw4cACDg4MF92nVf6GYnK4kSxCT2oRer8+bU49bL4hIEVxevYzLq5dNyzu8HXp64mPXf6yk1ES1ZG48z+eUtbHIeGVlBZOTk5AkKUNaVY9NB/XekWbEnj17wPM8fu/3fq/sfQWDQfziF7/AN77xDQBbRG8H6usTrhLs1tgCmtvW+Pg4wuEw+vr6cPjwYcuRsmXStZBekCQJyWQSPM9X1VimUo/Q64l1rC+u438W/wf3nLwHHF/a66nlI77L5UJra2tGVd34mGzUudbTpAlFURpGxZFIJHLKLYvFxMQEOjs78eEPfxiXLl3CqVOn8NhjjxU1UzAXrirSzUa25X4ZI5EIxsfHEY/H0dfXh2PHjhW9TztI10i21ZY7qaoKSZJAOAKe4ytmt0exvUYF5ZJWMQvH9EkTxlHxTGNc6c+5XgxvrGBzc9M2dzFZlnHx4kV85StfwZkzZ3Dffffh0UcfxWc/+9my931VkK7dDQ2A9vgxPj4OWZbR399vMoApFpYLaVnSC7UkW2a4zSrKiqxAUiX9cZ75n3IcB47nqq5aaETks3A0ytrW1tYyZG3sx84nnEZKL9ipXOjp6UFPTw/OnDkDALjzzjvx6KOP2rLvbU26djc0AMDGxgbGx8cBZLbJlgqrka7ACeAIB5WqdUO2bo8bAU8gY44WpRSKqkBVUp9BcuszYCTMc3xR87Ku5qGYVmRt0WgUy8vLiMfjAKAbm7N8cSmytkYjXbsaI/bs2YN9+/bhypUrOHToEH7+85/j6NGjtux725FuJRoaKKVYW1vD+Pg43G43hoaGbDVJtqpeoJQCChCJRWpGtqqqIhqN6mYnuewdCSEQeAHgzdvr/qeqZkZt/IzSo+KMyQIlphfYuX3mPz+DYDKI/rZ+k364HjvqrKKQrC2XW5sxTZFP1lZLyVix2NzctFWj+5WvfAUf+MAHIIoi+vv78fWvf92W/W4b0q1EQwMzoRkfH0cgEMDRo0czZEJ2oFCka2wZ5mltI1vmd5vLUDrf+80+D47jIBguPUopVKpCVVQ9P2xMUbCoOJFMwMOX5q1LCMFGYgOXli/hhaUXTH/bHditE3F/Wz/62vpwoOUAXHx9SaWKgVHWtnv3bn25Uda2sbGB2dlZ/Ykpm3Vmo0W6dpLuyZMn8fzzz9u2P4aGJ11jQ8OLL76I48eP20K2i4uLmJycREtLC44fP25bVTQbeJ7PGEECaGQ7Pz+P6elp7NixA6dOnUL7XDuWo8sVOxcjKKVIiknIkqxHttFoNOt7m2topRUQQsATPkPWZoyKZVnG66+/Dk7RhhYycrBDarUUXcJSdAnPzD2jL+MIh/0t+01E3NfWh71Nexs6N51L1ibLsp4vNsra4vG4HnTUq6yNoRHMboAGJt1sGttkMglVVUu+KIwk19HRgeuuuw5eb2nm38UgPdJl5zE1NYXOzk7ccMMNukawEqY36WAVdEmSMtII7O/VmsnGZl/BBe3m5/KbLAWN/q4ej8dExulPBMXcFFSqYjI4icngJJ6eflpf7uE9WhNHiohZE8cOX/nG7rWEIAhZZW3PPfccOjs7M2RtbrfbFBXXUtbGEAwG0dXVVdNzsIKGI918DQ0ulwuSJBVNuoqiYHZ2FrOzs9i1a5eJ5KoBpl5QVRWzs7OYmZnBrl27cPr06YzzqKSJjJFsXW5X1jRCLfWuLKfrcrnQ3t6eIbVKJpOIRCIZ1X2v14tEIoF4Im5q7y4FSSWJK+tXcGX9iml5i7vFRMJ9bX3oa+1Dk9v+dFQ1wXFc1vfaaJ1Za1kbQyP4LgANSLqKomia0CwNDYIg6LIwK5BlGdPT05ifn0dXV5fJhKba2NzcxDPPPIPdu3fnPY9KRLomsnVlJ9t6QKFZbMxBLL26Hw6H8dprryEpJpFIJCoiaQuJIVxavoRLy5dMyzv9naZ8cb06sBWDfCN/mKyNtUJHo1FQSuHz+UyRcSUadxzSrRDYvPtsYJFuIbCxM4uLi+jp6cHZs2dr8mjEIuypqSkIgmCJ9O20d2wUsmUoRb3AqvVutxvNzc3wJ7XcfCUkbdmwElvBSmwFz81vObAREOxr2WfKF9fKgc1O5JK1GSdCRyIR22VtDA7p1gCFIt1kMonJyUmsrq5i//79+JVf+ZWayGEURcHMzAzm5ubQ1dWFEydOYGJiwlKUbUekSymFKImQxNLItlo53WzHtQuVkLRZBQXN68CWni+ulQObXe93vonQTNYWCoVMsrZ060wrE6FDoVDde+kCVwnpxuNxTExMYHNzE729vRgaGqoJ2TJP3bm5OXR3d+s2j4lEwpJOFyiPdJl+WRTFkiPbeo6EC6FQpFyupI1FxaW+R7kc2JrdzSZdMfvd7G7OsSd7UGmNbj63NmYQlE3Wlm6dyeBEuhVCvgs6Pb0QjUYxPj6OSCSCvr4+HDlypCakYcwdZ0tnWG0DBgC3UHwu0A6yrQeU2xxRaUmbqqr6MYxRMfOjKBVhMYyXVl7CSysvmZbv9O00EfFA2wB6W3ttyxfXSqObbyI0yxevrq7qsrbFxUX84Ac/gCRJeP7553HNNdeUTb69vb1obm4Gz/MQBMGyXtdwMyap/8646BqOdPNBEAQkk0mEw2GMj48jkUigv78fO3furBnZTk1NYXFxEd3d3Tlzx1bbgIHiIl0j2QouwTayrVU7bjnHJYTYbpiTLmljMEbFsiQjqST19U3FuzLzxavxVazGV3Fh4cLWOYGgu7nbFBH3tfZhX8u+ovPF9eYwlkvWFgwG4Xa78ZnPfAZPPPEEXnnlFdx666146KGHyjre008/bcpN5wN7KjB0aJouNkIIYcsajnTzXaTJZBKzs7PY2NjQTWhqAUmSMDU1haWlJfT09OQ1MAe012Q10rVCuiayFewj21qjXNKs1s2CIxw4wUxWLF9sLN6xz5wRsDEqLidfPBuexWx4Fr+Y+YW+3MW5sL9lP/rb+vHmA2/GzT03F9xXoziMtba24q1vfSv+9E//FH/9139dk3PgOA5zc3N47bXXEIlEcMcdd/w6gFUAU5TSZSMJNxzpApndT+vr6xgbG4Oqqmhubsb1119fk/OSJAmTk5NYXl7Gvn37cPbsWUuRQjFfsEKkK4piRck21/4UVQEBybtOuWhkwxtjvtj4rWMpClVV87u05VHtWIGkShjbHMPY5hi8gtcS6TZSC3AikYDHY5+yhxCCX/u1XwMhBPfccw8++tGPZl1venoa//AP/4Af//jHCIfD+vcPwF8CWAPwOiHklwCeAfA6pVRuSNIFtIt1dXUV4+Pj8Hq9OHz4MDiOw8jISOGNbYZxNM/+/fstk20pyCUZEyURYrI6ka0x4lQUBclkMiNvamdOs1zo51aHfrzGFIVRvZItKjY2AxkJudj31itY67JsJNK102EMAP7rv/4L3d3dWF5exm233YbDhw/jjW98o2mdu+66C9/73vf0gZlpGEj93AjggwDGAPwlIeT/a0jSXVpawtjYGJqbm3HNNdfoDktMc1otiKKIiYkJrK6u4sCBAxUlWwYXZ5aVMYtHQRCq1/1DzWTr8XrAc7xOCipVgfUNkGAQqssFSRCguF2AxwvO5SpZdlUOaZbjDVELZJO0AdCjYlVVc0vaCuSLrRbaGslhzG7lQnd3NwBg165duOOOO3D+/PkM0v3Wt74FSZJw+vRp3H777Th06BA6OjqQSCRwxx133A2gG8AhAEMATgJ4DMA9DUm6qqri5MmT8Pl8puXFdqSVCqb3XVtbw4EDB6oqQWNRSi39dJNJrTDk8Xj0lmsjoXGEAwgBJ4ngJVGrMVGAAqA8D9XlguJyQXIJoC434PGAF4SKm53XY6RbLLKlGdIlbRn54jRJm9VibCNFuqFQyLZIl7U0Nzc3IxqN4mc/+xk+/elPm9ZRFAV/8Ad/gI9+9KNob2+H1+s1pTcopboPJCHkWgBvB/C/AbimIUl37969WQtPHMdVNJpJJpOYmJjA+vp6RfS+hZoOKKWIR+KIRKrvp6uq6pZxudtdOH+W/joIQAAQVQGXVCAkEwAA9nGpLhdUlwty6jf1eEBcLlP0Vu5n20iRbjEoVtK2MLOAK+4rJnOgbH4ljUS6m5ubtpHu0tIS7rjjDgCaAun9738/br/9dtM6PM/j4YcfNi1jBWxKKbxerwfaJS9SSl8G8DIh5M8AfLghSbfa+UHjhN++vj4cOnTI9nNgsrFsFz/z9R0bG0Minqgq2Rq9dFlXUKnDIbOBvY28LIGXJbji2IqKOU4nY9Hlwuvf/xd0HLwOgZ07dVF9PgPuqx25JG0DvQPYvXs3IpGInqozDsRk7229Scbywc70Qn9/Py5dulR4xTQQQvRmDUppMv3vlFIZwN82JOlWC6yTLRgMFj3ht1hka5BgxcKxsTE0NTXh5MmTiCxGwE1X/ouQPpKnyat1DSmKAktP6eW8Tywqpio4MQmI2vV74t++hp2iC+LOnYju3YuNXbsQ3rMH0oEDcPX1IZAii3SbQfYEsR3SC3Yg4Amgra3NRFLs8zbOYNvc3NTM3zc2TFFxOf4IlYLdUyNKwbPPPosnn3wSoijic5/73B8C2Ez9RADMUkpfIoSQhiTdfB84G31TzmMRM24OhULo7++vSiebsUHCqMzw+/0mE3UvX1l/30Jeupbfh0q8XxwBTwh8a2vwra2BydapqkL2eJDo6kK4qwuzu3Yh1tUFOjgI365d8Hg8UBRl26YXikW2nK7RpY0NxJyYmNDbbiORCMLhMBYWFkxjf4xkXE071HSEQiEMDg7W5NiSJOHv/u7v8Cd/8ieIx+NIJBIAkD7F8hkANwPgGpJ084EV00oh3VgsprcN9/f34+jRo1W7o3McB1mWsba2htHRUfh8Plx77bUZEyvsdBkzohEcx3JRJuE4uCQJrulpNE9PYy8ApPKZYns7Il1dkNvaEN47grgrAuLxgEvlinm+tnK2WqDYQlq2GWxs7A+LiqempvRrJ32qRzXywrXwXWBPUC+//DK++MUvYmNjw/haRQDGu9BC6ndjRrr5wPwXihFKx2IxjI2NIRqNor+/H8eOHav6F1GSJLz88stoamoyyeDSYbefbim+DLV6TC/qqISA43l4QyG4NzfhTyTQ+sYlrPmTACF6rlgSBCguF5ClcLddybgYyVguwsw19kcURX2qR7q5eZMh/WO3n24tRvWw92d0dBSTk5O488470d/fjy984QsA8BSAnwD4BLTOtG+xzRqSdPN9GYqRjUWjUYyNjSEej9fMo4F108XjcQwMDOj6wFyw0/y6pIYKgsrndHOAlpcm1n5rNiTg88jZVJcLoiBAdWtyNk7gdclVoxSW8sHq01IpaTq3242Ojo6s5uZsqsfy8jJisZg+5NRYvCu1MGp3c0QxWF7WZhbecsst6OnpYYufoZT+JSHkKIBbAVwEAEppY5JuPgiCULBBIhKJaEqARAIDAwPYsWNH1cl2Y2MDo6OjcLlcOHz4MBYWFizlxOxILxgbKopNIxDUMKdbJnLeK9LkbEiTs9GUrlgWhC05m9ttm9F5tWG1LmCXZMxobp7up5ttMjG7Lo1+uoW8poPBYM28dJkhe3NzM9bW1tjiAUJIEwAJWoPEOwF8hRDCNyTpFrJ3zBXphsNhjI2NQRRFDAwMoKOjo+pfls3NTYyOjoLneRw+fFi3r1teXrZkelNOekGSJSQTtWmosANlG94U+VGzS4PIEjhZ0lVXlAJIydkUY1Ts9Wq5YhuMayoJO9ILdoDjuKwWjpIk6WS8tLSkt9kaJW2MjNk1XEvSZTeESCRi/LzfAuCfALDqXpT9oSFJNx+ypRcY2UqSpJNttREMBjE6OgpCCA4ePJiRC2Oqi0IohXRlWUYymdQd+cslW0sqgEqkF8re3q5JCACoCl5Mgk/J2SgFlAMHoHBchnFNPflQAMV5L9TixuxyuQpK2tj8tdnZWTz++OOIxWL42c9+huPHj6O3t7fs81YUBTfccAO6u7vxox/9KOs67BjHjx/HzTffjGg0ittvvx3XXXcdLl682A2tDRgAXgbwQurf2y+94HK59HA/FArpwu+BgYGa3AlDoRBGR0dBKcXg4GDOvJNVT91icrqyIiOZSOpyIFuiFqs53QqgfNKtHAhJtdy6XKZHYdPECVFCUt3y1rVr/E+xsHoN1VNHWjZJGwCcPHkSvb29+PjHP46LFy/i8ccfx4kTJ/CZz3ymrOM99thjOHLkCEKhUN5zAoA3vOEN+O53vwsA6Orqwhe/+EW85S1v+VsAXdDSC39FKX0B0Hx2G5J0CxXSotEoLl68CEop+vv7a0K24XAYo6OjUFUVAwMDBSurPM9bMuuxEqUwMxoA9pFtCgTZmwxUqoKqdCu/WWeFtKogy2vO5a2bb/yPXROKc8Hq01KtZuEVA5fLhZMnT4Ln+bJNyxlmZ2fxr//6r/jjP/5jfOlLX7J0Dl1dXfp//+qv/ioA/B8AdlBK59PXb0jSBbK7Rm1ubmJsbAzJZNB+8vkAACAASURBVBInT56sSTWTkW2x0TXP80xUnRdu3p2T+NKdv4QKpeyNx2aVaUVRTJ+JIIrwqFtf2nr47lactC3XGAt4JShK1gnFdsnZrKYX2LHrHUwfbBfuv/9+fOELX0A4HC5p+5RhfRJABuECDUy6RmxsbGBsbAwcx6G3txdLS0tVJ9xIJILR0VFIkoTBwcGio+ti5qS5eBdERdT/m5nRqKpqcv6qGGgqxyYmIUtam7DX691qt01NSQBS+V9KoSLFSYSk/S7qsOWedoVRHkGZvBJSME0oVpSCDmKFSJKAQNhmPVHBYDCjRlIqfvSjH2HXrl04deoUzp07V9I+tgKNrRE9RjTsu08Iwfr6OkZHRyEIgl6cSiQSmJubq9p5MPlZMpnE4OBgyUU6q4U0QHs8FBVRJ1tFUeDxeuAS7Lvb5wPrRjK2CRuvLUIICMchNZsPANG1sKCpOFlVTUTMtjP8ZwZKLoTpJuYVRgWiQtOEYsPNNJeDWCGT82Ki3EaBnd1ov/zlL/GDH/wA//Zv/4ZEIoFQKITf/u3fxje/+c2i98UIN518G5Z0X331VSQSCZPsCsicCFwpGBsrGNmW8yhWzHBKF+dCIpHQZTTpvsKVgiiKSCaTuqg97+vNYe2oR7hgA/yQ+k0LRsVlpQe2meFNzqGYOUzOGRH7iR/xeLygaU0j+VTY2RjxyCOP4JFHHgEAnDt3Dn/2Z39WEuEakR7tNizpHj58OOtyjuMsP6aXAtYyHIvFbG2ssEK6bAZbLBwDx3Foamoq+7hWYGym8Hq9kGXZtlyfroUlhaPiWCyGRIKC5zjdO6Guco51cC45Tc5TRMxRbaRVIpEAz/Om1lyjr24jFNEY6sFhLB2EEDeAzwJYAnAZwFMpa8fGJd1cJFWpC8VohjMwMGB7y3A+0pVlGdPT01hYWMD+/fuxZ+ceJIKFi27lIpu+1/JkDhusHdOjYq/XC0EWUvIrEUpaFKeTcQ71RL0U0qoNY1Tc0dKB48ePA9A+X6Z7Nfrqer1e+Hw+PY3k9/vrmoAr1Rjxpje9CW9605tK3XwXgE+m/q0CeIAQ8g+U0mDDkm61YLR5HBgYqJgZTrZCmqqqmJ6extzcHLq7u/VR7t4XK5uXUxQFiURC00b6vKYqe02/fGl5TSCVllBVKKlCk5iSX+kVfxYN02okF+qXmBiMGl1BENDa2mp6NGdqlM3NTaytrWFiYgKxWKzurByNCIVCthXSbAQPYBJACMA+AH8B4CyA9zUs6eb78hNCyh6ql0gkMD4+jmAwiIGBgYrbPBojXVVVMTc3h+npaezZswdnzpwxkU2l7B0VVUEykWTjRsrT91ZJp0sIAeF5cDwPGKfppohYTVX8ZUWBrChQVapvZ/tp1nE0yFBIo8t8EiiluuMdkNvK0e12Z1g5VruLLRgMYv/+/VU9phGsHpH2ulcBfAiaveM+AO8G4AMaOL2QD6wVuJQ7sZFs+/r6qmJgDmz56c7Pz2NychKdnZ248cYbs+oP7bZ3NKogvF5vQclZ7dqArceqhOMgcBwgCFAVBRBFcDwHQlQt6s1TtCv59BuBdEt0GMtm5cg8mJmV48zMDKJRzWLA7/eb3MPYqKdKYHNzU0+Z1ALZpHqU0iiAXxgWfZ2kVmpY0rVi71gM6SaTSYyPj2NjY6Nq0yIY2Ay0SCSCUCiEG264Ie+522XvmD7/zJIKooa8UnZ6gBCNF/MoKHRChklvyTbPt/Nyz64qsHMSMCEEHo8HHo/H1JqrqipisRgikQiCwSDm5ub0QiwjYTsNzmvhpcuwtLSEb33rW7q0rKenB48//jjuueeev4dWRPtnSukFQghPKVWABibdfChGNmac8FvpOWjpMM5Aa25uht/vz6nKMKLc9EL6SJ5iVBCNbe2YnbazKSj0YN6ClE3zhazYadsKq7aO5aTnmLIm/bqSJEmPiufn5xGJRKCqKnw+X4bBeTHfwVpMjWDvz6VLl/DQQw/h+PHjuPfee/HDH/4Qv/u7vwtoqQUAOE0I+QCldIltuy1J14qRuSiKmJiYwNraGnp7eysy4TcfWGOHz+fTZ6D993//t6VtS00vlDIlop5QqvqApv22gi3+NUTF6VI2SqGyqJjjIIpi3Y8AqqXZjcvlQnt7u0lpQClFPB7XDc4XFxcRj8dNxM3SFLlafWth68iehl5//XWEQiFcf/31aGlpwT//8z+zVX4BYADAGwD8KoBvE0K4hjYxL9VTVxRFTE5OYmVlBb29vRgaGqpq4j8YDGJkZASCIODo0aMlaW1LId2SpkTkQMNaO5Z7ShlSNm0h89cFtmR22mq1cxLLhUpOjSgFhBD4/f6MWYBGOdvKygomJib0NBgjYr/fD4/Hg1AoVDPSnZ6eBgBcf/31CIfDeOmll3DDDTfgwoUL9wP4PwH87wDYm97YOd18yDY9gjUWLC8v48CBAzh79mxVyTYcDmNkZASU0qx+usWgmJyu0bjcjsi2lqRRr94LhACUEFMenlW0FUXJcBKrpb9uMTndWprc55KzJZNJPSp+5ZVX8Id/+IcIBoP42Mc+hhMnTuDWW2/FDTfcUNIxE4kE3vjGN+p1jjvvvBMPPvhg1nXZZ8ZMqmKxGH76059ieHgY99xzD86fP/8iIYRFfibnnIYl3UKFNEa6kiRhamoKS0tL2L9/f9XJNhqNYnR0FKIoYmhoKG/uyarUzUr/vN3G5YWgqAoUWakokVBSx62paa/X6JlgBKUUiqpAVXL767L0RCXeQ6ukq6pq5Y2TioTRU3fnzp3o7e3F888/j1tuuQUPPPAAXn75ZczNzZVMuh6PB0899RSampogSRJuueUW/Pqv/zpuuummrOcCAG1tbeA4Do8++ii8Xi8URcGZM2fYakzHxvK5FGhg0gWy2zsCWnqBGdEsLi5i3759VSfbeDyuTxgeHBw0VXdzgTVIFDrPfJGuqbHBZi9dBuN7rqoqEokEKKVaLl2SkVSSgKoiwDSxQEmuYnajopRdhK2jwAuadD6FfP66RiK2o+25mKkRxUzUriU4jsPJkydx8uTJsvZDCNHTfZIkQZKknO83+47edddd+Pa3v42JiQkAwHXXXYfbb78dhJCz0PK5a9gav974pJsNsixjeXkZS0tLGBwcrDrZJpNJjI2NIRgMYnBwsKh2YdYgUSjCyEa6RvKrpJcug1Fuxsjd2K9PVdXcCZbmKlaKJrb8nG5FZ0eUviUh4INh8IQAHg/g9Woz2AxEnM28ppRZbI04NSIfZFm29TwVRcGpU6cwOjqKj3/848aoNSv6+/vxk5/8BL/85S8hyzJuvvlmNnwzAOD/hTaqZwHYMr7ZNqQryzJmZmYwNzeHzs5O7Ny5EwcOHKja8Y1qiFJ1vqXMSauFvSOlVLN29LjR5G0yLWfQ7Q9s1MSWQ5mkzO0tHaCczUMhkHhM3xl1u8F7POA8ni0iFgSTeU3WWWwFJhQXk15ohMGldrcA8zyPF198EZubm7jjjjtw+fJlvSuPId0MaGhoCENDQ+nr/DuAfwc0a0fj3xqadAkhuhkM8yc4e/YsRFHEa6+9VpVzkGUZExMTWF5eRm9vLw4ePFjyI6BVe0ev4C2tsaFMiKIIUdTM0wsX5bL/LZ8mNsPekRjo2gZrxspmhMtkXVOajIKISUBMghhKMJQXQFJETD0euDwewOcDUrUA5q+bLSpm6Qk7myPqAZXS6La1teHNb34zfvKTn2SQ7le/+lXcfffdevqFTU1hnsYp72OOUqoCWoRLCPFTSmNAg5Pu3NwcxsbGdDMY9lieTzJmFxRFwdTUFBYWFmzLGVshXUVRsLGykWEiXkkYrR0DgQCi0WjhYxZxTtmIOJu9YyyRQCwG3U2M5ziQerF3LPccLMjwiCIDMRmIRc2SNY9Hj4oFFhWn0j2MjJnR+cToBF4Mv2jqDMtWaG0k0rXLS3dlZUWfRByPx/Hkk0/iU5/6lGkdSinuvfdePPzww7j77rvx3ve+F0eOHAEA03tIKVUJIT4ABwDsAfBOAA8ADU66LS0tJrJl4Hm+YqSrqipmZmYwOztrcv6yA/lIV1VVzM/PY2pqCgJfvtbWCtg04dIVEGU81GdoYjUjFl/SoxvZJA2ja4xEzOVoTqhj7YMl0s2xIUgyASQTplibCgKIx7sVFXs9gMeHE8dO4EjnEV12tba2hlhMS2sYvRJkWW6I9IKdpLuwsIDf+Z3f0WV+v/Vbv4Xf+I3fMK2TTCZx/PhxvPTSS/j85z+Pv/3bv8WePXtw/Phx7Nu3Dy0tLSCE4JOf/OT/Dc3e8QwAJqfYHqRbLU9dI+llc/6yA9nsHSmlWFpawvj4OHbu3Ikbb7wR3BIH8nrlCNc2tzGbE6kU1GRkoy83zBCTZBlqMgmtXyE1xDF1GhUtpFUh0i0GRJYBOQJEIzoZU8Kh5cKX0dxzBoGBAXQODkIdGAACAaiqanIRCwaDuHTpkqlFl0XFdfFkkYKd6YXjx4/jhRdeyLuO1+vFCy+8gK9+9at49NFHMTs7i8XFRbz44ovpqzKBLwXwIoCnAC2/29CkW40Pn1KKhYUFTE5O6qRn5+RRI9ILaWtraxgZGUFLSwuuv/56eL2a3KdSc66KdRsrBJqaW2wXcu0p20BH6DPEtEdrRVGgKCpUotrjKJZ5EmXuoApuv1SFd3wKwosr5iPv2QN1cBCewUG0DQxAHRpCyO/HiZMnIcuyHhWvrKwgHo+DjWsyknGlvhOFUAvfBUIIPvaxj+H9738/nn76afz4xz/GxYsXsby8DFVV0draildfffVJaKqFiwD+hVI6zWalNTTpFkI5I0copVheXsb4+Dja29tx6tSpiusWWXohGAxieHgYbrdb92Uwwm5rR+NkX1uLcoTYG+kW81ESok2RSBWQJFEE4TmQ1BNzeY5iFUCVZpJ51MyUAVlcBL+4CP6//ktfdkJREDhxAnRoCE0DA1AHB6EODgKpBoBsLbper9dExMUa15SCzc1NdHd3V/QYudDW1oY77rgDt99+OzY2NrC5uYlYLAae53H99dd/iFK6YFx/20nG0mFV85oOSinW1tYwOjqK5uZmXHfddXqEWWnIsoy5uTl4vV4cOnQopxTGTmtH3QDHXf8GOHbQUlHqidQGlpo7yk4vlLe5VXhUi5rxRAL85cvA5ctbCwmB2tOjRcUDA2hPETE9eFCfOMGi4qWlJcTjcdMcNuaZYGdaLhQK4dixY7btrxgsLS3h2WefRWdnJwKBAAKBAFpaWljUrxJCuqCN6kkAkACIlFK5oUm3kOmNJElFfcDM+cvr9WaNMCsF1r22sbGBjo6OgheRHaSbrkioCNnWC4FTqknOcpyOVfVE3uaOOsvp5kK2SNcyKAU3MwNuZgZ4+umtxS0tUAcG4B4c1KNievgwkFIRRaNRhMNhLC4uIhKJQFGUjFxxoenEuRAKhWwrpFkFU3acO3cO73vf+9DT06O79vn9fvak+PfQPBdCqd8cgP8A8C8NTbr5YMXekcEO569SIIoixsfHsb6+rnevMef9fCgnp2u3J0Nk7DiIIMG7ewrEHSlrX4VQVcObLOoJIHdzh6ooUMqxdqwS6XoV+2+EJBQC/8IL4I1FKJ6Hun8/1MFBeFlUPDAAHDpksnMMh8NYWFhAIpEoyeS8FjldhtnZWQCadDWLHcGvZ9mkDY1OulYi3XwIh8MYHR2FqqplO38VA1mWMTU1hcXFRfT19elevisrK0V3pFmFoii65aCdngybL74JkSunAQBC6wq8u6fg2T0F7+4peONBuD0LBfZgHeVaM1YyPUE4DgpyWDtaaNcl1SLdciLdYqAo4CYmwE1MAE8+qS+m7e1QBwfhHhhAM0tPpEbtMJPzSCSCubk5RKNRUEozomLj6J9aeOmy78573vMenD9/Ht/73vfwhje8AWfOnEEkEsG5c+fw8ssvA1oRbQ1AO4BBAK8B2zinmy/SNTp/DQ4OVu1DM2p8e3p6MhoqrHakFZNeYIoEVVXh8Xhsl7kRfut8pWAnpGAnwsOaLJEkPw7evQFv+wi87cPwdgzD0zYCd8sMSAnyrbI70iqU7SAEAMdlWDsy7wRju64+pZjZO6YkbdVK6nqsRLoVvAGQjQ3wFy6Av3BBWyAIiP385wBym5xnG/2TTCbxrW99C+vr65ienkZXV1dZhe6ZmRl88IMfxNLSEggh+OhHP4r77rsv67qsRfrKlSs4d+4cPvzhD+Ohhx5CIBCALMu4fPky3vCGN7wEYBjA/wVABNAEbVjl1UW6pTh/2QEmO5uYmMir8bVKulaMqNPbhO2W9DACIZyiaWAphdb1mJpDlhrsoyTbEF08jejiaX1bTojD0zamEXH7CDwdw/C0joPj8z+Z1HVzQxZrR57w4Dne9N7r7bqpKcWqqg3KDFRySnEKHCXgLbQrU1TPDI4WqJsweVogEMDu3bv15YyEP/e5z+FrX/saPvGJT+Atb3kLHnnkkZLOQxAE/Pmf/7luRn7q1CncdtttOHr0aMa6jHS///3vY2VlBadPnzad2y233AIArwB4H4CvUkr/03Ssks6wTmA1vWB0/hoYGEBnZ2fVNL4rKysYGxtDe3s7Tp8+nXfgpFXS5QgHF+eCpGaSVPr8s0oUyYz7U6nW52/0jtUCJZqKTA0mOKn/V2Uf4qvXIL5q6GknCjytkxlRMe8unOO2inIj5byw+B6z6NZ406Wp3v2y1BMWYCnKBfTCY1VQYrG6qakJb3/72/H5z38e3/jGN3LavFpFV1cXurq6AADNzc04cuQI5ubmspIuu/4XFrTU2Ze//GVIkqQ7kp0/fx4ArkutHkhts30GU+Z6swVBQCwWw5UrV8py/ioV6+vrGBkZQSAQwMmTJy1pX7N1pOWCm3dnkC4byVON+Wcsd0lJersoNUS65uOnkzADAQEoj+TmAJKbAwhO3J76A4UrsKBHxC8Gg+iTFtHp2izpnOs1Umbih7LUExZgNZ9b1Ug3ECh7H+ka63IxOTmJF154IaetI7veb731Vvzwhz/Ea6+9hs9+9rPYvXs3XC4XhoeHAeAwgE0Ac6nN9C92w5NuNjBP3dXVVRw6dKgs569iEQqFMDIyAo7jcOzYsaKUEFYjXUAj3aikRYGyLCORSNg2kicftM4uBZIkwef3IexOP1buY2ebJEz1iDjLupRAiuyFFNmL8Myb8Fjcja/JHNqFIA55pnDQO4UhzxQOeyexz7VYkHwqldPVTriMnWeL0IpUT1hp7rCqXKBVjHTLIV32hGUnIpEI3vWud+Ev/uIvchbW2THvvfderK6u4qGHHsLi4iIWFxeNqy1AawUeBrYaI4BtQLrGSFdRFExPT2N+fh47d+5EZ2dn1bpVYrEYRkZGIIoiDh48WJJ20KqfLqDJxoxTIio9ksdYkOM4TtdVEpInMrfgvZB9pHt2IqZUW74ht+JZ+TiejR7X/+bjEhjyTOOgdwqHvRM46JlCv2cOLiKXNA24qijisbic5g6rjREoo5OzaJShhQ+Hw2hubrbtVCRJwrve9S584AMfwG/+5m9a2uZTn/oU3v72t+OZZ57Rg6329nY8+OCDH6KUPpltm4YnXcCsCti7dy9uuukmJJNJFuZXFIlEAuPj4wiFQrrWtlRYjXTj8TjikTgSiUTFRvIwGHPETP0Qj8e1biOBB4WYe2OOB6ACVC1aJJvdI4yCqhQqqJ7fJKmQMK568VL8IF6KH9S34ImCfs8sBl0TGBDGEFt+Bp6OK+Bc8WJOxuIp2xzplnDoQukJXlQRi8Uy1RNp517Nm1M5ke7m5qZtGl1KKe6++24cOXIEDzzwgOXtfD4fTp8+jVOnTiESicDtdsPr9eJ//a//9STzWkjfpuFJd2lpCcPDwxmqgGwTge2EJEmYmJjA6uqqbfniQtuzZoqNjQ20BFqwidJym1bBTMtZjpjB5/PpwxVVyKY8NDEQITUqNCgFSVXqoZZIxIRo/rnaDlP/o4CKrEQsUx4jiQMYTuwHpW/E5M8/ABAKd9McPG0j8HaMwNt+Bd72EQi+9bLeq1qSblZkSU8EOBd8Xq9ujSnmsMYsx7OkaJQR6drZGPHLX/4S//iP/4hrr71Wn7X28MMP421ve5ul7TmOM6UjchEusA1I1+fzZXX+KqYjrRgYUxgHDhzATTfdVPlJuwbD9N7eXhw6dAhf/7evA/YV9k1gOWLWIpwNPJeSQ3mME28paIoIU575ABgfEYDnYMr5Uqo1BZjIuBABGf6uEyzZImKwfCe28p2mzQnEcA/EcA/CM2/WFwveNXg6RrbUE+3DcDXNVSe1WcXGCCvWmIosa8sUxeRRzGnVPlvPqZBkLB/s9NK95ZZbylI/pCMX4QLbgHRbW1uzPpJzqTu2XVBVFXNzc5ientZTGJV21qeUYm5uDlNTUxnHrIS9ozFH7PP7wJHCNxPCGd97g0bX+N1M5WK14NZAhIQAHAE442VIAZWCUBVQqRYRq0bZGTGuad6OraOZIaTOQbsRsHPIOP/U/uTEDsjzOxCd3xq3zbmi8LaPwpNST3jbh+FpnUh7zfpBs7w7FlEl0nXnKKSlW2NKoggKQOB5KCkyFqU8zR3lvPYy0gu1bAEuBw1PupV+DKKUYnFxERMTE+js7Kyon67xmEzf29HRkfWYdto7pnetFXMzyUpAGSsR/ZFf/7RSuUaq0gwiJoSAcuYx5VBVUF87qEi0KQnJJIhqJKtsZJwlKjYhFxFr/69KAcSWTyC2fMLweiV4WifgaR+Bt0NLTXjbRsuTWNVZCzCTjDFrTBiuPZryKM6Xnsg1uSPrscqIdGthdmMHGp50KwVm8TgyMoLW1taq+OkCWrR54cIF+Hy+vPpeO5zG0n10S2oRtkK62ZAu+tdOyNDdlknEalMTqFuT4FEASIraqJpEAkgktX8r6takBGylGgiXjQRyidj0IxjW1P6fqi4kNg4isXEQwfG3p/6owtWxgr2/9f/Au2eq+Peiag5jRTRH5EiZ5UtPqOmTO1KexvnSE+UW0lLjzhsKDU+6+e6oTIJVbBpgc3MTIyMj8Hg8OHHiRFUsHqPRKEZGRvQZTIXu4OWSrl4kc7ty5m2tgHDWmjms7SwLEbM8MaVIJhLgJJJ6rOXB8xz45maQVAGDAoAkAYkElGgUNB6HIMvgFLWIml1uEVvuxg4OUrAbarLE66TOSJdC8yG0CmN6Qo+JDZM78qUn1DIM85liqNHQ8KSbD2wqcCHSXV4G5uYIurvDmJsbgaqqOHz4sK0awFxgLcqhUAhDQ0OQJMkSyZea07VSJCsGhKvs1GU9N0sIvF4vAu6AwbtAGzeufZk58LxGFRLh4GrvgGePdmNSZVmLhqNTWzniIomuUGMHpRRJKQqhHHvHCsOrWKRSO5ojCLGUnnh9Zgbh559HIBBAc3Oz7iRm5amrFg5jdmBbky6TjRVKC4yPi/i93+NAqRv9/Sdx7bUCDh+mOHyY4tAhWo6qJSdkWcbk5CSWl5fR19enS86mpqagKErBvHGxOd30RgrbCMHOSNfqIVMFHONbJMsKkslE6u8EsixBUWQ9IuZ8PtCkuWBH1GKVE2aYiJgQuD186lxy2ztmVbrUYaRbqdtFenriyKlTkA4f1sf/LC0tYWxszGR0zsjYaOkIOIW0miEfeRSSjYmiiLGxMUxNJSEI18HlErC4CCwublmAEgL09GjkayTiUoNEVVUxOzuLmZkZ9PT0ZEjOrHalWU0vmIpkXg94zl7Fha3phRLA3NQURc1oFKGU6gMpJSmpF31yKycAqKpBxmZWThQC4VUILsFExgXtHXkOfJ1NjahqG7DfD57n0dLSYtK5Go3OjZaOTDN+7tw5bGxs2DJw4K677sKPfvQj7Nq1C5eN44kqhIYn3XzIZWTOosylpSX09fXh5pu78Nhj2d8KSoGZGYKZGYJ//3dtGSPigwfNRJzv8zeOUu/s7Czb3tEt5CddW4pkFlD59MIW0qlJFCVIkqh3AaWDEAJBECAIAIULXFyfSllAOcEBpsnCqhYVG2VsaSejVfspthTDW54IPM+bXcUo1ZtLJEkCF4/Dq1RoSrEBxUyNqFpiJMeXhj2R+f1+7Nq1S18uiiLW19exuLiIsbExvOc974HH48G73/1ufOITnyjpFD70oQ/h3nvvxQc/+MGSti8WDU+6xUS6qqpienoac3Nz2Ldvn24ivllkY5eRiFP+yyAE6O7OjIibmrYcx5qamkyj1LPBKul6+dz7sKtIZgl8ieqFMsBSCYIgwO8PFE9OxSonQEB5DiYNW3qHXcpbWDecMZcCTb9BoBee3HCDiCKIbotZvJGNVdRrpFsM3G439uzZg09/+tP4+c9/jvPnz0OWZaytrZV8Dm984xsxOTlZ8vbFouFJFzCb3hjBSJdSivn5eUxOTmY1EbeDlygFZmcJZme3iFhVVbS0hNDbq+LMmetw4oQHkkSRb7iwVXvHbEbmkiwhmUjaViSzAks6XZtAqYp4PA5KtU7EojoBCz3BF1BOmIkYKSI2p2pypVqyErHxejW1UZtJlq1rMrIpkYjd28zwRlVV/SmiVmPYS8G2IN1cEAQBq6urmJmZwY4dO3KaiLtcgNsNiHm8W4qBMY+qqs24fLlNn2RNCLB371ZEfOiQ9sPSWZbTC9zW66hYkcwCqpXTpVR7T/3+AASh+Lx0aVnTLeVEBhGnzHf0NQkgq0kIKRe2gns2fEapQ+Ssp+nrpkfmRRqee4sopFUD1OcrOaK2s9u02tgWpJst0l1bW8P4+DgIITh16lTeR3pAi3bLJV0rI3Io1eRpc3METz21tXzvXo2Ed+xox+HDwM03A/nmZHoET8WLZJZQ4UiXRXmEEHg93pII115ka3Wmqe5lCYlEIqNYxjSpOcFxKaLMQiQ0BwnmSZHkMjy3PDkCle/0BFDWI2Z6+qWRsC1I1wjjOPWBgQFsbGwUJFwAaGqi2Ngo9a6LlP2hmBqR01T0jEINmAAAIABJREFUDXx+nmB+niCZ7AAhgNvtQlfXVjTM8sQtLVohcHVxFbFYrKJFMiuoWHqBUqhUcw7TCau8XlsbTioXtP5mj1eAK/W4zJQTxhsjAJN8jCktaFsbaGsrIIogiVSHXTLVYaeqWV921kAva4rEkCeOJRGNbjUmsN+1Iq5yWoAjkUhVdPSVwLYhXdbRJcsyhoaG0NraikgkguXlZUvbl/r5i6IEUUzaNiLH+Ji5sECwsEDw9NNbf29ri6OzcxVtA8cA+QzInglAqJDdmAVUIr1AVa2DjOM04xp9eZ0/URonIzPlhBFGNy9RFHUi1iNiQQDX0gKS6kakwBYRJ5NAIqH9W1Wy3tQLpSdaXX74/T6oigJFVSGJIpS0c6i0Y54JZTqM5ZrsUCze97734dy5c1hdXUVPTw8efPBB3H333bbsOxu2BelOTk5iYWEhY8JvMfaOxcr9tM6upO0jcrRUSSaRSZImuE8keGxs9CD2SgdmQn8EAHC1rsC7ZxLePRPaT9cEeF/ElvMpCBsjXWMqgWvAx8ZCN6B0Ny8GRVGgqNoIJJbPN5JgtlZnkiJhJDTzH8hSQSL2KNoC1lyC9CnFhgnF0Wg00zvBZkKuF1vHb3/727bsxyq2Ben29PRg3759GcSXS6ebDYGAtT4crWiVTBWtiqygWwIxfVG2imQc/H5f1kq4FOyEFOxE+MrWmHNX62qKhDUy9uwZh+C3n4htSS+wVAJBdSMtu0FKey90IjaUAIxj2rdanbdImwv4wTUZzH8UJRURa+Y/SCRAJAmEbF1MPipo0jRq8FZLXWwcIeAEAQK0a87v9+stu+neCSZHsVQ+uiSU0dhg59SIamNbkK7b7c4a0XIcZ3m6bqGbrqqqSCSSoDSz88lOMA9YYy7Q6/VmkFEhr1spuBNScKeZiFvW9EjYs1uLioVAuLzzLYt0KVSV5W3NqYRc65d+pMrDzlTLVquzORpVVEVPTxiJmOM48D4fOENxiqqqlhtORcTuA/3A6CIgy1vvNCPM1I1PTCa1ay21b4HnMx3FUukJWZK28u5pEbGVJ79yI12HdOsQxTzy57rpbikSlJQiofJvmSzLKQWEN5UXzOJuVUJ0IYV2QArtQHj4Bn1Z86EL6P7Nx0o/Wb60jjQ9b5vF7q9RQSrcKKKnBUxTkFir81ZqANiKnjmPG3zKyYvc91UkuQDI+Di4kRGQ4WFww8MgY2OQw2EkRa0QbCR6XRVk0Cmz/LNhpS0TG0OuupDHbjm2jg7pbgNo6YUtGAcyut0eNDXZP6nBfDxAkkQkkyIIIQgEWBSQPUazK4esRMsrRuSdBpwFjZ63zYsS0wtlHVJvdc7eZiwmt0hwYmQC0fZdaO7qQtPQEPh3vAPxeByvv/oqmtbWMCDL4MbHoY6MgBseBsJhk0bYuP/UP/RlPMeB5zi4DOsZRwDpHrup9Tieh5xFM28VDunWGPkIiBCijw3PB+NNN30gY6UlNVqRTGtr9fl8SCa3tJ65QIpyPM0NqpR3CVh+pM4mAasqKptgIISCcPUhr9DSAmmTNwD09vQiGolidnYWkUgkZRSkYM+ePdjZ3w+1uRmygbzJwoIWDY+MgFy5Am5kBFhdzU/E2n8A2MoTm9ZL5YllWcbsxgbmzp+H1+s1uYl5vd6C37lQKITe3t7S3qAaY1uQbj4wBUO2TjQjAgGqKwQEwV5FQi6Yi2T+lHIB4DgesVgcADVUsTWLQnZOAsejr7UXCSWBhKz9JJUk1CJ1VeWSbuH0Ak0Zy1jN21YIlebDIiP+asPFudDR3oGO9g6Ew2G89tpr6OrqQmdnJ6LRKJaWljA6OqoX0Zqbm9Hc3IyWs2fhvvXWrR2tr2spCUbGw8Mg8/PmT9WQJ04v2LE8sSAI2Hf4MPacOgVRFBGJRBAOh7GwsIBEIgGXy6UTcXNzc0anpRPp1hhWTG/yke7m5ibm5pYhy/0VUiSYkV6UMx6PEMDr9ZjW1fJ1EpJJFUYiFngBLe5WtHq2dJ1JOYmkkkCcEbGczDoHjIGqZUa6eciGUlW7iaTcu8pFfcSR2VFND4pS4BE8UBQFY2NjCAaDOHr0qG6LaCQvSilisRhCoRDW19cxNTUFURTh9XrR0tKikeCJE/CeObP1mUajen6YGxkBGRkBmZzUFBVsx1miYtXng6qqEAQBbW1taG9vB0ldK7Is60S8urqKeDwOQgiam5vx9NNPY3l5Oecoq2Lxk5/8BPfddx8URcFHPvIR/NEf/ZEt+82FbUG6+ZBPNsYaKhRFweHDR237EHPB3Cacu0hmRM4qtsJ0nUloRKxFwi5egMdExBRJOalFw0oSCTmOpCxuOV/J5Q3ZzFo80iVgJBXd2oU6Vi/U2Fe4EIhCcOHCBXR3d2NoaCjnTVCrJwQQCATQ1dUFQLtuE4kEwuEwQqEQ5ubmkEgk4Ha7tWi4pQXNBw/Cf/Lk1n5FEWRsbCsqHh4GGR/X1BSpdVxtbSBu99aMtdQPOw+2b0bEqqrqRHzlyhX8/u//PgRBwDve8Q585jOfKel9URQFH//4x/Hkk0+ip6cHp0+fxjvf+U4cPXq0pP1ZwbYg3WKNzNNH5OzYsQOXLlXusZcVyUSRtQmzBHJpVJCNiBVFharmImKXqRXaSMS8V8BgxyAmNychqyUoEUzFI6MErIH1tiWgXiNdRpgBTwAnT5601BKfDkIIfD4ffD5fhrdtKBRCOBzG8vIyYrEYeJ7fSk309CBw6NDWtaCqIFNTWn54eBh07179b0YJpqqqOhGz3wyBQAD3338/nnzySfzgBz9Aa2srVlZWSnx3gPPnz2NwcBD9/f0AgPe+9734/ve/75BuOWBz0gCzeXl/f78+IgfQvBcqga0iWWW9bXlemxGWTsQsIk4kjETM60TcFmjDN37jGxAVEaMbo7iydgWvr72OK+tXMLE5AUXNTyYs0t2OErBiUI+kK0kSRFGEx+PBro5dJRFuPrjdbuzcuRM7d+40HTMcDiMcDmNqagqRSESPWpubm9Hc3o7m224Df/vtOfebi4gBIJFI4Etf+hJmZmbg8XjgdrvLsnVk3toMPT09eO6550renxVsG9LN56kriiKmp6f1ETnMvNwIu+egsSIZx3FVt1tkYERsbHUyE3ECqqrglVdG0dLSgu6WbhwaPIQ7Dt0BABAVESPrI3h97XUMrw/j9bXXMbE5AdXQpiwpcaiq2gASsAonGOoovaDVDMzXXrEz9UqFy+VCR0cHOjo69GWKouhpgfn5eUQiEaiqikAgsJUnbm7OOxeQ4zi8+OKLuO+++/DOd74TExMTBecI1iu2DelmAysKLC4uoqenJ+eIHKCsjkQTMotkPOqpBJROxG63dncPhUK6lAiAnk/b17IPRw4e0W9SSTmJ4fVhXJy+iGfHnsVsYAPTHFdHr7A2qJdIl9UM0rsms5neVws8z6O1tdXklcD8HcLhMFZWVjA2NgZZlk3KiebmZni9XiSTSXzhC1/AuXPn8LWvfQ3Hjx+37dy6u7sxMzOj//fs7GzFDdG3DemmR7obGxsYHh4Gz/PYs2cPhoaG8m5f7pN/eufaVpGsvulIlpHxhWCRSSgUwszMjP6I2NTUpOmI15I4IZzAu9/xbvC8H2/+OzeSBulaQk5AVETbX3ldu4zVWDLGnqwEV/apIdWKdK2C4zidWBnYMMpQKITNzU389Kc/xSOPPAJRFHHw4EHcf//92L17t63ncfr0aYyMjGBiYgLd3d34zne+gyeeeMLWY6Rj25AuQyQSwfDwMAghOHbsGERRxNLSUsHtOA7w+YB4vLjjZRbJ2MVdzwyxBVXVfozZlmyRiSiKGB4exuzsLPx+PxKJBC5fvgyfrxmKfAxuzgOfd0v9oVKKpJww6YjLJ+L6fU8r3QKcC+xmr6pq3hFGVqdH1xLGYZTt7e144okn0NXVhQcffBCJRAIXL17Ejh07bCVeQRDwV3/1V3jrW98KRVFw11134dixY7btP+sxK7r3KiKZTOLKlSuIRqM4ePAg2tvbAQDhcLgIp7HiSLdaRbJKQxSRc24bpRSLi4uYnJxET08Pjh49qn+xFUVBKBTWW6ZVVQFAUikMHh7eA5/LSMQqEnISCSWBpKxpiaUKRMRZX0eF91/NqcgMsqw18+SahmxEvUW6+fA///M/uP/++/Hud78bv/jFL/SU4Dve8Y6KHO9tb3sb3va2t1Vk39mwbUh3eXkZnZ2dOHbsmKloVYynbiBAsbpq1d6xtkUyOyFJ2Uk3FApheHgYTU1NOHXqVEaDCc/zaG9vg9fr1mcrasNxNQOWrYkJW0Ts5T3w5yBiFhFXi4htRRULaSpVkYgXNxPPK1TWO8QOJBIJPPLII3jmmWfw+OOPVzzirBW2Denu378/60BHo2SsEAoFq1tFMlqXRbJSkf4gIIoiRkZGkEgkcOjQoYJjUQRha74cMYwXZ9CcsDQdsZmINR1xTiKW4zoZS4q1p5XcqLD3QpVIl5kwFTumqd4j3QsXLuCBBx7Ae97zHpw7d66mI6gqjW3zynLd7XmeL5t0s9k7akW7xidcYIt0VVXF9PQ0FhYW0N/fj127dlmKooykmw2aExYPowNLNiLWOthyE3E9T4AlFXYYU1QFibhmilRKKquW6oV8iMfjePjhh3HhwgV885vfxJEjR2p9ShXHtiHdXCjOU9c8PSJz4KQX2gju+v3ylwJJIlhdXcHo6Ch27dqFG2+8sSiT9lL83PMRseY1kU7EfErqVqeoYKSry8B83pInPtdjpPvcc8/hE5/4BN7//vfjqaee2tbRrRHb5lXakVc1NkhIkpRyHHOV3bZbz1BVFS+++AoOHFBKbhO1q+OXEbFxzLqZiCWdiLec13hr/g4V/ugqoV5ghTJmMVoO6km9EI/H8bnPfQ4XL17EE088gUOHDtX6lKqKOg4d7IWV6LSpSSuSRaNRXajt8dTPxWontJ78JOLxBLq69uHaa68tuU20kgEKI2KPxw2fz4dAIACfzwee5/XOq2g0ilgshmQyCUmSoaqZnzXP8ej070Szuwkuzv4TtjO9wPSqoijC5/cVtCW1gnoppD3zzDO47bbbsH//fjz11FNXHeECV0mkKwgCFEXJ+/gSi8WwsbGJRKJzWxXJsoH15GtSIw+83sJuZ/lQoXFxObE1LWFrmXFsjSRJoHQrImZR8Q7f1qRoRVVMiomEnIBUiuEPg03pBaNfgp2P27VOL8RiMTz00EN46aWX8J3vfAcHDx6s6fnUEtuGdPNBEARIkpT1IpYkCePj41hfX8fevccRCPi3VZHMCDbJmOd5k9TIoow5JwTB2iTlSiLb2BpVpbp8TRQZEXO6fM3H+xFwbT22y6qiN3TEZU1LbJWIy20DzuaXYCdqVUijlOKZZ57BJz/5SXz4wx/Gl7/85YoNdW0UXDWkm65gUFUVMzMzmJ2dxYEDBzA0NITlZW7bFckAjXy0EUCAz5c5WViWCRop0rUKjiPguHxEzCbqbhGx//9v78zDmyrTPnyfdN9SbAsFylKglJal0BWQRRYRRBQGRkFQZHFgQAUFGRQUURQFdURwBhgLlEUsoIXxk0UUFdGBLiyKUGhpWcvalqZ7kybv90eaQwOUhm7pcu7rynU1ycnJe9JzfnnyvM/7e+yccbE3F+LC21bW3dUCsxKRbpG2iGLdnX4JVYk1It28vDzefvttTp48ybZt2/Dz86vxMdRG6k1O916RQelaXSEE165d49ChQ+h0OiIiImjWrBl6vR4np9rjFFVVFBVpKSjIx87OrsyuGJWPdCv3+ppEpTJGxA4ODjg7O+Pi4iKnk/R6PQUFBSU54gKKirRgELjYOePl5EkLNx/8HmiH3wPtaOHmg5ezF672LtiqbCoU6ZrmD8DoE1udEWBNiq4QggMHDjBo0CACAgL4/vvvKyy4GRkZdOvWjW7dutG0aVN8fHzk+9p71SkC48aNo0OHDnTu3JlJkybdsTI1Pj4eW1tbvvrqqwqNraLUoculfO5l76jT6cjKyiIpKQlnZ2dCQkKws7OT3eqNnp91e2VZaXS6YrRay2a+yzl3y6W2RrqWYoqIS18OBsOtHLEpIjaaxxtL11zsXHC1v2VN1yPQmxH9m3Eq45TRjzjjNDcLb971/Sz1S6hKakp0c3Nzeeutt0hKSuLrr7+mXbt2ldqfp6cnx44dA2DhwoW4urry6quvAkZTq3tNMo4bN45NmzYBMHbsWCIjI5k2bRpg/MKbO3cujzzySKXGVxHqleiWhRCClJQU7O3tCQwMxMXFxUxsTVGys3PdTy3o9QaKisybXZZHZSPdui66d+PuQmxArzdQXHynENsKie5Nu9O7ZW95++t512UBPp1pNIdPz0232C+hKqnunK4pun3ttdeYMmUK//rXv6r9yyQsLIwePXowefJk+vfvf8e5XtpPISIigkuXLsn3V6xYwahRo4iPj6/WMd6NeiW6t0e6Op2OlJQUrl27RuPGjQkMDJSjF2PRvflJUVWeutbAFD3p9YaS3KDlJ7yFC/bKpC6lFyrDrTZJdwqxVlvA778nyaWGarUatVpNz2Y96duqL4WFhZw+fZqs4iyKGxWTkp0id+nQFGmqfezVGenm5OSwYMECUlNT2bFjR421Rk9KSmL37t189tlnvPDCCzz77LNMmDCB5s2bm22n0+nYuHEjn376KWDsFrF9+3Z++uknRXSrCtNy1rS0NHx9fXF3d+fmzZtotVpUKpVZdFuauhrparU6edVcRaInna5yaZX6GOlaikmImzTxICwszKybrsmcu6CgAL1ej7e3N11adkGtVjPQdqC8jyu5V4zRcIkIn8o8RU5RTpWOszrqdIUQ7N+/n9dff51p06axcuXKGu2NZ2Njw7Bhwxg2bBg3btzg9ddfp1WrVvzvf/8jIiJC3m769On07duXPn36APDyyy+zZMkSq/Xxq1eia7IhTE1Nxdvbm+4lbaILCwu5ceMGhw8fRpIkOQpxd3c3+wnu5QVLl+pITFSRmChx8qSK7GwrH9Q9KC7Wl1hLVmw9vonKpxesXzJmbUxfPKW76bq5uZGdnU3Tpk1p2rQpubm5XL9+nZSUFPR6vdyuRq1W09unN/1a95P3dyX3ill++HTm6UoJcVWvSMvJyeGNN97gwoULfPPNN7Ru3bpK928pGo2G6OhooqKisLe3Z+3atWadJd5++21u3LjB6tWr5ccSEhIYM2YMAOnp6ezatQtbW1tGjBhRI2OuV6J75swZCgsLzSbJhBA4ODjINnFGD9hsNBoNKSkp5OXlYW9vL4twSIiaBx+8FRWkpUmcPCmRmKji5EkVyckSRUXWOkIjphIwACcn50q3OW9I1QvVReloX6/Xc/bsWW7evElAQIDs0qZWq+WfvkII8vLyyM7O5tq1a5w5cwa9Xo+rq6vcN6xvi770b91f3u/lnMskZiTKUXFSZhI5WsuEuKrSC0IIfv75Z+bNm8eLL77I6tWrrRYxPvPMMxw8eJAnn3ySDRs23NEdJjIyku+++459+/aZjfHs2bPy3xMmTGDYsGE1JrhQz0TXz88PvV5/10kyE0YP2Adkk3MwmtpoNBq5T1hRUZGcl3N3d6d/fzWDBhk/Kr0eUlONUXBiolGMz56VZD/Z6sZkfuLg4GjmUVAZlIm0ymP6DDIzM0lKSqJ58+aEhYWVOZFpan/k6uoqC7Gpb1h2djZXr14lOTkZg8EgC7FaraZfy34M9DWmJoQQpOWkyd2bTUKcq8294/2qIr2QnZ3NG2+8QVpaGv/3f/9Hq1atKr3PyvDUU08RFRVV5sq9v//977Ru3ZqePXsCMHLkSBYsWFCTQ7wrUjmLAepUknPOnDm4uroSFhZGaGgobm5uFVrZY1r7bhLi7OxsOQpxd3eXIxHTt2dBAZw6JZWkJYy3q1er9thMy0Pt7Oyxt6/aLqiTJxczcWLFV1TNnWvHb7/Vm5LvCvHkk1oefvhPtFotAQEBODk5lf8iCzAJsUajkVubCyHMhNjV1VWu8RVCcCnnkpyWOJVxiuTMZPaM2VPhVW5CCH788Ufmz5/PzJkzmThxotWi2zpEmR92vRLd06dPc+jQIWJjYzly5AharZbOnTsTGhpKeHg4nTp1qnDbZoPBIDdr1Gg0crNGUzSsVqvN8sOZmXDqlKrS+eFbS3dVODg4VEuXivHj9UyZUvEShvnz7di/v+FehDqdjoceOs/s2fYWexBXhtLnYnZ2Nrm5uXcIcemgwLjqrmJj0mg0zJs3j+vXr7Nq1SpatmxZlYdSn2kYons7hYWFHDt2jEOHDhEfH8+JEydwdnYmNDSUsLAwwsLCaN26dYW/tYuLi+UTX6PRkJ+fj4ODg5kQOzjcyqXdT37Y5AImhAEHh/srAbtfnn5azwsvVFx0Fyyw48cfG57olvZLmDxZxbRp1rtcDAYDOTk5ZkJsSmGUjogtPdeFEHz//fcsWLCAWbNmMX78eCW6vT/KFN16ldO9HUdHR3r06EGPHj0A44mUmZlJfHw8hw4dYuvWrZw/f54WLVoQHh4ui/EDDzxgYccEWzw8PPDw8JAfKyoqQqPRoNFouHDhAlqtFmdnZ1mEb88Pp6RIcjRsyg8XFhpbstjbO5jVhFYXykTa/WNqm2PyS3BwKAas0xEYjKVrt3dw1uv1ckR88eLFO4TY3d0dFxeXO8Q0KyuL119/nczMTHbv3o2Pj09NH069pl5HupZgMBg4d+4csbGxxMbGkpCQQE5ODoGBgbIId+3atcKrh0rXbZpyxAaDATc3N1mITRHIjRs3OHnyLDk5LcjK8uH0adtqyQ/fzvDheubMqXik+957duze3TCiIFO6x+jfcKsM629/K+a556wnupai1+vviIhVKhV5eXn8/vvvODo6sm7dOubMmcMzzzyjRLcVp2FGupagUqlo27Ytbdu25emnnwaMObrjx48TGxvL+vXr+eOPP7C1tSUkJISQkBDCwsJo3769RQYlpes2mzVrBpj/FLxw4QLZ2dny0tA2bVri5dUIJycDkmQMQTMzKTVJZ8wP51Rh7XxlV6QZ63TrN0JQ4pegv6tTW13RJhsbGxo1akSjRo3kx/R6PceOHePAgQOkpKTg6OjI6tWrsbe3l+tZFaqOBi+6d8POzk4W2GnTpiGEICcnh4SEBGJjY1m0aJHcT6x0ftjb29uitITpp6CzszP5+fnY2tri7++PSqVCo9GQnJws54dN0XB4uDu9et2KrKqyflhZkXZvTG1zTKbvd6OufgZCCL777jvefvtt5s6dy9ixY1GpVOTk5JCfn2/t4dVLFNG1AFOVwoABAxgwYABgPFkvX74spyVWrlxJeno67du3l0vWQkJC7mo6Y3rthQsXaNWqFf7+/vI2pfPDhYWFZGdnk5WVJeeHXVxcZCEeMEDNoEHGq710ftgkxufOWVY/3NBdxsrCOJlpXIRSnnlQXcxrZ2ZmMnfuXAoKCti7d6/8SwzAzc1NXtShULU0+JxuVaLX60lMTCQ2Npb4+HiOHDmCXq8nKChIjobT0tK4cuUKvXr1ok2bNvdVwlZ6FZOpblMIgZub210nRvLz4fTp8uuHe/UysGRJxWfTli+3ZevW+qW8Jj8LS9vmvPJKMaNG1f6cLhjPo507d7Jo0SLmzZvHmDFjqr3MrQGi5HRrAhsbGzp37kznzp2ZPHmyvMji8OHD7N27l6eeegpJkmjbti3JycmEh4cTFhaGj4+PRRMWd1vFZJqh1mg0nD9/ntzcXGxsbORoOCDAnW7dHM3qh2/PDysuY7e4VQZmc19+FnXlM8jIyGDOnDkUFxfzww8/4O3tbe0hNTjqyKlSN5EkCWdnZ/r06cPWrVtZtWoVQ4YMIT09XU5LbNiwQXZDM0XDISEhuLu7WxR9mAS2dKmQTqeTo+GrV69SUFCAo6OjHA2Hh6vN8sM37+61bTH1Jb1gXGKtv29rTKj9n4EQgm+++YbFixczf/58Ro8erUS3VkJJL9QCDAYDZ86cITY2lri4OA4fPkx+fj6dOnWShbhz585mCy3uB5PXrqlkTaPRUFxcLLtcubu74+bmVuF2MWvW2LJuXS1XnXtgLAMrxM7OrsLtzufP1/Hoo7Wz3VN6ejqzZ89GkiQ+++wzmjRpYu0hNQQa5oq0uoxWq+XYsWOyEP/55584OjoSHBwsC3Hbtm0rXEdpyg+bhDinpAbt9vywJdFQVJQNkZF170fTrbY5oqRPWsUjvzff1DF4cO0SXSEEO3bs4IMPPuDNN9/kySefVKLbmkMR3bqOEIKsrCzi4+PlibrU1FR8fHwICQmRV9R5eXlV+MIyFc6bhDgvLw9bW1s5P+zu7n5X/4eEBBVffmlT5fXD1Ymph1xVrfpbuFDHww/XHtG9fv06s2fPxs7OjhUrVtC4cWNrD6mhoYhufcTURv7QoUPExcURHx9PVlYWHTp0kE1+unbtipOTU4WFWKvVmq2mKywsxMnJycxfonQFRm30Hy6NwSBKJsqkKjUQevddHf36WV90hRDExMSwdOlSFi5cyMiRI5Xo1jooottQKC4u5sSJE7LJz7Fjx5AkiW7duskLOTp06FDh/K2pdrV0fri0+bYpP2xKe1Smfriqud0voSp5/30dffpYV3SvXbvG7NmzcXJy4tNPP8XLy8uq42ngKKLbUBFCkJuby+HDh+W0RFJSEp6enoSGhhIaGkpERARNmzatcERU2vPVlB++V1skS+uHqwq93lgGZmtrU+HJyPJYulTHgw9aR3QNBgNff/01H330Ee+88w4jRoxQolvro4iuwi1MveTi4uLkiPjq1av4+fnJaYng4GBcXV0rlR8unZa4vS2SWq02MxGqLn+JwkKjX4Jxoqz6DBI+/lhH9+41L7pXr15l1qxZuLm5sWzZMjw9Pe+63cWLFxk/fjzXrl1DkiSmTJnCzJkz2bZtGwsXLiQxMZG4uDjCwsLKfc+MjAwGDhxtnPOeAAAUtUlEQVQov7+NjY2cM46Liyu3AmTcuHEkJCRgZ2dHREQEq1evNktRxcfH07NnT6Kjo/nrX/9q6UdR21BEV+He6PV6kpKS5Pzw0aNH0Wq1dOnSRRbijh07VtgEHszbImk0mjvaIqnVarPVX5XJD5v8Eqqj08bdWLZMR1hYzYmuwWBg69atfPLJJ7z77rs88cQT9/yCvHLlCleuXCEkJIScnBxCQ0PZsWMHkiShUqmYOnUqH330Ubmim5mZabZUfeHChbi6uvLqq6/e83Wm1I6Liwu7du3i0UcfBWDs2LH07duXadOmAcbzcNCgQTg6OjJp0qR6Kbp1r85HoVqwsbEhMDCQwMBAJk6cCBi9H44ePcqhQ4dYsWIFJ06cwNXV1czkp1WrVhZHkPb29jRu3FiOikq3RTK1K7+9LdLAgW4MGmR5fvh+/BKqkpp0Wrt69SozZ87Ew8OD/fv3m4lgWTRr1kz2VnBzcyMwMJC0tDQGDRpU7muFEPz0009ERkYSGxtLSkqKxWNNTEwkMjKSmJgYYmJiCA4OZujQofLzERERXLp0Sb6/YsUKRo0aRXx8vMXvUddQRFehTBwdHenZs6fc2E8IQUZGhmwCHx0dLZv2mEx+QkNDLTaBN63Yc3Z2NrO9NBlvX7p0iZycHFQqlRwN+/ioad/emeHD78wPx8Xlcfy4Hr1ebZFfQlVSE29nMBiIjo5m+fLlLF68mMcee6xCXyrnzp3j6NGjdO/e/Z7bXb58maioKDZu3EjHjh2ZNGkSGzduLHf/eXl5bN26lTVr1gAwceJEFi5ceIeBjk6nY+PGjXz66acApKWlsX37dn766SdFdBUUwCiSXl5ePProo/LPQ4PBwNmzZ4mNjeXHH39k6dKl5Obm0rFjRzkiDgoKstgE3iSwarWaFi1aAOZtkc6cOUN+fj729vby8mdfX1sMhlRCQ11p164d2dl6EhNFtfkP343qXgZ85coVZs6cSePGjdm/f79ZN+v7ITc3l1GjRrFs2TLUanWZ28XFxfHggw/y/PPPc+DAgfuqhGjWrBlBQUFERkYSEBBQ5nbTp0+nb9++9OnTB4CXX36ZJUuWVCj3bjAY6ozhuiK6CpVCpVLRrl072rVrx9ixYwFj/s5kAr9u3TqOHz+OnZ0dwcHBcn7Yz8/P4oukrLZIWVlZnDt3jtzcXOzt7bG1teXSpUuo1Wq6d1fTq1fV5IctobpE12AwsHnzZj777DPef/99hg4dWuGUiU6nY9SoUYwbN46RI0fec9ugoCDWrFnDmjVrGD58OBMmTGD06NH3FGoTX331FWvWrGHkyJGMGTOG5557jtatW5tt8/bbb3Pjxg1Wr14tP5aQkCCbpqenp7Nr1y5sbW0ZMWJEme9lKMkt1RXBBWUirUYoKipi/PjxHD58GE9PT7Zs2YKvr6+1h1VjCCHIzs6WTeDj4uJISUnB29vbLD98P510b968SVJSEt7e3rRq1QpJkixuiwRVXz8cFaXFz69qL5fLly8zY8YMmjVrxscff2zW7eF+EULw3HPP4eHhwbJly+54vl+/fmVOpCUnJ7NmzRq2bdtGz5492bRpk/zcvSbSMjIy2LRpE+vWrcPLy4vIyEh8fX2JjIxk7dq17Nu3r8xW9RMmTGDYsGH3nEgr3eX4l19+4cMPP2TJkiV07Nix3M+jBlCqF6zJv//9b/744w9WrVpFdHQ027dvZ8uWLdYellURQpCWlkZsbKxctpaRkYG/v7+cHw4ODr5jMkyn05GcnExhYSGBgYFlXrRg3hZJo9HItpel64dLr9YrKIBTpypWP7xxo5Y2barmcjEYDGzatImVK1eyZMkSBg8eXOkJwV9//ZU+ffrQpUsX+Ytn8eLFFBUV8dJLL3Hjxg0aNWpEt27d+O677+66j+LiYnbu3Mnw4cPlxyytXoiLi6NZs2a0bNkSW1tbWrduLed4R44cyYIFC8y2t0R0TcyaNYsNGzYwY8YMpkyZQpMmTWpD5KuIblUxYsQILl68SGFhITNnzmTKlCnlvmbw4MEsXLiQnj17UlxcTNOmTblx44ZSwH4ber2ekydPytHw0aNHEUIQFBREaGgoGRkZZGRkMGPGjAov5tDpdGb+Ere3RXJ3dzerMy1dP3zqlDE/nJ19536//FJLy5aVv1wuXbrEjBkzaNmyJR999JGZZafCnbnb48ePM336dL755hs5z63RaGrD56aIblVhqlMsKCggPDyc/fv3M336dE6fPn3HtrNmzWL8+PF07tyZPXv2yBND7dq1IzY2VlmmWQ6mTsq7d+/mrbfeQgijG5ibm5tcKREeHo6Pj0+lvsBMbZFMQnx7WyS1Wm22bPhu+eGNG7U0b17xy8VgMLBhwwZWr17Nhx9+yKBBg5Qv5VKYdOr2z+TIkSPMmjWL0aNH4+rqyrJly2jevDm9e/dm0qRJ1jT6Uep0q4rly5ezfft2wLjKJzk5ucGnCqoLUyfl/Px8VqxYwYABAxBCcOPGDdkEfv369Vy+fJk2bdqYmcCr1WqLRcvR0RFHR0fZZ9Yk9iYT+OTk5DvaIg0c6GJWP1wZfbx48SIvvfQSbdu25cCBAxZNVjU0TP/Lffv2sWbNGoKDg+natSuPPPIIQ4cOZf/+/aSnpzN37lyuXbvGypUrGTJkSK10V1NE9z74+eef+eGHHzh48CDOzs7069ePwsJCRo8efc9I18fHh4sXL9KiRQuKi4vRaDRlLtdUuJPx48fLf0uSRJMmTXj88cd5/PHHgVsm8IcOHWL37t28++67FBYW3mECb6lBuUnsXVxcLG6L5O7ujqOj431FpwaDgaioKD7//HM+/vhjBg4cqES3JQghMBgM2NjYyBNmK1as4L333uOFF14gNzeXcePGsWHDBv7xj3+Ql5eHk5MTKpWKa9eusW3bNlxdXa19GHdFEd37QKPR8MADD+Ds7MypU6c4dOgQQLmR7hNPPMH69evp2bMnX331FQMGDFAuripEpVLh7++Pv7+/LNBFRUWyCfzq1atlE/iQkBBZiNu0aWPxhEtF2iKp1eoyhf7ChQu8+OKL+Pv789tvv9VagbAGhYWF5Obmyuk3SZLQarXs27ePLVu28NBDDwGwdu1ali1bJvtAnD9/nm3btrF48WKmTJlCu3btrHYM90IR3ftgyJAhrFq1isDAQDp06ECPHj0set3kyZN59tln8fPzw8PDg+jo6GoeqYKDgwPdu3eXV12ZTODj4uKIjY0lJiaGs2fP4uPjI4twaGgonp6eFn8h2tnZ4enpKf9qKd0W6ebNm5w7d86sLdK1a9cICAhgy5YtrFu3jo8//lj5Ar6NoqIiXnjhBc6dO8e+fft49tlnmTp1Kn5+fhQVFdG2bVu2bNnCzJkz6d+/P1988YUcFe/YsYOdO3cSExPDgAEDrH0oZaJMpCk0WAwGAxcuXDAzgddoNAQEBNxhAl9RSrdFevPNNzl48CCFhYU8/vjj9OrVi3HjxlW4L1t95dy5c/Tp0we9Xk+XLl3473//ixCCiIgIbGxsyM7O5pNPPpFL1xYtWsSLL74olwPWEpSJNAWF21GpVPj6+uLr6yuvhNLpdLIJ/BdffMGcOXNQqVTyarqwsDD8/f0tNkGXJAknJyc2b97MqVOnWL9+PeHh4fz+++8kJCTUuEdEbeT2MrDU1FTS0tLw8vJiz5498i+BN998k2eeeYb//e9/hIWFkZmZyVNPPYWLiwvFxcUVXhpd0yiRroLCPShtAm9axJGUlETjxo1lEQ4PD8fb2/uuaYKzZ8/y0ksv0aVLFxYvXoyLi4sVjqL2UlxcLH/x5Ofn4+zsDBhX482bN4+srCx27Nghbz9r1iySkpLQ6/WcOnWKIUOGsHLlSquMvRyUOl0FhapCCMGVK1fMTOCvX78um8CHhYXRtWtXvvzyS9lFq0+fPkrutgwMBgNTpkwhKysLb29v5s+fT/PmzYmNjWXcuHHMmzePSZMmAcZfIpmZmaSmptKoUSMCAwOtPPoyUURXQaE60ev1nD59Wq4f3rNnDxEREURFRcnRm4IRvV4vp2eOHDnCk08+SceOHZk8eTJLliyhW7duvPTSS3Ts2JHly5fz/vvv8+effxIdHY1Go2HevHlWPgKLUERXQaEmKW3GomDk119/xcXFheDgYHQ6HXZ2dnz++efk5ubyyiuvAPDee+/xySefMG3aNF577TWcnJwYP3488fHxFBYW8sUXX9C7d28rH4lFKKKroKBgPc6ePcv8+fPx9PSkb9++zJkzh507d+Lk5ESjRo2wt7dnzJgxnD59mqFDh3LkyBFeeeUV2YLy6NGjBAcHW/ko7gulekFBQaHmMaUS2rRpw2OPPcaMGTPYvHkzGzZsoFOnTvJ28+fPx9nZmeTkZHJycmjZsiXr1q3Dz8+PoKCguia490QRXStSlomHgkJ9oHTu9vjx4zg5OdG0aVP8/f3p1q0bYKxe0Gq17Nq1i1mzZgGwfv16hg8fzujRowkKCrLa+KsLJb1QS7C03YjBYECSpAYj1L/88gsvv/wyf/zxR11vyd0g0Wq1PP300xw8eJCtW7diY2PD+vXrkSSJ5cuXy92lp06dyq5du2jcuDF5eXl8++23tG/f3sqjrxRKeqG2cf36deLi4vD09CQ8PPyuRfJ6vR4hBCqVShbkWmDOXKO0atWKqKgoPvroI2sPReE+yc7OZvz48bi5uZGcnCzXKF+6dIno6GjWrl3L1KlTAaPR/2+//ca5c+fMDI7qIw3rCq5FHD16lC1btvDss8/SpEkTPvzwQ3Q6ndk2NjY22NraykL7ww8/8Pzzz5ORkSGnJnbu3MmwYcNqfPwVYcSIEYSGhtKpUyf+85//WPQaX19fgoKCGtyXTX1Aq9Vy5coV3njjDVxcXIiOjpaj3K5du7Jjxw6+/PJLHnroIT755BP69u1b7wUXlEjXaoSGhjJ48GDAKMB/+ctf6N27Nz179mTHjh1s2rSJ1NRUIiIimD17Nu3bt+e3334jIyPDzBayR48ect6sdA6tNrJ27VozA/hRo0aVawCvUHfx8vLC39+fp556Ch8fHxITE+nVqxfnz5/noYceYsiQISxatIjHH3+83HY/9QohxL1uCtVEamqq2LZtm/jzzz+FEEJ06tRJ7Nu3T/znP/8RrVq1ElFRUeLbb78VkydPFnv27BFLliwRHh4ewtfXV7zyyivi+PHjZvsrLi6W/9br9UKn0wmDwVCjx1Qeb731lggKChJBQUFCrVaLgwcPWvza5557Tmzbtq0aR1dzpKeni65du4quXbsKb29v0bx5c/l+UVHRPV87duxY4e/vLzp16iQmTpwotFqtEEKIxMRE0aNHD2Fvby8+/PDDmjgMi8jLyxN79+4V27dvlx8bPXq0+Pbbb4UQQmRmZlpraNVNmbqqRLpW4LvvvuOdd97BxcWFxMREsrKy8PDw4MKFC0RGRvL+++/L7cwfe+wx0tPTGTx4MKtWrWLo0KFcuXKFb775hubNm/Pwww/z9ddf06ZNG3n/pXPApbHmJFxFDeDrI56enhw7dgywrLGjVqtFp9Ph4uLCuHHj5G68Y8eOJTIykmnTpuHh4cHy5cvNfApqA87OzgwaNEi+/89//pO4uDi5UqGumNRUJYroWoF169bh7OzM3r17Adi8eTPr168nLS0NR0dHQkNDgVvpAi8vL1JTU5Ekiblz59KyZUsATpw4wbFjx2TBPX78OB988AEnT56kb9++zJ07V+58AOaTcKKGV0xV1AC+IZOYmEhkZCQxMTHExMQQHBzM0KFD5ecjIiK4dOkSAE2aNKFJkybs3LnTWsMtk4KCAn766Sfmz5+PVqslJiZGLhlriCizE1YgLCwMnU5HTEwMv/76K8uWLcPFxQUfHx8aNWpETk4OcCv1AxAbG0vjxo3l+wAHDx6kbdu2ACQkJDB8+HBat27NP//5TzQaDUuXLgWM4r1hwwb27NnDiRMn0Gg0dxVcg8FAcXExer2+yo95yJAhFBcXExgYyGuvvWaxAXx8fDwtWrRg27ZtTJ061aygvj6Sl5fHunXr6N27N3/729/o2LEjf/zxxx2LA3Q6HRs3bmTIkCFWGqnlODk50bFjR+bPn8+JEycatOCCEulahQkTJpCTk8M777xD//79EULg7e3NoEGDiI6OZt++fYSFhWFra0teXh4uLi4cPHiQgIAAGjVqJO9n//79snitWbOGgIAAFi9eDBgF9MUXX+TixYu4urqyfft2Tp48Sffu3dm6dSvDhw8nKirKzKC7rLREVeDg4MDu3bvv+3Xh4eFyNNcQaNasGUFBQURGRhIQEFDmdtOnT6dv37706dOnBkdXcUy+xQqUO5Gm3GrgBngBbUv+7gUcBc4B/wV6lzy+DXjzttclAZNK/o4H/l7quU7AdmAo0Bo4AKwvea4V8DMwvuR+B2A1cBrYDAwD7Kz9uTSEG7AQeLXU/UeALcBJYAHQ+i6veQvYAajK259yq303Jb1QCxBCpAshUkv+/k0IEQz8FYgBzpZsFg08J0nSOkmSgiRJegCjeP5S8nwrIKXUbpsCjTCKtx+QC/y75D0uYEwt+ZRs+w7GFTTjgKvAAKBJ1R+pQnkIIfYKIUYDfQAN8F9Jkn6QJMkXQJKk54HBwNNCCIPVBqpQYZT0Qi1FCJEAJJR6aDtwE3gIo1iqAHshxBnJmKDdATwJfF+y/avAKYzRcD8gnxIBlyRJhTG6XVSyrTtw2vSekiS5oiwBtypCiAzgU+BTSZIiAFOifRVwHjhYkpePEUK8I0lSU4znixowSJL0MtBRCJFd86NXuBfleS8o1GIkSfIsuTiRJKkV8CXQEqPQ6oDJwA2M4vqAEGJqqW3PAe5CiBxJknpj/FnqAfxLCLGmhg9FQaHBoES6dRiT4Jb8fQHoJUlSF6A5kCCEyJAkqQ3gD5wp9dIw4HKJ4LoJIX6VJOmvGCPlv0qSdFIIcbAGD0VBocGgRLoNAEmS/DBOuiSV3F8FdBVC9JQkaRSQI4TYK0mSLbAXOCaEmCVJkiSUE0RBoUpRIt0GgBDizG0P/QMwLQXyAeaV5AfPYswD174KewWFesL/A0qJupgGmnKSAAAAAElFTkSuQmCC\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- }
- ],
- "source": [
- "import matplotlib.pyplot as plt\n",
- "\n",
- "acts = ['a=0','a=1']\n",
- "stas = ['T<21', '2124']\n",
- "colors = ['b', 'g', 'r']\n",
- "\n",
- "fig = plt.figure()\n",
- "ax = fig.add_subplot(111, projection='3d')\n",
- "ax.set_xlabel('actions', labelpad=6, fontsize=12)\n",
- "ax.set_ylabel('states', labelpad=10, fontsize=12)\n",
- "ax.set_zlabel('$\\mathbf{q(s,a)}$', labelpad=0, fontsize=15)\n",
- "plt.xticks(ticks=range(len(acts)), labels=acts)\n",
- "plt.yticks(ticks=range(len(stas)), labels=stas)\n",
- "\n",
- "for i, s in enumerate(stas):\n",
- " x = np.arange(len(acts))\n",
- " h = model.q[i,:]\n",
- "\n",
- " # Set color\n",
- " color = [colors[i]]*len(acts)\n",
- "\n",
- " # Plot the 3D bar graph\n",
- " ax.bar(x, h, zs=i, zdir='y', color=color, alpha=0.8)\n",
- "\n",
- "plt.show()"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "V0AIl-HeVyqs"
- },
- "source": [
- "Does it sound familiar? this is actually the [q-function that we had conceptually introduced before](#qFunctionConcept), but for our specific case!\n",
- "\n",
- "We observe that the state with the highest value is the one in the middle (green bars 🟢👌, `2124`), there is more value on `a=0`, so there is a preference for the agent to turn heating off. \n",
- "\n",
- "Sometimes it is useful to know what is the value of being on a specific state, independently of the action to be taken. This is represented by the so-called state-value function, which relates to the action-value function as follows:\n",
- "\n",
- "\\begin{align}\n",
- " v(\\pmb{s}) = \\max_{\\pmb{a}} q(\\pmb{s},\\pmb{a})\n",
- "\\end{align}\n",
- "\n",
- "At this point we can easily compute and plot the value function for our case:\n",
- "\n",
- "\n"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 22,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/",
- "height": 288
- },
- "id": "urJOkjSNoa-h",
- "outputId": "57f5b577-d685-431b-9296-649aaba81837"
- },
- "outputs": [
- {
- "output_type": "display_data",
- "data": {
- "text/plain": [
- ""
- ],
- "image/png": "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\n"
- },
- "metadata": {
- "needs_background": "light"
- }
- }
- ],
- "source": [
- "# Compute the state-value function\n",
- "v = np.amax(model.q, axis=1)\n",
- "\n",
- "# Plot state-value function\n",
- "fig = plt.figure()\n",
- "\n",
- "ax = fig.add_subplot(111)\n",
- "ax.set_xlabel('states', labelpad=10, fontsize=12)\n",
- "ax.set_ylabel('$\\mathbf{v(s)}$', labelpad=0, fontsize=15)\n",
- "plt.xticks(ticks=range(len(stas)), labels=stas)\n",
- "x = np.arange(len(stas))\n",
- "ax.bar(x, v, color=colors, alpha=0.8)\n",
- "plt.show()"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "clFn8dd7obRI"
- },
- "source": [
- "Notice that we have trained our agent following an off-policy method: the actions were driven by a policy different than that one that our agent would follow. This is because the agent was using an epsilon-greedy policy to explore more rewarding actions. If we conclude we are happy with the learned policy, we can test it by setting `deterministic=True` with the `predict` method. For example, let's test our learned agent for the first day of February: \n",
- "\n"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 23,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "WuYEBf9nsmH6",
- "outputId": "d63ce6c1-f06d-49a6-c16e-e3fa226dbbe5"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "-------------------------------------------------------------------\n",
- "State [Bin #] = 1\n",
- "Action [ - ] = 0\n",
- "-------------------------------------------------------------------\n"
- ]
- }
- ],
- "source": [
- "env.stop()\n",
- "env = BoptestGymEnvCustomReward(url = url,\n",
- " testcase = 'bestest_hydronic_heat_pump',\n",
- " actions = ['oveHeaPumY_u'],\n",
- " observations = {'reaTZon_y':(lower_setp,upper_setp)}, \n",
- " random_start_time = False,\n",
- " start_time = 31*24*3600,\n",
- " max_episode_length = 24*3600,\n",
- " warmup_period = 24*3600,\n",
- " step_period = 3600)\n",
- "env = DiscretizedActionWrapper(env, n_bins_act=1)\n",
- "env = DiscretizedObservationWrapper(env, n_bins_obs=3, outs_are_bins=True)\n",
- "\n",
- "done = False\n",
- "obs = env.reset()\n",
- "\n",
- "from IPython.display import clear_output\n",
- "while not done:\n",
- " # Clear the display output at each step\n",
- " clear_output(wait=True)\n",
- " # Compute control signal \n",
- " action = model.predict(obs, deterministic=True) \n",
- " # Print the current operative temperature and decided action\n",
- " print('-------------------------------------------------------------------')\n",
- " print('State [Bin #] = {:.0f}'.format(obs))\n",
- " print('Action [ - ] = {:.0f}'.format(action))\n",
- " print('-------------------------------------------------------------------')\n",
- " # Implement action \n",
- " obs,reward,done,info = env.step(action) # send the action to the environment"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "sLBD3joyxe9Z"
- },
- "source": [
- "Now there is no randomness involved. The agent exploits its policy by ALWAYS picking action `a=1` when `s=0` because it has learned that that is the action with the highest value in that state. \n",
- "\n",
- "We can now evaluate our learned policy by calculating the core KPIs with BOPTEST:"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 24,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "eLzZaaNzyeZv",
- "outputId": "ab1c7d0d-8eb3-4192-cfe5-6f6848dce054"
- },
- "outputs": [
- {
- "output_type": "execute_result",
- "data": {
- "text/plain": [
- "{'tdis_tot': 1.7415794961384694,\n",
- " 'idis_tot': 0,\n",
- " 'ener_tot': 0.17501300879744733,\n",
- " 'cost_tot': 0.044365797730152895,\n",
- " 'emis_tot': 0.029227172469173696,\n",
- " 'pele_tot': 0.01990768126278055,\n",
- " 'pgas_tot': None,\n",
- " 'pdih_tot': None,\n",
- " 'time_rat': 0.0003869439121605693}"
- ]
- },
- "metadata": {},
- "execution_count": 24
- }
- ],
- "source": [
- "env.get_kpis()"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "WbDAlStV2Tvx"
- },
- "source": [
- "This prepares the ground for different RL configurations to be evaluated and compared between each other and to other types of controls like classical rule based controllers or more advanced model predictive control. Recall that there are specific [scenario periods for each test case in BOPTEST](https://github.com/ibpsa/project1-boptest/tree/master/testcases#test-cases) that are set for these comparisons. "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "FBL289bfsmcJ"
- },
- "source": [
- "**Food for thought: 🤔**\n",
- "- If the agent never receives a reward when the temperature is out of the comfort bounds (states 0 🔵 and 2 🔴), why is the q-function not 0 for those states?\n",
- "- Could you think of measures to improve learning?\n",
- "- Could you think of measures to improve performance?\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "eKqvn5yb_mqJ"
- },
- "source": [
- "# **Gearing up** 💪"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "X1sdwYm5b66G"
- },
- "source": [
- "The previously stylished example had a very limited representation of the state space. It was useful to illustrate how we can configure and train a RL agent without needing too many interactions with the environment (our building). However, using RL for solving this environment may feel like overkilling the problem. Our `SimpleController` was already enough to decide when to turn on heating based on indoor temperature readings. You should note, however, that you have developed a general agent capable of learning from any environment and the potential to infer way more complex relationships between environment observations and actions. Examples of what this RL agent could infer for building control are the following:\n",
- "- Dynamic energy pricing\n",
- "- A heating schedule based on user inputs. \n",
- "- A heating curve based on ambient temperature.\n",
- "- The variable heat pump COP based on condenser, evaporator, and ambient temperature reaadings. \n",
- "\n",
- "We could for example extend our reward function as to minimize the building energy use or the greenhouse gas emissions while keeping comfort.\n",
- "And all this can be inferred without the need of a model that requires domain knowledge. On the downside, learning more complex dynamics from higher dymensional observation spaces requires more training data. This means that more interactions with the environment (the building) are required, which sometimes are unavailable. For this reason, sample-efficiency is key in RL and there exist several tricks to expedite learning. \n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "V3ZF29MChF4F"
- },
- "source": [
- "To finalize, we are going to instantiate a more complete building environment by extending the observation space with the time of the week as well as information about the ambient temeprature, solar irradiation, internal gains, electricity pricing, or temperature setpoints. With BOPTEST-Gym we can also establish a predictive and a regressive period that include predictions of the boundary condition data and past observations of the measured data, respectively. "
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 25,
- "metadata": {
- "id": "Bdx3qCDvhFSX"
- },
- "outputs": [],
- "source": [
- "env.stop()\n",
- "\n",
- "from boptestGymEnv import NormalizedObservationWrapper\n",
- "from stable_baselines3 import DQN\n",
- "\n",
- "env = BoptestGymEnvCustomReward(\n",
- " url = url,\n",
- " actions = ['oveHeaPumY_u'],\n",
- " observations = {'time':(0,604800),\n",
- " 'reaTZon_y':(280.,310.),\n",
- " 'TDryBul':(265,303),\n",
- " 'HDirNor':(0,862),\n",
- " 'InternalGainsRad[1]':(0,219),\n",
- " 'PriceElectricPowerHighlyDynamic':(-0.4,0.4),\n",
- " 'LowerSetp[1]':(280.,310.),\n",
- " 'UpperSetp[1]':(280.,310.)}, \n",
- " predictive_period = 24*3600, \n",
- " regressive_period = 6*3600, \n",
- " random_start_time = True,\n",
- " max_episode_length = 7*24*3600,\n",
- " warmup_period = 24*3600,\n",
- " step_period = 900)\n",
- " \n",
- "env = NormalizedObservationWrapper(env)\n",
- "env = DiscretizedActionWrapper(env,n_bins_act=10)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "Is0vIEcP-lmS"
- },
- "source": [
- "This new environment has a way higher dimensional state-action space than the ones we treated before:"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 26,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "GSg90XCe-26Q",
- "outputId": "9a21fa91-b278-4295-a125-a6e1e93ce08a"
- },
- "outputs": [
- {
- "output_type": "stream",
- "name": "stdout",
- "text": [
- "Observation space of the building environment (dimension):\n",
- "(608,)\n",
- "Action space of the building environment:\n",
- "Discrete(11)\n"
- ]
- }
- ],
- "source": [
- "print('Observation space of the building environment (dimension):')\n",
- "print(env.observation_space.shape)\n",
- "print('Action space of the building environment:')\n",
- "print(env.action_space)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "NW3VhsGq-3Ee"
- },
- "source": [
- "Because of this high dimensional state-action space, an agent will probably require many more interactions to solve this environment. Luckily, there are readily available state-of-the-art RL algorithms that implement all sort of tricks to expedite and stabilize learning while maintain the learning principle that you have learned above. For example, we can access the advanced Deep Q-Network (DQN) algorithm from Stable-Baselines3 to learn this more complex environment. We set here our agent to learn for `10` steps to show how this learning process would be initiated. \n",
- ""
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 27,
- "metadata": {
- "colab": {
- "base_uri": "https://localhost:8080/"
- },
- "id": "4NvXkkFh-qH5",
- "outputId": "2c368341-a92a-406a-935b-58258f5577cb"
- },
- "outputs": [
- {
- "metadata": {
- "tags": null
- },
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "Using cpu device\n",
- "Wrapping the env in a DummyVecEnv.\n"
- ]
- },
- {
- "output_type": "execute_result",
- "data": {
- "text/plain": [
- ""
- ]
- },
- "metadata": {},
- "execution_count": 27
- }
- ],
- "source": [
- "model = DQN('MlpPolicy', env, verbose=1, gamma=0.99, seed=seed, \n",
- " learning_rate=5e-4, batch_size=24, \n",
- " buffer_size=365*24, learning_starts=24, train_freq=1)\n",
- "\n",
- "# Main training loop\n",
- "model.learn(total_timesteps=10)"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "07ZJaTPzEo8O"
- },
- "source": [
- "However, this is clearly not enough! Solving an environment of these dimensions would probably require millions of steps or other tricks to accelerate learning. Could you think of any?"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 28,
- "metadata": {
- "id": "EDJHCuQ2NFN6"
- },
- "outputs": [],
- "source": [
- "env.stop()"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "F-_f2qRTB0Nw"
- },
- "source": [
- "# **Further resources** 📚"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "p54OK_TGrtfp"
- },
- "source": [
- "- For RL, check out the resources page from Stable-Baselines 3 [here](https://stable-baselines3.readthedocs.io/en/master/guide/rl.html) and the [open access book of Richard S. Sutton and Andrew G. Barto](http://incompleteideas.net/book/the-book-2nd.html) \n",
- "- For BOPTEST, check out the websites of the [BOPTEST framework](https://ibpsa.github.io/project1-boptest/), its [GitHub repository](https://ibpsa.github.io/project1-boptest/), and its overarching project: [IBPSA Project 1](https://ibpsa.github.io/project1/). "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "Jblq_C7CHQHj"
- },
- "source": [
- "# **Feedback** 💬 "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "jJ9lmUndHLMq"
- },
- "source": [
- "Please help us improve by filling out [this form](https://forms.gle/JdprK6tgxQtwvhFV8). It'll only take a couple of minutes!"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "WNB5MoRmOWc9"
- },
- "source": [
- "#**Annex I: Formal Reinforcement Learning theory** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "G-ZId2TdCngy"
- },
- "source": [
- "In RL we aim to derive an optimal control policy from the direct interaction of an agent (the RL algorithm) and an environment (the process to be optimized).\n",
- "A policy is a mapping from environment states to actions that the agent \"decides\" to take. \n",
- "This control method is based on the principle of dynamic programming. Unlike\n",
- "classical dynamic programming, RL does not assume the existence of a perfect\n",
- "system model and uses function approximations to build a policy from samples\n",
- "of historical data. Hence, the agent performs empirical learning and decides on\n",
- "actions to drive the environment towards favorable trajectories according to a reward function that the environment delivers every control step.\n",
- "\n",
- "The process of the RL agent interacting with the environment is a sequential decision-making problem formalized as a **Markov Decission Process (MDP)**. A diagram summarizing the RL approach is shown in the following figure:\n",
- "\n",
- "\n",
- "\n",
- "*Figure: Diagram of the RL approach. The RL agent decides an action. After the action is implemented, the environment returns the new state $\\pmb{S}_{k+1}$ and associated reward $R_{k+1}$.*\n",
- "\n",
- "In an MDP, the agent and the environment interact during a sequence of discrete-time steps indexed here as $k=0,1,2,...,K$, with $K$ being the terminal sample that could be $K=\\infty$. \n",
- "Every time step $k$ the agent receives a representation of the environment named state: $\\pmb{S}_k \\in \\pmb{\\mathcal{S}}$, where $\\pmb{\\mathcal{S}}$ is the state space. \n",
- "Note that the agent's observation of the state-space may or may not fully characterize the environment state. \n",
- "In the latter case where the agent can only see a partial observation of the environment's state-space, we refer to **partially observable Markov decision processes (POMDPs)**.\n",
- "\n",
- "Upon receiving the state representation, the agent computes its control logic and in turn sends back to the environment a control action $\\pmb{A}_k \\in \\pmb{\\mathcal{A}}$, where $\\pmb{A}_k$ is the most appropriate action chosen from the action space $\\pmb{\\mathcal{A}}$. \n",
- "One time step later, the agent observes a new state from the environment $\\pmb{S}_{k+1}$ along with a scalar value indicating its reward $R_{k+1} \\in \\mathcal{R} \\subset{\\mathbb{R}}$. Notice that the reward $R_{k+1}$ is an indicator of the agent's performance when taking action $\\pmb{A}_k$ from state $\\pmb{S}_k$.\n",
- "\n",
- "The environment $\\mathcal{E}_{\\pmb{f}}$ is governed by the natural laws of the system dynamics $\\pmb{f}$ and it is defined by $\\mathcal{E}_{\\pmb{f}}:\\pmb{\\mathcal{S}}\\times \\pmb{\\mathcal{A}} \\rightarrow \\pmb{\\mathcal{S}}\\times \\mathcal{R}$. \n",
- "The goal of RL is to infer an **optimal control policy** $\\pi_{*}:\\pmb{\\mathcal{S}} \\rightarrow \\pmb{\\mathcal{A}}$ that maximizes the **expected cumulative return** $G$ when the agent acts according to it. \n",
- "The cumulative return is defined as some function of the rewards sequence, and a typical definition is to discount the rewards with a **discount factor** $\\gamma \\in [0,1]$ as shown in the following equation:\n",
- "\n",
- "\\begin{align}\n",
- " G_k = R_{k+1} + \\gamma R_{k+2} + \\gamma^2 R_{k+3} + ... = \\sum_{i=0}^\\infty \\gamma^i R_{k+i+1}\n",
- "\\end{align}\n",
- "\n",
- "The **action-value function** $q(\\pmb{S},\\pmb{A})$ estimates the expected return when being in a specific state $\\pmb{S}$ and taking an action $\\pmb{A}$.\n",
- "The **state-value function** $v(\\pmb{S})$ directly estimates the expected return for being in state $\\pmb{S}$.\n",
- "Frequently, the policy and value functions are approximated by **function approximations** to cope with high-dimensional state-action spaces. \n",
- "Examples of commonly used regressors are neural networks or randomized trees. \n",
- "\n",
- "\n",
- "\n",
- "A **trajectory** of an MDP is defined as a sequence of states, actions and rewards.\n",
- "Most of the RL algorithms learn from finite trajectories of experience called **episodes**.\n",
- "Sometimes, the trajectories are broken down into tuples of the form $(\\pmb{s}_k,\\pmb{a}_k,r_k,\\pmb{s}_{k+1})$ and stored in a **replay memory** $\\pmb{\\mathcal{D}}$. \n",
- "Using a replay memory allows to serve the historical data in random batches of tuples to preserve as much as possible the independent and identically distributed assumption that is typically taken to parametrize policies and value functions. \n",
- "\n"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "_RIO07aKaQHG"
- },
- "source": [
- "# **References** "
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {
- "id": "cQM-3Ra5BYM7"
- },
- "source": [
- "\n",
- "- **[1]** *Blum, D., Arroyo, J., Huang, S., Drgona, J., Jorissen, F., Taxt Walnum, H., Yan, C., Benne, K., Vrabie, D., Wetter, M., and Helsen,\n",
- "L. Building Optimization Testing Framework (BOPTEST) for Simulation-\n",
- "Based Benchmarking of Control Strategies in Buildings. Journal of Building\n",
- "Performance Simulation 14, 5 (2021), 586–610. https://doi.org/10.1080/19401493.2021.1986574*\n",
- "\n",
- "- **[2]** *Arroyo, J., Manna, C., Spiessens, F., and Helsen, L. An OpenAI-Gym\n",
- "environment for the Building Optimization Testing (BOPTEST) framework.\n",
- "In Proceedings of the 17th IBPSA Conference (Bruges, Belgium, September 2021) [https://doi.org/10.26868/25222708.2021.30380](https://www.conftool.pro/bs2021/index.php/30380_Arroyo_Javier.pdf?page=downloadPaper&filename=30380_Arroyo_Javier.pdf&form_id=30380)* \n",
- "\n",
- "- **[3]** *Drgona, J., Arroyo, J., Cupeiro Figueroa, I., Blum, D., Arendt, K., Kim, D.,Ollé, E. P., Oravec, J., Wetter, M., Vrabie, D. L., and Helsen, L. All you need to know about model predictive control for buildings. Annual Reviews in Control 50 (2020), 190–232. https://doi.org/10.1016/j.arcontrol.2020.09.001*\n",
- "\n",
- "- **[4]** *Vázquez-Canteli, J. R., and Nagy, Z. Reinforcement learning\n",
- "for demand response: A review of algorithms and modeling techniques.\n",
- "Applied energy 235 (2019), 1072–1089. https://doi.org/10.1016/j.apenergy.2018.11.002*\n",
- "\n",
- "- **[5]** *Chen, B., Cai, Z., and Bergés, M. Gnu-RL: A Practical and Scalable Reinforcement Learning Solution for Building HVAC Control Using a Differentiable MPC Policy. Frontiers in Built Environment 6 (2020). https://doi.org/10.3389/fbuil.2020.562239*\n",
- "\n",
- "- **[6]** *Sutton, R. S., and Barto, A. G. Reinforcement Learning: An Introduction, second ed. The MIT Press, 2018.*\n"
- ]
- }
- ],
- "metadata": {
- "colab": {
- "collapsed_sections": [],
- "provenance": [],
- "include_colab_link": true
- },
- "kernelspec": {
- "display_name": "Python 3 (ipykernel)",
- "language": "python",
- "name": "python3"
- },
- "language_info": {
- "codemirror_mode": {
- "name": "ipython",
- "version": 3
- },
- "file_extension": ".py",
- "mimetype": "text/x-python",
- "name": "python",
- "nbconvert_exporter": "python",
- "pygments_lexer": "ipython3",
- "version": "3.10.4"
- }
- },
- "nbformat": 4,
- "nbformat_minor": 0
-}
\ No newline at end of file
diff --git a/docs/tutorials/CCAI_Summer_School_2022/Building_Control_with_RL_using_BOPTEST.ipynb b/docs/tutorials/CCAI_Summer_School_2022/Building_Control_with_RL_using_BOPTEST.ipynb
new file mode 100644
index 0000000..10b995d
--- /dev/null
+++ b/docs/tutorials/CCAI_Summer_School_2022/Building_Control_with_RL_using_BOPTEST.ipynb
@@ -0,0 +1,3087 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "colab_type": "text",
+ "id": "view-in-github"
+ },
+ "source": [
+ ""
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "V1CcDG8FanTw"
+ },
+ "source": [
+ "# **Key Learning Objectives** 🎯"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "oT2QjTu24zwV"
+ },
+ "source": [
+ "\n",
+ "This is an introductory, hands-on tutorial to guide you through the main concepts of Reinforcement Learning (RL) for controlling Heating, Ventilation and Air Conditioning (HVAC) systems for buildings.\n",
+ "We are going to apply RL to a building emulator from the Building Optimization Testing (BOPTEST) framework **[1]** using the BOPTEST-Gym interface **[2]**.\n",
+ "BOPTEST is a framework for performance benchmarking of control algorithms.\n",
+ "Further information and documentation can be found here:\n",
+ "\n",
+ "[https://ibpsa.github.io/project1-boptest/](https://ibpsa.github.io/project1-boptest/)\n",
+ "\n",
+ "You will learn:\n",
+ "\n",
+ "- What RL is, how it works and how it can be used in the application of building energy management.\n",
+ "- The most popular standard for representing general RL problems: OpenAI-Gym.\n",
+ "- The BOPTEST API and its Gym interface.\n",
+ "\n",
+ "📌 **Note**: This tutorial is prepared for use with BOPTEST v0.4.0.\n",
+ "and uses a web-based version of BOPTEST (called \"BOPTEST-Service\") as not to require installation of any BOPTEST software on a user's own device. It is also possible to use BOPTEST on a user's own (local) device. Both the web-based and local versions have the same functionality, and will produce the same results, with only small changes in the API (changing the BOPTEST-service url to your localhost url, that is, to: `http://127.0.0.1:5000/`).\n",
+ "\n",
+ "**EDIT**: This tutorial was originally developed with BOPTEST v0.2.0. and has been continuously updated to work with the latest BOPTEST versions. Specifically, the following updates have been implemented:\n",
+ "\n",
+ "- **BOPTEST v0.4.0.** *Jul 13, 2023*. BOPTEST-Gym internally updates for new BOPTEST API changes when getting results and forecast. Update to Gym v0.26.2. and stable-baselines3 v2.0.0. Import Gymnasium instead of Gym. Change from `compute_reward` to `get_reward` not to fall into Stable Baseline's convention for goal environments. Use `terminated` and `trunctated` outputs from Gym instead of directly `done`. Return `info` upon calling the reset method of Gym.\n",
+ "- **BOPTEST v0.3.0.** *Oct 25, 2022*. There are just small changes required for this update, basically retrieving the `'payload'` after each request. That is the origin of the differences between the notebook explained in the recording and this updated notebook. \n",
+ "- **BOPTEST v0.2.0.** *Aug 18, 2022*. Initial version."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "TSTpxm2GrjhR"
+ },
+ "source": [
+ "# **Outline** ⏰\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "VUbaQ5GqrvIl"
+ },
+ "source": [
+ "[Part 1: Background](#background)\n",
+ " 1. [Introduction to Reinforcement Learning](#introRL)\n",
+ " 1. [Application of Reinforcement Learning in buildings](#applicationRL)\n",
+ "\n",
+ "[Part 2: The OpenAI-Gym standard](#openAIGym)\n",
+ " 1. [What is OpenAI-Gym?](#whatIsOpenAIGym)\n",
+ " 1. [Example using an OpenAI-Gym environment](#exampleOpenAIGym)\n",
+ "\n",
+ "[Part 3: The Building Optimization Testing (BOPTEST) Framework](#boptest)\n",
+ " 1. [What is BOPTEST?](#whatIsBoptest)\n",
+ " 1. [Selecting a building test case](#selectBuilding)\n",
+ " 1. [Obtaining general information about the building](#obtainInfo)\n",
+ " 1. [Getting control input and measurement points](#gettingIOs)\n",
+ "\n",
+ "[Part 4: Implementing RL for a building with BOPTEST-Gym](#implementingRL)\n",
+ " 1. [What is BOPTEST-Gym?](#whatIsBoptestGym)\n",
+ " 1. [Starting up a BOPTEST-Gym environment](#startingUpBoptestGym)\n",
+ " 1. [Interacting with a BOPTEST-Gym environment](#interactingWithBoptestGym)\n",
+ " 1. [Developing a basic RL algorithm](#developingRlAlgo)\n",
+ " 1. [Testing our RL algorithm in BOPTEST-Gym](#testingRlAlgo)\n",
+ "\n",
+ "[Gearing up](#gearingUp)\n",
+ "\n",
+ "[Further resources](#furtherResources)\n",
+ "\n",
+ "[Feedback](#feedbackForm)\n",
+ "\n",
+ "[Annex I: Formal RL theory](#theoryRlFormal)\n",
+ "\n",
+ "[References](#references)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "oEzP9ZW4MXPv"
+ },
+ "source": [
+ "# **Part 1: Background** 📖 "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Fas232CyMX6_"
+ },
+ "source": [
+ "## **Introduction to Reinforcement Learning** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "TAy9fRjUTSdb"
+ },
+ "source": [
+ "Could you imagine a magic oracle able to decide on the best actions to optimize any process? Could you imagine this oracle not needing any prior information of the process but just learning from interacting with it? Powerful, right? Well, that is exactly what RL is meant for.\n",
+ "\n",
+ "Reinforcement Learning (RL) is one of the categories of machine learning, along with unsupervised learning and supervised learning. The main difference from the other categories is that RL learns from dynamic data, that is, data that are obtained while learning."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "e853vYumSx08"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "*Figure: The categories of machine learning. Source: [Mathworks](https://www.mathworks.com/discovery/reinforcement-learning.html)*\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "2_qdE6Ab4aE9"
+ },
+ "source": [
+ "In RL the goal is to learn the actions to be taken to achieve a predefined objective. RL relies on the principle of *repetitive experimentation*, that is, an approach where we roll out several **episodes of experience** where an agent 🤖 (the RL algorithm) interacts with its environment 🌎 (the process to be optimized) to learn based on a **reward** signal that is returned for every **action** taken from a specific **state** of the environment.\n",
+ "\n",
+ "Let's take the example of teaching a dog to grab a stick. In this case, the dog is the agent and all its surroundings conform the environment. Whenever the dog observes that there is a person throwing a stick it will perform an action. In case it grabs the stick and brings it back, the person will provide a cookie as a reward to encourage that behavior. In case the dog does not go for the stick but just runs around or goes chasing other dogs, the person will not provide the cookie. Eventually, the dog will associate the actions that bring the most rewards to specific observations and will be taking those actions accordingly.\n",
+ "\n",
+ "A more formal introduction to RL and its associated terminology can be found at the end of this tutorial, in [Annex I: Formal RL theory](#theoryRlFormal)."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "mMppgppKX4Fy"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "*Figure: RL notation when teaching a dog. Source: [Mathworks](https://www.mathworks.com/discovery/reinforcement-learning.html)*\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "pNy7uRzo8foI"
+ },
+ "source": [
+ "⚠️ **Important note:** ⚠️ It is common to find in the RL literature that the same term indistinctly designates the\n",
+ "state and the observation. This is not strictly correct for partially observable environments (most of the cases) where the observation only conveys part of the information that defines the state. For example, the state of the Tic-Tac-Toe game can be fully observed because there is a finite number of possibilities that define the state of a game. On the contrary, the thermal state of a building is only partially oservable. We can observe the indoor air temperature, but we cannot measure all temperatures from walls, ground, furniture... which also influence the building's thermal state."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "pNC0UnC2WYyE"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "What is particularly extraordinary of RL algorithms is that the same algorithm can be successfully used for a variety of tasks, from [robotic motion control](https://www.technologyreview.com/2021/04/08/1022176/boston-dynamics-cassie-robot-walk-reinforcement-learning-ai/) to [defeat the human world champion in the game of Go](\n",
+ "https://www.youtube.com/watch?v=WXuK6gekU1Y&ab_channel=DeepMind).\n",
+ "The latter is an astonishing achievement. It is true that the IBM supercomputer Deep Blue could previously [defeat Garry Casparov in chess](https://en.wikipedia.org/wiki/Deep_Blue_versus_Garry_Kasparov), but Go is to chess what chess is to Tic-Tac-Toe ([*Chris Wiltz*](https://www.designnews.com/design-hardware-software/googles-ai-beat-go-champion-mimicking-human-intuition)). And what is more important, professionals of Go state that this game has so many possible combinations that mastering it requires certain intuition and creativity, qualities that have only been attributed to humans so far... if AlphaGo defeated the best human player of Go, could machines resemble these qualities? Well, that is more a philosophical question. This tutorial is limited to investigate whether machines can efficiently control buildings, which you will see is already an enormous challenge!\n",
+ "\n",
+ "\n",
+ "\n",
+ " "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "XOcvaJpkNbho"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "*Figure: Netflix documentary that explains how AlphaGo, a RL algorithm developed by [DeepMind](https://www.deepmind.com/), could defeat Lee Sedol (4-1) and Fan Hui (5-0), the human world champions in the game of Go.*\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "o1jhKBOeOciu"
+ },
+ "source": [
+ "📌 **QUICK FACTS:**\n",
+ "- RL is a **category of machine learning** algorithms, together with supervised and unsupervised learning.\n",
+ "- Contrarily to other machine learning techniques, RL learns from **dynamic data**, that is, data that are obtained from interactions with the environment.\n",
+ "- Particularly, it learns from **state-action-reward** samples, so there is no need of domain knowledge to model the environment.\n",
+ "\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "6G1nWECgbmuW"
+ },
+ "source": [
+ "## **Application of Reinforcement Learning in buildings** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "xoTh8XvAM_OR"
+ },
+ "source": [
+ "During the last decade, there has been a clear interest growth in using optimal control for HVAC systems **[3]**. The figure below underlines this increased interest by showing the number of yearly peer-reviewed scientific publications related to optimal control in buildings.\n",
+ "RL algorithms have\n",
+ "gained particular popularity for their application in a **demand response** setting.\n",
+ "An extensive review for this application was written by Vázquez-Canteli et al. **[4]** This review is\n",
+ "not limited to HVAC systems but also demand response for charging electric\n",
+ "vehicles or thermal energy storage."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "7NywyXo6hD5n"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "*Figure: Evolution of the number of scientific publications about optimal control in buildings during the last decades. Data obtained from the Clarivate Web of Science.*"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "8KR-sTeJiJG2"
+ },
+ "source": [
+ "RL has already attracted the attention of the building control community for\n",
+ "many years. The figure below is obtained from the popular paper of Chen et al. **[5]** who graphically summarized the application of RL in buildings indicating the amount of data required by each research work to train the implemented RL algorithm.\n",
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "nCH0DBBbPEC7"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "\n",
+ "*Figure: Summary of the data required in the history of RL applications to buildings. Chen et al.* **[5]** ."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Ae18iXNKWV5I"
+ },
+ "source": [
+ "You can see from the figure that the feasibility and potential of applying RL for HVAC control\n",
+ "was first investigated by Liu and Henze back in 2006. Then, the interest was lost for a period, probably because Model Predictive Control (MPC) has been typically preferred for optimal control in buildings because it is much more data-efficient (it does not need as much data to be implemented). A comprehensive and complete review on the application of MPC for building energy management is provided by Drgona et al. **[6]**. \n",
+ "The reasons why RL is gaining momentum again are clear: \n",
+ "\n",
+ "- Evolution in deep learning\n",
+ "- We have much more data than before\n",
+ "- We have much more computational power than before\n",
+ "\n",
+ "In fact, there exist very recent developments for the application of RL in buildings, most of them using the OpenAI-Gym standard that is introduced in the next section. It is worth mentioning:\n",
+ "\n",
+ "- [CityLearn](https://github.com/intelligent-environments-lab/CityLearn) ➡️ Gym environment for providing demand response scenarios at an urban scale. That is, the goal of the RL agent is to flatten the energy demand of a district. It considers static\n",
+ "building heating and cooling load data and simplified models for energy storage.\n",
+ "- [Gym-Eplus](https://github.com/zhangzhizza/Gym-Eplus) ➡️ Gym environment wrapper around EnergyPlus simulation models.\n",
+ "- [Sinergym](https://github.com/ugr-sail/sinergym) ➡️ Extension of Gym-Eplus.\n",
+ "- [Energym](https://github.com/bsl546/energym) ➡️ Gym wrapper around building simulation models to assess controller performance.\n",
+ "- [Beobench](https://github.com/rdnfn/beobench) ➡️ A Toolkit for Unified Access to BuildingSimulations for Reinforcement Learning.\n",
+ "- 👉🏻[BOPTEST-Gym](https://github.com/ibpsa/project1-boptest-gym) ➡️ Gym environment for the BOPTEST Framework. The goal of the RL agent in this environment is to efficienty control an individual building. It allows testing against high-fidelity building models.\n",
+ "\n",
+ "The last of which is the focus of this tutorial.\n",
+ "\n",
+ "These RL frameworks for HVAC control bring hope\n",
+ "to the adoption of this technology in buildings. However, there is still a clear\n",
+ "need to different techniques and understand the best practices of RL\n",
+ "for this particular application. Let's investigate how!"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "7YnuNAQdM_L2"
+ },
+ "source": [
+ "# **Part 2: The OpenAI-Gym standard** 🤖 "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Sv728rc3M_Ir"
+ },
+ "source": [
+ "## **What is OpenAI-Gym and Gymnasium?** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "AhQfyrBCUigq"
+ },
+ "source": [
+ "OpenAI-Gym is a software library that gathers a **collection of tasks** called environments with a **unique Python interface** to control all of them. This unique interface is a key feature in the software package, and has given rise to a standard for the format in which RL agents are developed and treated, independently of\n",
+ "their internal functioning. The tasks defined in the Gym environments involve\n",
+ "a wide variety of fields like video games, classic control theory problems, or high dimensional robotic locomotive tasks.\n",
+ "[Gymnasium](https://gymnasium.farama.org/) is a maintained fork of the OpenAI-Gym library. You can find a list of available environment [here](https://gymnasium.farama.org/environments/classic_control/).\n",
+ "\n",
+ "\n",
+ "\n",
+ "The OpenAI-Gym philosophy heavily relies on the episodic aspect of RL, i.e.\n",
+ "the agent’s history is broken down into a series of experiences called **episodes** that may be of\n",
+ "variable length. The agent interacts with the environment until it reaches a\n",
+ "terminal state when the episode is finished. The goal is to maximize the total\n",
+ "cumulative reward per episode.\n",
+ "\n",
+ "The main methods of the OpenAI-Gym interface are the following:\n",
+ "\n",
+ "- `obs = env.reset()` ➡️ The `reset` method is the one called first to initialize the environment `env` (whatever it is). The environment returns the first observation `obs` (state).\n",
+ "- `next_obs,reward,terminated,truncated,info = env.step(action)` ➡️ The `step` method is used iteratively to interact with the environment. The RL agent computes an `action`, and the environment returns the next observation, associated reward, whether the episode is done (=terminated), and some other optional information."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "ntNSOBzJPJuF"
+ },
+ "source": [
+ "## **Example using an OpenAI-Gym environment** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "mQ2879zIPOJ0"
+ },
+ "source": [
+ "Now that we understand the main concepts of OpenAI-Gym we are going to illustrate its typical usage with a quick example. We're going to use the [CartPole environment](https://gymnasium.farama.org/environments/classic_control/cart_pole/), which is one of the classic control problems available in the OpenAI-Gym framework.\n",
+ "Let's start by installing the dependencies that we require:\n",
+ "\n",
+ ""
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "E0sfte45O8iN",
+ "outputId": "2bb75d65-ae6d-4559-f950-66d462bb3fd0"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Requirement already satisfied: gymnasium==0.28.1 in /usr/local/lib/python3.10/dist-packages (0.28.1)\n",
+ "Requirement already satisfied: numpy>=1.21.0 in /usr/local/lib/python3.10/dist-packages (from gymnasium==0.28.1) (1.25.0)\n",
+ "Requirement already satisfied: jax-jumpy>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from gymnasium==0.28.1) (1.0.0)\n",
+ "Requirement already satisfied: cloudpickle>=1.2.0 in /usr/local/lib/python3.10/dist-packages (from gymnasium==0.28.1) (2.2.1)\n",
+ "Requirement already satisfied: typing-extensions>=4.3.0 in /usr/local/lib/python3.10/dist-packages (from gymnasium==0.28.1) (4.7.1)\n",
+ "Requirement already satisfied: farama-notifications>=0.0.1 in /usr/local/lib/python3.10/dist-packages (from gymnasium==0.28.1) (0.0.4)\n"
+ ]
+ }
+ ],
+ "source": [
+ "!pip install gymnasium==0.28.1"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "RnmDsSAnM_F-"
+ },
+ "source": [
+ "**Cartpole environment description:**\n",
+ "\n",
+ "\"*A pole is attached by an un-actuated joint to a cart, which moves along a frictionless track. The pendulum is placed upright on the cart and the goal is to balance the pole by applying forces in the left (-1) and right (+1) direction on the cart. A reward of +1 is provided for every timestep that the pole remains upright.*\"\n",
+ "\n",
+ "\n",
+ "You can also check out the physics of the environment in the [GitHub repository of OpenAI-Gym](https://github.com/openai/gym).\n",
+ "See below an example of the evolution of an episode of the Cartpole environment. Note that most of the Gym envronments can be rendered to show how the RL agent is performing.\n",
+ "\n",
+ "\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "eMa2F3q_pV-C"
+ },
+ "source": [
+ "First, we are going to import `gym` and then `make` our Cartpole environment (version 1). Note how it is possible to instantiate the registered Gym environments by referring to their names with a string.\n",
+ "After that, we are going to interact with the environment for a maximum number of episodes of experience that we are going to indicate with `max_num_episodes`. In each episode, the environment is initialized with the `reset` method, and then we interact with the environment until the episode is `done`, which happens when the pole is down.\n",
+ "\n",
+ "It is important to note that in this example we are not implementing RL just yet. Instead, we are using the `sample` method from the action space of the environment to compute a random control action. This is useful when we want to quickly check how an environment behaves, but we should aim to replace that line by some intelligent RL agent able to compute a control action that optimizes the performance of the environment."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "LBxXhZc5nGb3",
+ "outputId": "85dc735f-6580-461e-d676-d13735bfa70d"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "\n",
+ " Episode #1 had 24 steps and total_reward=24.0\n",
+ "\n",
+ " Episode #2 had 13 steps and total_reward=13.0\n",
+ "\n",
+ " Episode #3 had 12 steps and total_reward=12.0\n",
+ "\n",
+ " Episode #4 had 27 steps and total_reward=27.0\n",
+ "\n",
+ " Episode #5 had 30 steps and total_reward=30.0\n",
+ "\n",
+ " Episode #6 had 38 steps and total_reward=38.0\n",
+ "\n",
+ " Episode #7 had 34 steps and total_reward=34.0\n",
+ "\n",
+ " Episode #8 had 22 steps and total_reward=22.0\n",
+ "\n",
+ " Episode #9 had 25 steps and total_reward=25.0\n",
+ "\n",
+ " Episode #10 had 23 steps and total_reward=23.0\n"
+ ]
+ }
+ ],
+ "source": [
+ "import gymnasium as gym\n",
+ "\n",
+ "env = gym.make('CartPole-v1')\n",
+ "max_num_episodes = 10 # maximum number of episodes\n",
+ "\n",
+ "for episode in range(max_num_episodes):\n",
+ " done = False\n",
+ " obs = env.reset()\n",
+ " total_reward = 0.0\n",
+ " step = 0\n",
+ " while not done:\n",
+ " action = env.action_space.sample() # Compute random action. This is to be replaced by a RL algo\n",
+ " obs,reward,terminated,truncated,info = env.step(action) # send the action to the environment\n",
+ " done = (terminated or truncated)\n",
+ " total_reward += reward\n",
+ " step += 1\n",
+ "\n",
+ " print('\\n Episode #{} had {} steps and total_reward={}'.format(episode+1,step,total_reward))\n",
+ "\n",
+ "env.close()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "ai9dHGWksZeu"
+ },
+ "source": [
+ "Notice how every episode lasts for a different number of steps because we are applying random forces to the cart. Also, notice how the total reward of each episode is equal to the number of steps, because the Cartpole environment gives a reward of +1 every timestep that we get to maintain the pole upright.\n",
+ "\n",
+ "\n",
+ "\n",
+ ""
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "tpeeOA8BM-5L"
+ },
+ "source": [
+ "# **Part 3: The Building Optimization Testing (BOPTEST) framework** 🏠 "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "z0ry1NQwuMXa"
+ },
+ "source": [
+ "Now that we understand how RL and OpenAI-Gym work, let's use that knowledge for the particular application of buildings.\n",
+ "In this tutorial we are going to connect with a BOPTEST building emulator that we will use as our environment to control through RL.\n",
+ "This emulator is a simulation model that was configured based on detailed physics and that has been peer-reviewed to ensure that it represents the behavior of an actual building as realistically as possible. Hence, although it is a simulation model, we are going to consider this emulator as the real building for control, testing and benchmarking.\n",
+ "\n",
+ "In this section we explain what BOPTEST is and how it can be generally used. Next section will exclusively focus on BOPTEST-Gym, the OpenAI-Gym interface of BOPTEST, to learn how we can use it to implement and assess RL algorithms for building control."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "OcPk7llkJP4m"
+ },
+ "source": [
+ "## **What is BOPTEST?** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "HtDrzTFuJU0e"
+ },
+ "source": [
+ "BOPTEST is a software framework enables the performance evaluation and benchmarking of advanced control algorithms for building HVAC control through simulations. The software is developed and is available on the BOPTEST GitHub respository at [https://github.com/ibpsa/project1-boptest](https://github.com/ibpsa/project1-boptest)\n",
+ "\n",
+ "and general information about BOPTEST can be found through the following link:\n",
+ "\n",
+ "[https://ibpsa.github.io/project1-boptest/](https://ibpsa.github.io/project1-boptest/)\n",
+ "\n",
+ "In the link below you can also find information about the overarching project that gave birth to BOPTEST, IBPSA Project 1:\n",
+ "\n",
+ "[https://ibpsa.github.io/project1/](https://ibpsa.github.io/project1/)\n",
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ ""
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "sNiHr2w0IFYI"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "*Figure: The BOPTEST concept.*\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Jj7kbbF8JXEG"
+ },
+ "source": [
+ "The main use case of the BOPTEST framework is the one where a control developer wants to evaluate the performance of his/her building control strategy. Testing in a real building may be very expensive, or just not possible. BOPTEST offers a menu of emulator building models so that the control developer can select one of them, interact in co-simulation, and eventually assess the performance of his/her controller with a set of Key Performance Indicators (KPIs) that are calculated by the BOPTEST framework. \n",
+ "\n",
+ "Note that using a standardized set of building emulators, testing scenarios, and KPIs enables benchmarking, allows to compare across different controllers, and throws light on what are the best practices for building control. In addition, making these emulators easily and rapidly available to use allows for control developers without expertise in building modeling to utilize them for controls testing and evaluation.\n",
+ "\n",
+ "In this section we are going to explain the basic BOPTEST functionality to connect to a building test case and obtain available control inputs and measurement points. For a more complete description on how to use BOPTEST please visit this [BOPTEST Colab tutorial](https://github.com/ibpsa/project1-boptest/blob/master/docs/workshops/BS21Workshop_20210831/Introduction_to_the_BOPTEST_framework.ipynb). "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "l5Fgw7eJHEjy"
+ },
+ "source": [
+ "## **Selecting a building test case** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "owb2Z2rqHEjz"
+ },
+ "source": [
+ "BOPTEST test cases are developed as [Functional Mock-up Units (FMU's)](https://fmi-standard.org/) and deployed within a containerized environment using the [Docker](https://www.docker.com/) software with:\n",
+ "\n",
+ "* A detailed emulator **building model**.\n",
+ "* Yearly **boundary condition data** for weather, schedules, pricing, and emission factors. These data are representative of the building location.\n",
+ "* An **API** that allows for, among other things, initializing a simulation or testing scenario, advancing a simulation with a control input, receiving forecast data, receiving emulator data, and receiving computed KPIs. The full API is described [here](https://github.com/ibpsa/project1-boptest/tree/boptest-service#test-case-restful-api).\n",
+ "\n",
+ "The basic workflow to test a controller is:\n",
+ "\n",
+ "1. Select a **test case** from the menu of those available.\n",
+ "2. Select one of the **testing scenarios** defined for the given test case. Testing scenarios are standardized for each emulator.\n",
+ "3. Set **parameters** for the interaction with your test controller, such as the control step or forecast horizon, if required. \n",
+ "4. Run the test case scenario in a **co-simulation** loop with your controller.\n",
+ "5. Request the KPIs and **evaluate** your controller's performance.\n",
+ "\n",
+ "We start by selecting and launching a BOPTEST building test case from the [repository of currently available test cases](https://ibpsa.github.io/project1-boptest/testcases/index.html). In this example, we are going to work with the test case called `bestest_hydronic_heat_pump`, which is a single-zone residential building with radiant floor heating and a heat pump. This is a high-fidelity, yet, relatively simple test case that allows us to focus on fundamental aspects. You may want to note the other test cases available in the repository as well as the fact that there are more under development. \n",
+ "\n",
+ "We can launch our chosen test case as follows. First, import the Python `requests` library so that we can make HTTP requests to the BOPTEST API at the address indicated by the `url`. Then, use the `POST /testcases//select` BOPTEST API endpoint to launch the test case and receive a corresponding `testid`. While the `url` is the common gateway for everyone to access the BOPTEST web-service, the `testid` is a unique identifier for you to address the test case that you have selected and launched."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "metadata": {
+ "id": "V_qU6ukZghTb"
+ },
+ "outputs": [],
+ "source": [
+ "import requests\n",
+ "\n",
+ "# url for the BOPTEST service\n",
+ "url = 'https://api.boptest.net'\n",
+ "\n",
+ "# Select test case and get identifier\n",
+ "testcase = 'bestest_hydronic_heat_pump'\n",
+ "\n",
+ "# Check if already started a test case and stop it if so before starting another\n",
+ "try:\n",
+ " requests.put('{0}/stop/{1}'.format(url, testid))\n",
+ "except:\n",
+ " pass\n",
+ "\n",
+ "# Select and start a new test case\n",
+ "testid = \\\n",
+ "requests.post('{0}/testcases/{1}/select'.format(url,testcase)).json()['testid']\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "eRZGKWDlHEj2"
+ },
+ "source": [
+ "Please do not get distracted by the `try-except` statement. We are using that one to stop already created test cases if we are revisiting this cell. This prevents from having several dangling test cases that can overwhelm our server.\n",
+ "\n",
+ "Once you have successfully obtained the `testid`, it is possible to start interacting with your selected test case using the rest of the BOPTEST API. You will need this `testid` for all further interactions with this test case. For example, use the `GET /name` BOPTEST API endpoint, along with your `testid`, to request the name of your test case and check that it matches the one we want."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "8mdK5JtNI-e_",
+ "outputId": "e7ab2747-b9e7-4678-b3e9-cf36ee9adc06"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "{'name': 'bestest_hydronic_heat_pump'}\n"
+ ]
+ }
+ ],
+ "source": [
+ "# Get test case name\n",
+ "name = requests.get('{0}/name/{1}'.format(url, testid)).json()['payload']\n",
+ "print(name)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "gUOQXYjlHEj3"
+ },
+ "source": [
+ "With our unique `testid` in-hand and having some practice using the BOPTEST API, we are ready to move on to start using our building emulator. For this tutorial, we are going to explain only how to obtain information about the building using the BOPTEST API before moving to learn BOPTEST-Gym.\n",
+ "Note that the test case will timeout after 15 minutes of no requests. If the test case times out, you can simply select and start a new one by repeating the steps described above.\n",
+ "\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "jmglIZGFHEj3"
+ },
+ "source": [
+ "## **Obtaining general information about the building** \n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "mJ6leLGvRJya"
+ },
+ "source": [
+ "The first thing we want to do is learn about the building and system that we want to control. All building information can be found under documentation provided for each specific test case on the [Test Cases tab](https://ibpsa.github.io/project1-boptest/testcases/index.html) of the BOPTEST website.\n",
+ "\n",
+ "The building information includes a description of the building envelope, the HVAC system design, the functioning of the baseline controller, available control inputs and measurement outputs, and available testing scenarios. Understanding how the system works is an important practice for control design, so take as much time as needed to understand the equipment, the points that can be measured, and the points that can be overwritten by your controller.\n",
+ "We briefly summarize the `bestest_hydronic_heat_pump` case here for completeness, but it is strongly recommended to have a deeper look into the [documentation](https://ibpsa.github.io/project1-boptest/testcases/ibpsa/testcases_ibpsa_bestest_hydronic_heat_pump/).\n",
+ "\n",
+ "The building represents a residential dwelling of 192 $m^2$ for a family of 5 members.\n",
+ "An air-to-water modulating heat pump of 15 $kW$ nominal heating capacity extracts energy from the ambient air to heat up the floor heating emission system, as shown in the figure below.\n",
+ "An evaporator fan blows ambient air through the heat pump evaporator when the heat pump is operating.\n",
+ "The floor heating system injects heat into the floor using water as the working fluid."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "SMQcNDl1HEj4"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "\n",
+ "*Figure: Schematic of HVAC system and control for the `bestest_hydronic_heat_pump` test case.*"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "lNLRXp2eHEj4"
+ },
+ "source": [
+ "A baseline controller is embedded in every test case emulator that is meant to be representative of a typical controller for that type of building. The baseline controller includes local loop control such that supervisory set points may be the focus of a test controller, although many of those local loop control signals are also available for overwriting if a user chooses. The baseline controller can also be considered an initial benchmark for control performance.\n",
+ "\n",
+ "In our selected test case, the baseline controller consists of a PI controller with the zone operative temperature as the controlled variable and the heat pump modulation signal for compressor frequency as the control variable, as depicted as C1 in the figure above and shown in the figure below.\n",
+ "The control variable is limited between 0 and 1, and it is computed to drive the zone operative temperature towards its set point, which is defined as a function of the occupancy schedule."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "RqVtoDgTHEj4"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "\n",
+ "*Figure: Primary PI controller C1.*\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "ryOicW_KHEj5"
+ },
+ "source": [
+ "All other equipment (fan for the heat pump evaporator circuit and floor heating emission system pump) are switched on when the heat pump is working (modulating signal higher than 0) and switched off otherwise. This is depicted in the figure of the HVAC schematic as controller C2."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "-J5j60bRHEj5"
+ },
+ "source": [
+ "## **Getting control input and measurement points** \n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "AGG7G6VeR4WB"
+ },
+ "source": [
+ "While control input and measurement points are described in the documentation, they are also available to retreive from the BOPTEST API. This is especially useful to store for later when requesting data for a specific point.\n",
+ "\n",
+ "Retrieve the control input and measurement outputs using the `GET /inputs` and `GET /measurements` BOPTEST API endpoints."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 5,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "0IKRxBykJY6u",
+ "outputId": "536ba021-d0f6-4d9f-f4a5-8eda9fc16541"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "TEST CASE INPUTS ---------------------------------------------\n",
+ "dict_keys(['oveTSet_activate', 'ovePum_activate', 'ovePum_u', 'oveHeaPumY_u', 'oveTSet_u', 'oveHeaPumY_activate', 'oveFan_activate', 'oveFan_u'])\n",
+ "TEST CASE MEASUREMENTS ---------------------------------------\n",
+ "dict_keys(['weaSta_reaWeaPAtm_y', 'reaPFan_y', 'reaQHeaPumCon_y', 'reaTRet_y', 'weaSta_reaWeaNOpa_y', 'weaSta_reaWeaTBlaSky_y', 'reaQHeaPumEva_y', 'weaSta_reaWeaNTot_y', 'weaSta_reaWeaSolAlt_y', 'reaTZon_y', 'weaSta_reaWeaHHorIR_y', 'weaSta_reaWeaLon_y', 'weaSta_reaWeaSolTim_y', 'weaSta_reaWeaCloTim_y', 'reaPPumEmi_y', 'weaSta_reaWeaHGloHor_y', 'weaSta_reaWeaHDifHor_y', 'weaSta_reaWeaRelHum_y', 'reaTSetHea_y', 'reaCO2RooAir_y', 'weaSta_reaWeaSolDec_y', 'reaPHeaPum_y', 'weaSta_reaWeaHDirNor_y', 'reaTSetCoo_y', 'weaSta_reaWeaWinDir_y', 'reaTSup_y', 'weaSta_reaWeaSolZen_y', 'reaQFloHea_y', 'reaCOP_y', 'weaSta_reaWeaTDryBul_y', 'weaSta_reaWeaTWetBul_y', 'weaSta_reaWeaTDewPoi_y', 'weaSta_reaWeaWinSpe_y', 'weaSta_reaWeaLat_y', 'weaSta_reaWeaCeiHei_y', 'weaSta_reaWeaSolHouAng_y'])\n"
+ ]
+ }
+ ],
+ "source": [
+ "# Get inputs available\n",
+ "inputs = requests.get('{0}/inputs/{1}'.format(url, testid)).json()['payload']\n",
+ "print('TEST CASE INPUTS ---------------------------------------------')\n",
+ "print(inputs.keys())\n",
+ "# Get measurements available\n",
+ "print('TEST CASE MEASUREMENTS ---------------------------------------')\n",
+ "measurements = requests.get('{0}/measurements/{1}'.format(url, testid)).json()['payload']\n",
+ "print(measurements.keys())"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "A4L6Gw6YJU5L"
+ },
+ "source": [
+ "The naming convention is such that the extension `_y` indicates a measurement point, `_u` indicates the value of an input which can be overwritten by a test controller, and `_activate` indicates the enabling (with value 0 or 1) of a test controller to overwrite the corresponding input value.\n",
+ "Hence, `_u` is enabled for overwriting by the test controller when `_activate=1`.\n",
+ "`weaSta_` indicates a measurement for a weather point, so that historical weather data can be easily retrieved.\n",
+ "\n",
+ "Notice that the jsons returned from the `GET /inputs` and `GET /measurements` BOPTEST API endpoints also include a description and unit of each variable, as well as the minimum and maximum value for inputs variables:"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "U7guJ_I10QOF"
+ },
+ "source": [
+ "Now let's stop the test case since we are not going to use it for a while. We do this to not overwhelm the server."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "v5_1Q_H80Z5k",
+ "outputId": "95577327-d7d9-4870-893b-7bdc1e164ac8"
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ ""
+ ]
+ },
+ "execution_count": 6,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "requests.put('{0}/stop/{1}'.format(url, testid))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "UHYHM9MjSz_C"
+ },
+ "source": [
+ "# **Part 4: Implementing RL for a building with BOPTEST-Gym** 🤖 🏠 "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "BEC76h9HT7gL"
+ },
+ "source": [
+ "## **What is BOPTEST-Gym?** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "z7A9k7GFUBsK"
+ },
+ "source": [
+ "BOPTEST-Gym is the OpenAI-Gym interface of BOPTEST that helps to train RL agents for the application of building climate control.\n",
+ "The BOPTEST-Gym interface accomodates the BOPTEST API to have BOPTEST building emulators as environments that follow the OpenAI-Gym standard.\n",
+ "Therefore, the BOPTEST-Gym interface facilitates the development of RL agents as it allows interacting with the BOPTEST building emulators with a standard that is very well known by the machine learning community. Or even better, it allows us to directly use existing RL agents that have been developed following this standard, like those from the [Stable Baselines 3](https://stable-baselines3.readthedocs.io/en/master/) repository.\n",
+ "\n",
+ "You can find more information about BOPTEST-Gym in [this paper](https://publications.ibpsa.org/conference/paper/?id=bs2021_30380), but here we summarize the main points you should know:\n",
+ "- BOPTEST-Gym enables the interaction of RL agents with a set of physics-based and highly **detailed building models** to assess RL for the application of building climate control.\n",
+ "- All **hyperparameters** of the environment are initialized when the environment is instantiated. A particularly relevant hyperparameter is `testcase`, a string specifying the BOPTEST emulator of choice. This string selects the building model from the [menu of BOPTEST building emulators](https://ibpsa.github.io/project1-boptest/testcases/index.html).\n",
+ "- The **state** of any building emulator environment can have a *time* component e.g. a weekly schedule, a *measurement* component with a subset (or all) measurements available in the building, and an *exogenous* component including disturbances of any kind of boundary condition data to the building such as electricity prices, ambient temperature, or temperature set-points.\n",
+ "- The **action** space is defined based on any subset (or all) inputs available to the emulator. These can be either building set-points, like zone\n",
+ "operative temperature set-points, or lower level actuator signals, such as heat\n",
+ "pump modulating signal or a pump stage.\n",
+ "- The **`reset()`** method is called at the beginning of every episode to return the environment to a logical initial state.\n",
+ "- The **`step()`** method is called every time step to take the action computed by the RL agent, overwrite the building inputs with the vector of action values and advance the building simulation model during one time step period. BOPTEST-Gym also has wrappers for discretization of the state and action spaces. This functionality comes in handy when training RL agents.\n",
+ "- A default **reward** function is implemented in the `compute_reward` method of the BOPTEST-Gym environment that can be overwritten. It is convenient to use the BOPTEST `/kpis` API to obtain the KPI values at the present time for defining custom reward functions."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "kvMgiRhLX2i8"
+ },
+ "source": [
+ "## **Starting up a BOPTEST-Gym environment** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "WcrOX1Z_UvTY"
+ },
+ "source": [
+ "BOPTEST-Gym uses RL algorithms from the [Stable Baselines 3](https://stable-baselines3.readthedocs.io/en/master/) repository to exemplify and test its functionality. Therefore, we need to install stable-baselines3.\n",
+ " "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "jpZk6qJKTuYl",
+ "outputId": "1246faeb-3e8e-47e9-cb19-dcb2756c921f"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Requirement already satisfied: stable-baselines3==2.0.0 in /usr/local/lib/python3.10/dist-packages (2.0.0)\n",
+ "Requirement already satisfied: numpy==1.25.0 in /usr/local/lib/python3.10/dist-packages (1.25.0)\n",
+ "Requirement already satisfied: gymnasium==0.28.1 in /usr/local/lib/python3.10/dist-packages (from stable-baselines3==2.0.0) (0.28.1)\n",
+ "Requirement already satisfied: torch>=1.11 in /usr/local/lib/python3.10/dist-packages (from stable-baselines3==2.0.0) (2.0.1+cu118)\n",
+ "Requirement already satisfied: cloudpickle in /usr/local/lib/python3.10/dist-packages (from stable-baselines3==2.0.0) (2.2.1)\n",
+ "Requirement already satisfied: pandas in /usr/local/lib/python3.10/dist-packages (from stable-baselines3==2.0.0) (1.5.3)\n",
+ "Requirement already satisfied: matplotlib in /usr/local/lib/python3.10/dist-packages (from stable-baselines3==2.0.0) (3.7.1)\n",
+ "Requirement already satisfied: jax-jumpy>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from gymnasium==0.28.1->stable-baselines3==2.0.0) (1.0.0)\n",
+ "Requirement already satisfied: typing-extensions>=4.3.0 in /usr/local/lib/python3.10/dist-packages (from gymnasium==0.28.1->stable-baselines3==2.0.0) (4.7.1)\n",
+ "Requirement already satisfied: farama-notifications>=0.0.1 in /usr/local/lib/python3.10/dist-packages (from gymnasium==0.28.1->stable-baselines3==2.0.0) (0.0.4)\n",
+ "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch>=1.11->stable-baselines3==2.0.0) (3.12.2)\n",
+ "Requirement already satisfied: sympy in /usr/local/lib/python3.10/dist-packages (from torch>=1.11->stable-baselines3==2.0.0) (1.11.1)\n",
+ "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch>=1.11->stable-baselines3==2.0.0) (3.1)\n",
+ "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch>=1.11->stable-baselines3==2.0.0) (3.1.2)\n",
+ "Requirement already satisfied: triton==2.0.0 in /usr/local/lib/python3.10/dist-packages (from torch>=1.11->stable-baselines3==2.0.0) (2.0.0)\n",
+ "Requirement already satisfied: cmake in /usr/local/lib/python3.10/dist-packages (from triton==2.0.0->torch>=1.11->stable-baselines3==2.0.0) (3.25.2)\n",
+ "Requirement already satisfied: lit in /usr/local/lib/python3.10/dist-packages (from triton==2.0.0->torch>=1.11->stable-baselines3==2.0.0) (16.0.6)\n",
+ "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->stable-baselines3==2.0.0) (1.1.0)\n",
+ "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib->stable-baselines3==2.0.0) (0.11.0)\n",
+ "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib->stable-baselines3==2.0.0) (4.40.0)\n",
+ "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->stable-baselines3==2.0.0) (1.4.4)\n",
+ "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib->stable-baselines3==2.0.0) (23.1)\n",
+ "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib->stable-baselines3==2.0.0) (8.4.0)\n",
+ "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->stable-baselines3==2.0.0) (3.1.0)\n",
+ "Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib->stable-baselines3==2.0.0) (2.8.2)\n",
+ "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas->stable-baselines3==2.0.0) (2022.7.1)\n",
+ "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.7->matplotlib->stable-baselines3==2.0.0) (1.16.0)\n",
+ "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch>=1.11->stable-baselines3==2.0.0) (2.1.3)\n",
+ "Requirement already satisfied: mpmath>=0.19 in /usr/local/lib/python3.10/dist-packages (from sympy->torch>=1.11->stable-baselines3==2.0.0) (1.3.0)\n"
+ ]
+ }
+ ],
+ "source": [
+ "!pip install stable-baselines3==2.0.0 numpy==1.25.0"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "4ljXzd7W4R0H"
+ },
+ "source": [
+ "Now that we have all package dependencies, let's clone the BOPTEST-Gym repository. We are going to clone the `boptest-gym-service` branch which works in the same way as the `master` branch but allows us to directly use the web-based version of BOPTEST that is readily available such that we do not have to deploy the building test case Docker containers locally.\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "2adMIlJu_ZZ8",
+ "outputId": "cba698ae-874e-427c-8284-393621d85a11"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Cloning into 'boptestGymService'...\n",
+ "remote: Enumerating objects: 3305, done.\u001b[K\n",
+ "remote: Counting objects: 100% (851/851), done.\u001b[K\n",
+ "remote: Compressing objects: 100% (427/427), done.\u001b[K\n",
+ "remote: Total 3305 (delta 456), reused 764 (delta 378), pack-reused 2454\u001b[K\n",
+ "Receiving objects: 100% (3305/3305), 47.56 MiB | 17.14 MiB/s, done.\n",
+ "Resolving deltas: 100% (1733/1733), done.\n"
+ ]
+ }
+ ],
+ "source": [
+ "try:\n",
+ " !rm -rf boptestGymService\n",
+ "except:\n",
+ " pass\n",
+ "!git clone -b boptest-gym-service https://github.com/ibpsa/project1-boptest-gym.git boptestGymService"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "cs9guwYo5w50"
+ },
+ "source": [
+ "Now we move our working directory to our recently cloned repository, import the `BoptestGymEnv` class, and instantiate our first BOPTEST-Gym environment!"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "mZsXUZIQ5iIj",
+ "outputId": "0f249174-2535-4b03-f372-09fc7b2e2ae0"
+ },
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "/usr/local/lib/python3.10/dist-packages/gymnasium/spaces/box.py:130: UserWarning: \u001b[33mWARN: Box bound precision lowered by casting to float32\u001b[0m\n",
+ " gym.logger.warn(f\"Box bound precision lowered by casting to {self.dtype}\")\n"
+ ]
+ }
+ ],
+ "source": [
+ "import sys\n",
+ "sys.path.insert(0,'boptestGymService')\n",
+ "from boptestGymEnv import BoptestGymEnv\n",
+ "\n",
+ "# Instantiate environment\n",
+ "env = BoptestGymEnv(url = url,\n",
+ " testcase = 'bestest_hydronic_heat_pump',\n",
+ " actions = ['oveHeaPumY_u'],\n",
+ " observations = {'reaTZon_y':(280.,310.)},\n",
+ " random_start_time = False,\n",
+ " start_time = 31*24*3600,\n",
+ " max_episode_length = 24*3600,\n",
+ " warmup_period = 24*3600,\n",
+ " step_period = 3600)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "8XVI61rnU4QZ"
+ },
+ "source": [
+ "You have connected to a BOPTEST building emulator and wrapped it around a Gym environment. Let's examine more in detail the arguments that you have used:\n",
+ "- `url`: the domain where your test case lives. In this case it is the url to BOPTEST-service, but it could be your localhost if you decide to spin a test case in your machine using Docker.\n",
+ "- `testcase`: The string identifier of the testcase.\n",
+ "- `actions`: List of strings indicating the action space.\n",
+ "- `observations`: Dictionary mapping observation keys to a tuple with the lower and upper bound of each observation. These bounds define the typical operational range for discretization and normalization purposes. Observation keys must belong either to the set of measurements or to the set of forecasting variables of the BOPTEST test case.\n",
+ "- `max_episode_lenght`: Maximum duration of each episode in seconds.\n",
+ "- `random_start_time`: Set to True if desired to use a random start time for each episode. That is typically usefull when training an RL agent to run several episodes with different boundary condition data. In our case, we set it to False and specify the start time of the episode.\n",
+ "- `start_time`: start time of the episode. It is specified in seconds from the beginning of the year. To be used in combination with `random_start_time=False`. \n",
+ "- `warmup_period`: Desired simulation period to initialize each episode, in seconds. In our case, we simulate the testcase for one day right before the beginning of the episode.\n",
+ "- `step_period`: The period of each control step, in seconds. In this case is set to one hour."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "-ZcNOH0SYEiR"
+ },
+ "source": [
+ "Now you can interact with the building emulator following the Gym standard. Everytime you use one of the methods of your environment, BOPTEST-Gym will send the associated commands through the BOPTEST API that you have learned above as to provide the desired functionality. A schematic of this process is shown in the figure below. This figure illustrates the typical steps that take place when training an agent and the mapping between the BOPTEST-Gym interface and the BOPTEST API. It is important to note that a state can be returned not only with current measurements, but also with boundary condition forecast or regressive values."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "rcnaIJvhYDa5"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "*Figure: Sequence diagram for training an agent withthe BOPTEST-Gym environment.*"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "AZVIz69qXyCZ"
+ },
+ "source": [
+ "## **Interacting with a BOPTEST-Gym environment** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "dA9_wVo8bMxr"
+ },
+ "source": [
+ "Let's see what we can do with our building Gym environment. Recall that the first step is using the `reset` method to simulate the building right before the episode start time a time period specified in `warmup_period`. This will bring the building to a reasonable initial state and the environment will return an observation `obs` which, in our case, it is comprised of only the zone operative temperature (`reaTZon_y`). This temperature is in Kelvins, so we convert it to degrees Celsius."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "Z4n5GsjXV08x",
+ "outputId": "d4cc4181-d8e8-418b-8f40-f137ab122002"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Zone temperature: 21.37 degC\n",
+ "Episode starting day: 31.0 (from beginning of the year)\n"
+ ]
+ }
+ ],
+ "source": [
+ "obs, _ = env.reset()\n",
+ "print('Zone temperature: {:.2f} degC'.format(obs[0]-273.15))\n",
+ "print('Episode starting day: {:.1f} (from beginning of the year)'.format(env.start_time/24/3600))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "IeJyBvLYqC11"
+ },
+ "source": [
+ "📌 **Note: About initialization**\n",
+ "\n",
+ "The initial state in the emulator consists of all states after simulation during the warmup period without any external input from an external controller. This particular emulator has 63 continuous time states comprising temperatures of walls, floor, roof, water, etc. During the warmup period, the baseline controller embedded in the emulator is used. After initialization the baseline controller will also work at any time unless some of the control variables are intentionally overwritten by an external controller."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "9i9VfDrdYJ0e"
+ },
+ "source": [
+ "We can inspect the observation and action space of any environment as follows:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "a_PC0YAEYR5U",
+ "outputId": "fd621a64-2550-4a61-f5c4-d38a28cd06b7"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Observation space of the building environment:\n",
+ "Box(280.0, 310.0, (1,), float32)\n",
+ "Action space of the building environment:\n",
+ "Box(0.0, 1.0, (1,), float32)\n"
+ ]
+ }
+ ],
+ "source": [
+ "print('Observation space of the building environment:')\n",
+ "print(env.observation_space)\n",
+ "print('Action space of the building environment:')\n",
+ "print(env.action_space)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "SBVCnncbePIQ"
+ },
+ "source": [
+ "So this environment has a Box (continuous and bounded) observation space which is the indoor building temperature. The operational range of this variable goes from $280$ $K$ to $310$ $K$. That is, from ~$7$ $°C$ to $37$ $°C$. On the other hand, the action space is a continuous variable that goes from $0$ to $1$. The latter variable represents the heat pump compressor frequency with $0$ meaning no heating, and $1$ meaning the heat pump working at full capacity. "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "pkx5Os6-Yltb"
+ },
+ "source": [
+ "But actually, the BOPTEST-Gym environment can be directly printed to show a lot of useful information to control the building:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "FGzL_ZskfoyO",
+ "outputId": "82b22b21-7480-4cc8-f2d4-fc9122d34332"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "\n",
+ "========================\n",
+ "BOPTEST CASE INFORMATION\n",
+ "========================\n",
+ "\n",
+ "Test case name\n",
+ "--------------\n",
+ "{'name': 'bestest_hydronic_heat_pump'}\n",
+ "\n",
+ "All measurement variables\n",
+ "-------------------------\n",
+ "{'reaCO2RooAir_y': {'Description': 'CO2 concentration in the zone',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'ppm'},\n",
+ " 'reaCOP_y': {'Description': 'Heat pump COP',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': '1'},\n",
+ " 'reaPFan_y': {'Description': 'Electrical power of the heat pump evaporator '\n",
+ " 'fan',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'W'},\n",
+ " 'reaPHeaPum_y': {'Description': 'Heat pump electrical power',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'W'},\n",
+ " 'reaPPumEmi_y': {'Description': 'Emission circuit pump electrical power',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'W'},\n",
+ " 'reaQFloHea_y': {'Description': 'Floor heating thermal power released to the '\n",
+ " 'zone',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'W'},\n",
+ " 'reaQHeaPumCon_y': {'Description': 'Heat pump thermal power exchanged in the '\n",
+ " 'condenser',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'W'},\n",
+ " 'reaQHeaPumEva_y': {'Description': 'Heat pump thermal power exchanged in the '\n",
+ " 'evaporator',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'W'},\n",
+ " 'reaTRet_y': {'Description': 'Return water temperature from radiant floor',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'K'},\n",
+ " 'reaTSetCoo_y': {'Description': 'Zone operative temperature setpoint for '\n",
+ " 'cooling',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'K'},\n",
+ " 'reaTSetHea_y': {'Description': 'Zone operative temperature setpoint for '\n",
+ " 'heating',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'K'},\n",
+ " 'reaTSup_y': {'Description': 'Supply water temperature to radiant floor',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'K'},\n",
+ " 'reaTZon_y': {'Description': 'Zone operative temperature',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'K'},\n",
+ " 'weaSta_reaWeaCeiHei_y': {'Description': 'Cloud cover ceiling height '\n",
+ " 'measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'm'},\n",
+ " 'weaSta_reaWeaCloTim_y': {'Description': 'Day number with units of seconds',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 's'},\n",
+ " 'weaSta_reaWeaHDifHor_y': {'Description': 'Horizontal diffuse solar radiation '\n",
+ " 'measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'W/m2'},\n",
+ " 'weaSta_reaWeaHDirNor_y': {'Description': 'Direct normal radiation '\n",
+ " 'measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'W/m2'},\n",
+ " 'weaSta_reaWeaHGloHor_y': {'Description': 'Global horizontal solar '\n",
+ " 'irradiation measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'W/m2'},\n",
+ " 'weaSta_reaWeaHHorIR_y': {'Description': 'Horizontal infrared irradiation '\n",
+ " 'measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'W/m2'},\n",
+ " 'weaSta_reaWeaLat_y': {'Description': 'Latitude of the location',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'rad'},\n",
+ " 'weaSta_reaWeaLon_y': {'Description': 'Longitude of the location',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'rad'},\n",
+ " 'weaSta_reaWeaNOpa_y': {'Description': 'Opaque sky cover measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': '1'},\n",
+ " 'weaSta_reaWeaNTot_y': {'Description': 'Sky cover measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': '1'},\n",
+ " 'weaSta_reaWeaPAtm_y': {'Description': 'Atmospheric pressure measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'Pa'},\n",
+ " 'weaSta_reaWeaRelHum_y': {'Description': 'Outside relative humidity '\n",
+ " 'measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': '1'},\n",
+ " 'weaSta_reaWeaSolAlt_y': {'Description': 'Solar altitude angle measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'rad'},\n",
+ " 'weaSta_reaWeaSolDec_y': {'Description': 'Solar declination angle measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'rad'},\n",
+ " 'weaSta_reaWeaSolHouAng_y': {'Description': 'Solar hour angle measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'rad'},\n",
+ " 'weaSta_reaWeaSolTim_y': {'Description': 'Solar time',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 's'},\n",
+ " 'weaSta_reaWeaSolZen_y': {'Description': 'Solar zenith angle measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'rad'},\n",
+ " 'weaSta_reaWeaTBlaSky_y': {'Description': 'Black-body sky temperature '\n",
+ " 'measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'K'},\n",
+ " 'weaSta_reaWeaTDewPoi_y': {'Description': 'Dew point temperature measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'K'},\n",
+ " 'weaSta_reaWeaTDryBul_y': {'Description': 'Outside drybulb temperature '\n",
+ " 'measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'K'},\n",
+ " 'weaSta_reaWeaTWetBul_y': {'Description': 'Wet bulb temperature measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'K'},\n",
+ " 'weaSta_reaWeaWinDir_y': {'Description': 'Wind direction measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'rad'},\n",
+ " 'weaSta_reaWeaWinSpe_y': {'Description': 'Wind speed measurement',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': 'm/s'}}\n",
+ "\n",
+ "All forecasting variables\n",
+ "-------------------------\n",
+ "['winDir',\n",
+ " 'TDewPoi',\n",
+ " 'LowerSetp[1]',\n",
+ " 'PriceElectricPowerConstant',\n",
+ " 'UpperSetp[1]',\n",
+ " 'PriceElectricPowerHighlyDynamic',\n",
+ " 'solTim',\n",
+ " 'solHouAng',\n",
+ " 'nOpa',\n",
+ " 'InternalGainsRad[1]',\n",
+ " 'nTot',\n",
+ " 'HGloHor',\n",
+ " 'winSpe',\n",
+ " 'TBlaSky',\n",
+ " 'solDec',\n",
+ " 'lon',\n",
+ " 'PriceElectricPowerDynamic',\n",
+ " 'HDifHor',\n",
+ " 'InternalGainsCon[1]',\n",
+ " 'solZen',\n",
+ " 'HHorIR',\n",
+ " 'relHum',\n",
+ " 'pAtm',\n",
+ " 'Occupancy[1]',\n",
+ " 'ceiHei',\n",
+ " 'lat',\n",
+ " 'InternalGainsLat[1]',\n",
+ " 'TWetBul',\n",
+ " 'TDryBul',\n",
+ " 'HDirNor',\n",
+ " 'EmissionsElectricPower',\n",
+ " 'cloTim',\n",
+ " 'solAlt',\n",
+ " 'UpperCO2[1]']\n",
+ "\n",
+ "All input variables\n",
+ "-------------------\n",
+ "{'oveFan_activate': {'Description': 'Activation for Integer signal to control '\n",
+ " 'the heat pump evaporator fan either on or '\n",
+ " 'off',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': None},\n",
+ " 'oveFan_u': {'Description': 'Integer signal to control the heat pump '\n",
+ " 'evaporator fan either on or off',\n",
+ " 'Maximum': 1,\n",
+ " 'Minimum': 0,\n",
+ " 'Unit': '1'},\n",
+ " 'oveHeaPumY_activate': {'Description': 'Activation for Heat pump modulating '\n",
+ " 'signal for compressor speed between 0 '\n",
+ " '(not working) and 1 (working at '\n",
+ " 'maximum capacity)',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': None},\n",
+ " 'oveHeaPumY_u': {'Description': 'Heat pump modulating signal for compressor '\n",
+ " 'speed between 0 (not working) and 1 (working '\n",
+ " 'at maximum capacity)',\n",
+ " 'Maximum': 1,\n",
+ " 'Minimum': 0,\n",
+ " 'Unit': '1'},\n",
+ " 'ovePum_activate': {'Description': 'Activation for Integer signal to control '\n",
+ " 'the emission circuit pump either on or '\n",
+ " 'off',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': None},\n",
+ " 'ovePum_u': {'Description': 'Integer signal to control the emission circuit '\n",
+ " 'pump either on or off',\n",
+ " 'Maximum': 1,\n",
+ " 'Minimum': 0,\n",
+ " 'Unit': '1'},\n",
+ " 'oveTSet_activate': {'Description': 'Activation for Zone operative '\n",
+ " 'temperature setpoint',\n",
+ " 'Maximum': None,\n",
+ " 'Minimum': None,\n",
+ " 'Unit': None},\n",
+ " 'oveTSet_u': {'Description': 'Zone operative temperature setpoint',\n",
+ " 'Maximum': 308.15,\n",
+ " 'Minimum': 278.15,\n",
+ " 'Unit': 'K'}}\n",
+ "\n",
+ "Default simulation step (seconds)\n",
+ "---------------------------------\n",
+ "3600\n",
+ "\n",
+ "Default scenario\n",
+ "----------------\n",
+ "{'electricity_price': 'constant'}\n",
+ "\n",
+ "Test case scenario\n",
+ "------------------\n",
+ "{'electricity_price': 'constant'}\n",
+ "\n",
+ "===========================\n",
+ "GYM ENVIRONMENT INFORMATION\n",
+ "===========================\n",
+ "\n",
+ "Observation space\n",
+ "-----------------\n",
+ "Box(280.0, 310.0, (1,), float32)\n",
+ "\n",
+ "Action space\n",
+ "------------\n",
+ "Box(0.0, 1.0, (1,), float32)\n",
+ "\n",
+ "Is a regressive environment\n",
+ "---------------------------\n",
+ "False\n",
+ "\n",
+ "Is a predictive environment\n",
+ "---------------------------\n",
+ "False\n",
+ "\n",
+ "Regressive period (seconds)\n",
+ "---------------------------\n",
+ "None\n",
+ "\n",
+ "Predictive period (seconds)\n",
+ "---------------------------\n",
+ "None\n",
+ "\n",
+ "Measurement variables used in observation space\n",
+ "-----------------------------------------------\n",
+ "['reaTZon_y']\n",
+ "\n",
+ "Predictive variables used in observation space\n",
+ "----------------------------------------------\n",
+ "[]\n",
+ "\n",
+ "Sampling time (seconds)\n",
+ "-----------------------\n",
+ "3600\n",
+ "\n",
+ "Random start time\n",
+ "-----------------\n",
+ "False\n",
+ "\n",
+ "Excluding periods (seconds from the beginning of the year)\n",
+ "----------------------------------------------------------\n",
+ "None\n",
+ "\n",
+ "Warmup period for each episode (seconds)\n",
+ "----------------------------------------\n",
+ "86400\n",
+ "\n",
+ "Maximum episode length (seconds)\n",
+ "--------------------------------\n",
+ "86400\n",
+ "\n",
+ "Environment reward function (source code)\n",
+ "-----------------------------------------\n",
+ "(' def get_reward(self):\\n'\n",
+ " \" '''\\n\"\n",
+ " \" Compute the reward of last state-action-state' tuple. The \\n\"\n",
+ " ' reward is implemented as the negated increase in the objective\\n'\n",
+ " ' integrand function. In turn, this objective integrand function \\n'\n",
+ " ' is calculated as the sum of the total operational cost plus\\n'\n",
+ " ' the weighted discomfort. \\n'\n",
+ " ' \\n'\n",
+ " ' Returns\\n'\n",
+ " ' -------\\n'\n",
+ " ' Reward: float\\n'\n",
+ " \" Reward of last state-action-state' tuple\\n\"\n",
+ " ' \\n'\n",
+ " ' Notes\\n'\n",
+ " ' -----\\n'\n",
+ " ' This method is just a default method to compute reward. It can be \\n'\n",
+ " ' overridden by defining a child from this class with\\n'\n",
+ " ' this same method name, i.e. `get_reward`. If a custom reward \\n'\n",
+ " ' is defined, it is strongly recommended to derive it using the KPIs\\n'\n",
+ " ' as returned from the BOPTEST framework, as it is done in this \\n'\n",
+ " ' default `get_reward` method. This ensures that all variables \\n'\n",
+ " ' that may contribute to any KPI are properly accounted and \\n'\n",
+ " ' integrated. \\n'\n",
+ " ' \\n'\n",
+ " \" '''\\n\"\n",
+ " ' \\n'\n",
+ " ' # Define a relative weight for the discomfort \\n'\n",
+ " ' w = 1\\n'\n",
+ " ' \\n'\n",
+ " ' # Compute BOPTEST core kpis\\n'\n",
+ " ' kpis = '\n",
+ " \"requests.get('{0}/kpi/{1}'.format(self.url,self.testid)).json()['payload']\\n\"\n",
+ " ' \\n'\n",
+ " ' # Calculate objective integrand function at this point\\n'\n",
+ " \" objective_integrand = kpis['cost_tot'] + w*kpis['tdis_tot']\\n\"\n",
+ " ' \\n'\n",
+ " ' # Compute reward\\n'\n",
+ " ' reward = -(objective_integrand - self.objective_integrand)\\n'\n",
+ " ' \\n'\n",
+ " ' self.objective_integrand = objective_integrand\\n'\n",
+ " ' \\n'\n",
+ " ' return reward\\n')\n",
+ "\n",
+ "Environment hierarchy\n",
+ "---------------------\n",
+ "(,\n",
+ " ,\n",
+ " ,\n",
+ " )\n",
+ "\n",
+ "\n"
+ ]
+ }
+ ],
+ "source": [
+ "print(env)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "-KYf1BksgfQj"
+ },
+ "source": [
+ "Note that this descriptive summary provides information not only about the Gym environment but also all information about the original BOPTEST test case. This may be useful, for example, if we want to extend our observation space or if we want to change our control action."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "zQ_y22lrg1cM"
+ },
+ "source": [
+ "BOPTEST-Gym comes along with other functionality that may be useful when training RL agents, like the capacity to discretize and normalize observation and action spaces. For instance, we are dealing now with continuous action environment meaning that the agent could decide to take any action between 0 and 1. However, it is probably helpful to the agent to decide on just whether the heating needs to be turned on (action=1) or off (action=0). For that, we can wrap our environment around a discretization wrapper with only one action bin (one bin has two extremes). The concept of wrappers is very powerful in Gym environments. With them, we are capable to customize observation, action, step function, etc. of an environment. No matter how many wrappers are applied, `env.unwrapped` always gives back the internal original environment object. Let's see how it works with BOPTEST-Gym:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 13,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "zIqfeNwgh9VK",
+ "outputId": "bb7087fd-19f3-4eee-828c-859b2ec641d0"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Action space of the wrapped agent:\n",
+ "Discrete(2)\n",
+ "Action space of the original agent:\n",
+ "Box(0.0, 1.0, (1,), float32)\n"
+ ]
+ }
+ ],
+ "source": [
+ "from boptestGymEnv import DiscretizedActionWrapper\n",
+ "env = DiscretizedActionWrapper(env,n_bins_act=1)\n",
+ "print('Action space of the wrapped agent:')\n",
+ "print(env.action_space)\n",
+ "print('Action space of the original agent:')\n",
+ "print(env.unwrapped.action_space)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Ghlx_zaf282q"
+ },
+ "source": [
+ "Another thing that we can do is to interact with the building environment for one episode of experience (one day). This is similar to what we did with the Cartpole example, but this time we are going to run just one episode and use a hysteresis controller that will turn on the heating the temperature is below a predefined temperature setpoint, and turn it off when the temperature goes above the setpoint. We first configure such controller:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 14,
+ "metadata": {
+ "id": "MrO0o7hNf5pB"
+ },
+ "outputs": [],
+ "source": [
+ "import numpy as np\n",
+ "np.set_printoptions(precision=3)\n",
+ "\n",
+ "class SimpleController(object):\n",
+ " '''Simple controller for this emulator.\n",
+ "\n",
+ " '''\n",
+ " def __init__(self, TSet=22+273.15):\n",
+ " self.TSet = TSet\n",
+ "\n",
+ " def predict(self, obs):\n",
+ " # Compute control\n",
+ " if obs[0]"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "JAIt_IfivAHN"
+ },
+ "source": [
+ "In this section we are going to develop a very simple RL agent based on the very well known *q-learning* algorithm. Although simple, this exercise will help us understand the main concepts of RL and how this machine learning technique can be helpful to mitigate climate change by enhancing building's operational efficiency. Recall that our objective is to develop an RL agent that can decide on the best action to take in each situation (each state) just from interactions with the environment (the building). Imagine we are at time $k$ in a certain state $\\pmb{s}$ and take an action $\\pmb{a}$. In return, we obtain a reward $r'$ the next time step and end up in a state $\\pmb{s}'$ :"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "nxO3gVx3vqrL"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ "*Figure: The backup diagram. Edited version from the book of Richard S. Sutton and Andrew G. Barto* **[6]**."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "XflFYx7lylyw"
+ },
+ "source": [
+ "In *q-learning* we aim to derive an *action-value function*, the q-function. The q-function indicates what is the **long-term** value of taking an action $a$ from a certain state $s$. With this information we not only have an estimation of the value of each state, but we can also decide to take the next action $\\pmb{a}'$ that leads to the highest value from the next state $\\pmb{s}'$. This principle relies on the so-called *Bellman optimality equation* that is presented below:\n",
+ "\n",
+ "\\begin{align}\n",
+ " q(\\pmb{s},\\pmb{a}) = r' + \\gamma \\max_{\\pmb{a}'} q(\\pmb{s}',\\pmb{a}')\n",
+ "\\end{align}\n",
+ "\n",
+ "This equation states that the total expected cummulative return of taking action $\\pmb{a}$ from state $\\pmb{s}$ equals the immediate reward $r'$ plus the maximum achievable reward that we can obtain from the following state $\\pmb{s}'$. Note that the q-function estimates the **TOTAL EXPECTED CUMULATIVE RETURN** of taking action $\\pmb{a}$ from state $\\pmb{s}$ (not just the immediate reward). So given the q-function we can know straight-away what is the best action to take for each state $\\pmb{s}$. You can imagine a q-function with one-dimensional state and action spaces as follows:"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "HBpa3qjuysK-"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ "*Figure: Example of how a q-function may look like for the case with one-dimensional state and action spaces. Note that, given the q-function, we can pick the action $a$ that leads to the highest expected cumulative reward $q_*$ from state $s$.*\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "db8AVf3GoCz9"
+ },
+ "source": [
+ "Powerful, right? now the question remains how to derive the q-function 😅.\n",
+ "\n",
+ "The q-function is inferred iteratively using the reward received by the agent each control step and bootstrapping with the Bellman optimality equation presented above. The sum of the immediate reward and the next-state q-function estimate is called the target. We use this target to recursively update the q-function at a learning rate $\\alpha$. The difference between the target and our current q-function estimate is called *temporal difference*. In summary, the q-learning method consists of recursively updating the q-function using the following formula:\n",
+ "\n",
+ "\\begin{align}\n",
+ " q(\\pmb{s},\\pmb{a}) = q(\\pmb{s},\\pmb{a}) + \\alpha [ \\underbrace{\\underbrace{r' + \\gamma \\max_{\\pmb{a}'} q(\\pmb{s}',\\pmb{a}')}_\\text{target} - q(\\pmb{s},\\pmb{a})}_\\text{temporal difference}]\n",
+ "\\end{align}\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "qjBqNfXd_pY2"
+ },
+ "source": [
+ "So in summary, the agent observes the reward once it has taken an action from a state. It has to explore the rewards from different state-action pairs and update its q-function as it goes.\n",
+ "\n",
+ "In our example we are going to use tabular state and action spaces to expedite learning and to easily store and visualize the q-function. Note, however, that we could use general function approximators like neural networks to configure the q-function.\n",
+ "\n",
+ "📌 **Note: The exploration-exploitation dilema** ⚖️\n",
+ "\n",
+ "RL always faces the so-called exploration-exploitation dilema. That is, how much of what we have learned we should exploit and how much we should explore to find even better solutions? In our case, we implement an *Epsilon-greedy* approach to balance exploration and exploitation of the RL agent. That is, the agent sometimes picks a random action (exploration), and sometimes picks an \"intelligent\" action (exploitation). The frequency at which the agent picks a random action is determined by *Epsilon* (`eps`) and it follows a linearly decaying schedule."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "ZQ7Um2UtLHk4"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "*Figure: The epsilon-greedy strategy for balancing exploration and exploitation.*\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "6aOHW96nLqOZ"
+ },
+ "source": [
+ "Our `Q_Learning_Agent` consists of only three methods:\n",
+ "\n",
+ "- `__init__` ➡️ The constructor.\n",
+ "- `predict` ➡️ Method to decide on an action given an observation.\n",
+ "- `learn` ➡️ Method for learning with the q-learning method explained above.\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 16,
+ "metadata": {
+ "id": "9U81QUVcUfoW"
+ },
+ "outputs": [],
+ "source": [
+ "class Q_Learning_Agent(object):\n",
+ "\n",
+ " def __init__(self, env, eps_min=0.01, eps_decay=0.01, alpha=0.05, gamma=0.9):\n",
+ " '''Constructor of a q-learning agent. Assumes discrete state and action spaces.\n",
+ "\n",
+ " '''\n",
+ " self.env = env\n",
+ " self.eps_min = eps_min\n",
+ " self.eps_decay = eps_decay\n",
+ " self.alpha = alpha\n",
+ " self.gamma = gamma\n",
+ "\n",
+ " # Initialize epsilon\n",
+ " self.eps = 1.0\n",
+ "\n",
+ " # Initialize q-function as a null function\n",
+ " self.q = np.zeros((env.observation_space.n,\n",
+ " env.action_space.n))\n",
+ "\n",
+ " def predict(self, obs, deterministic=True):\n",
+ " '''Method to select an action with an epsilon-greedy policy.\n",
+ "\n",
+ " '''\n",
+ " if deterministic:\n",
+ " # Use q-function to decide action\n",
+ " return np.argmax(self.q[obs])\n",
+ " else:\n",
+ " if self.eps > self.eps_min:\n",
+ " # Linearly decreasing schedule\n",
+ " self.eps -= self.eps_decay\n",
+ " if np.random.random() < self.eps:\n",
+ " # Explore with random action\n",
+ " return np.random.choice([a for a in range(env.action_space.n)])\n",
+ " else:\n",
+ " # Exploit the information of our q-function\n",
+ " return np.argmax(self.q[obs])\n",
+ "\n",
+ " def learn(self, total_episodes=10):\n",
+ " '''Learn from a number of interactions with the environment.\n",
+ "\n",
+ " '''\n",
+ " for i in range(total_episodes):\n",
+ " # Initialize enviornment\n",
+ " done = False\n",
+ " obs, _ = env.reset()\n",
+ " # Print episode number and starting day from beginning of the year:\n",
+ " print('-------------------------------------------------------------------')\n",
+ " print('Episode number: {0}, starting day: {1:.1f} ' \\\n",
+ " '(from beginning of the year)'.format(i+1, env.unwrapped.start_time/24/3600))\n",
+ "\n",
+ " while not done:\n",
+ " # Get action with epsilon-greedy policy and simulate\n",
+ " act = self.predict(obs, deterministic=False)\n",
+ " nxt_obs, rew, terminated, truncated, _ = env.step(act)\n",
+ " done = (terminated or truncated)\n",
+ " # Compute temporal difference target and error to udpate q-function\n",
+ " td_target = rew + self.gamma*np.max(self.q[nxt_obs])\n",
+ " td_error = td_target - self.q[obs][act]\n",
+ " self.q[obs][act] += self.alpha*td_error\n",
+ " # Make our next observation the current observation\n",
+ " obs = nxt_obs\n",
+ " # Print the q-function after every episode to show progress\n",
+ " print('q(s,a) = ')\n",
+ " print(self.q)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "RWPbW8WKaQET"
+ },
+ "source": [
+ "## **Testing our RL algorithm in BOPTEST-Gym** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "SV8bk8x75C_0"
+ },
+ "source": [
+ "Now that we have a RL agent ready, let's test it in BOPTEST-Gym! We are going to exploit the features of BOPTEST-Gym to:\n",
+ "\n",
+ "- Define a custom reward function of the enviornment.\n",
+ "- Instantiate the environment and define its state and action spaces.\n",
+ "- Train our RL agent.\n",
+ "\n",
+ ""
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "wy1TpSGEPxYr"
+ },
+ "source": [
+ "### Define a custom reward function of the environment"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "J9jwn5BQQCyj"
+ },
+ "source": [
+ "The definition of the reward function is **KEY**🗝 since it is what drives the learning of an agent.\n",
+ "The `BoptestGymEnv` Class allows to override its `get_reward` method that is called every control step as to freely define any reward function of choice.\n",
+ "\n",
+ "In our example, the goal is to implement a RL agent to identify the actions that keep comfort inside the building, and we should encode our reward function accordingly. We could implement this function by integrating the temperature deviations out of the comfort range. However, this approach is error-prone. We typically want to directly use signals from the environment to define the reward, preferrably those that are directly related to the function we want to optimize so that we make sure we strive for the ground truth optimum. In BOPTEST we use the `GET /kpis` API to obtain the so-called core KPIs at the present time, which are:\n",
+ "\n",
+ "\n",
+ "* **Thermal discomfort**: reported with units of [$K \\, h/zone$], defines the cumulative deviation of zone temperatures from upper and lower comfort limits that are predefined within the test case FMU for each zone, averaged over all zones. Air temperature is used for air-based systems and operative temperature is used for radiant systems.\n",
+ "* **Indoor Air Quality (IAQ) Discomfort**: reported with units of [$ppm \\, h/zone$], defines the extent that the CO$_2$ concentration levels in zones exceed bounds of the acceptable concentration level, which are predefined within the test case FMU for each zone, averaged over all zones.\n",
+ "* **Energy Use**: reported with units of [$kWh/m^2$], defines the HVAC energy usage.\n",
+ "* **Cost**: reported with units of [USD/$m^2$] or [EUR/$m^2$], defines the operational cost associated with the HVAC energy usage.\n",
+ "* **Emissions**: reported with units of [$kg \\, CO_2/m^2$], defines the CO$_2$ emissions from the HVAC energy usage.\n",
+ "* **Computational time ratio**: defines the average ratio between the controller computation time and the test simulation control step. The controller computation time is measured as the time between two emulator advances.\n",
+ "\n",
+ "The time series graph below shows how thermal discomfort and energy use are computed by the BOPTEST `GET /kpis` API call.\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "ZQagxsvtr3ow"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "*Figure: Integration of thermal discomfort (top) and energy use (bottom). In BOPTEST, the `GET /kpis` API can directly return these values every control step. Note that the integration step is significantly smaller than the control step.*"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "esAUAwHdr8y1"
+ },
+ "source": [
+ "The core KPIs are normally calculated at the end of the simulation to assess the controller performance, although they can be computed at any time. The warmup period is not taken into account for the calculation of the KPIs. See below how we define the `get_reward` method using the `GET /kpi`. Every control step we check whether there has been a discomfort increment. If there is not discomfort increment, we reward our agent with $1$, otherwise we return a $0$ (no reward). Clipping the reward is a good practice to accelerate learning."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 17,
+ "metadata": {
+ "id": "hTcc3XiVP-A6"
+ },
+ "outputs": [],
+ "source": [
+ "# Redefine reward function\n",
+ "class BoptestGymEnvCustomReward(BoptestGymEnv):\n",
+ " '''Define a custom reward for this building\n",
+ "\n",
+ " '''\n",
+ " def get_reward(self):\n",
+ " '''Custom reward function. To expedite learning, we use a clipped reward\n",
+ " function that has a value of 1 when there is no increase in discomfort\n",
+ " and 0 otherwise. We use the BOPTEST `GET /kpis` API call to compute the\n",
+ " total cummulative discomfort from the beginning of the episode. Note\n",
+ " that this is the true value that BOPTEST uses when evaluating\n",
+ " controllers.\n",
+ "\n",
+ " '''\n",
+ " # Compute BOPTEST core kpis\n",
+ " kpis = requests.get('{0}/kpi/{1}'.format(self.url, self.testid)).json()['payload']\n",
+ " # Calculate objective integrand function as the total discomfort\n",
+ " objective_integrand = kpis['tdis_tot']\n",
+ " # Give reward if there is not immediate increment in discomfort\n",
+ " if objective_integrand == self.objective_integrand:\n",
+ " reward=1\n",
+ " else:\n",
+ " reward=0\n",
+ " # Record current objective integrand for next evaluation\n",
+ " self.objective_integrand = objective_integrand\n",
+ " return reward"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "2hpd_svcOhDy"
+ },
+ "source": [
+ "### Instantiate the environment and define its state and action spaces"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "xszlVIQtOkiz"
+ },
+ "source": [
+ "Similarly to our `SimpleController` example, now we are going to use an agent that observes only the current indoor temperature and decides whether to turn heating on or off. However, instead of hard-coding such logic, we are going to use our very own implementation of the `Q_Learning_Agent` to see if it can learn how to do that.\n",
+ "For this, we are going to let our RL agent interact with the building for some episodes of experience.\n",
+ "Since we are now going to run several episodes for training, we want to stop our previous environment and start one that randomly initializes our building emulator throughout the year.\n",
+ "This allows to train our agent when using different boundary condition data in our building environment. We are also going to exclude the Spring, Summer, and Fall periods for training since we are only focused on learning the heating behavior.\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 18,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 35
+ },
+ "id": "24fsDMTv8tSF",
+ "outputId": "f65aaa3b-50d4-4232-98d9-75ef2f7c13a3"
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "env.stop()\n",
+ "import random\n",
+ "\n",
+ "# Seed for random starting times of episodes\n",
+ "seed = 123456\n",
+ "random.seed(seed)\n",
+ "# Seed for random exploration and epsilon-greedy schedule\n",
+ "np.random.seed(seed)\n",
+ "\n",
+ "# Winter period goes from December 21 (day 355) to March 20 (day 79)\n",
+ "excluding_periods = [(79*24*3600, 355*24*3600)]\n",
+ "# Temperature setpoints\n",
+ "lower_setp = 21 + 273.15\n",
+ "upper_setp = 24 + 273.15\n",
+ "# Instantiate environment\n",
+ "env = BoptestGymEnvCustomReward(url = url,\n",
+ " testcase = 'bestest_hydronic_heat_pump',\n",
+ " actions = ['oveHeaPumY_u'],\n",
+ " observations = {'reaTZon_y':(lower_setp,upper_setp)},\n",
+ " random_start_time = True,\n",
+ " excluding_periods = excluding_periods,\n",
+ " max_episode_length = 2*24*3600,\n",
+ " warmup_period = 24*3600,\n",
+ " step_period = 3600,\n",
+ " render_episodes = True)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "NU8aoMvV9AdE"
+ },
+ "source": [
+ "We have set the zone temperature as the only observation of the environment state. We have also set the lower and upper bounds of this variable to be $21$ and $24 °C$, respectively, which are the bounds of the comfort range during occupied periods. These bounds can be used by the environment for normalization or discretization purposes. In fact, we are going to discretize both the action and observation spaces to expedite learning. We decide to set only one bin for the action space (two possible actions: heating on or off). We split the observation space in three bins with the outer bounds of the comfort range as bins of the observation space (`outs_are_bins=True`). That is, the observation space is defined by $[-∞,21,24,+∞]$ as shown on the left hand side of the figure below. Note that only the middle bin is always comfortable whereas the other bins may lead to discomfort. If we had set `outs_are_bins=False` we would have had all our bins within the comfort range. The latter would give the agent a notion of what is the temperature within the comfort range (close to the lower bound, middle, or close to the upper bound), but it would raise an error if the temperature is out of the range. "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 19,
+ "metadata": {
+ "id": "uCUZKrOMOIEN"
+ },
+ "outputs": [],
+ "source": [
+ "from boptestGymEnv import DiscretizedObservationWrapper\n",
+ "env = DiscretizedActionWrapper(env, n_bins_act=1)\n",
+ "env = DiscretizedObservationWrapper(env, n_bins_obs=3, outs_are_bins=True)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Ab6WP3zLEvnb"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "*Figure: Possibilities for the discretization of the state space.*\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "GTvGxERwOOI6"
+ },
+ "source": [
+ "### Train our RL agent"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Sc3XqYSDOuGq"
+ },
+ "source": [
+ "The only missing step is to let our RL agent learn by rolling out episodes of experience with the environment. We use the previously defined `learn` method for this. Note that, since we set `render_episodes=True`, we will be seeing a plot with relevant variables after each episode is finished. This is helpful to check if the agent is learning as expected from early stages. If the agent is not showing any sign of life we can prematurely stop the learning process to use new learning settings while saving some valuable time and computational cost. "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 20,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 1000
+ },
+ "id": "jtOpX5y_RTsV",
+ "outputId": "dd84cf65-cc65-4f43-8a23-8c9fd1283c5e"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "-------------------------------------------------------------------\n",
+ "Episode number: 1, starting day: 11.4 (from beginning of the year)\n"
+ ]
+ },
+ {
+ "data": {
+ "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnUAAAHVCAYAAACXAw0nAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAACpKklEQVR4nOzdd3hT9ffA8Xe696KLQlv23huVvQQFEXGiMkT8KqCC+FNEQBAEUVARFRURF0NlOQFFpoAyLcieLdCW0b3b5P7+uDRSOpI0SZO05/U8eSC569w04+R8xtUoiqIghBBCCCEcmpOtAxBCCCGEEOaTpE4IIYQQohKQpE4IIYQQohKQpE4IIYQQohKQpE4IIYQQohKQpE4IIYQQohKQpE4IIYQQohKQpE4IIYQQohKQpE4IIYQQohJwsXUAVZ1Op+Py5cv4+vqi0WhsHY4QQghhNEVRSE9PJyIiAicnqRPZnGJHtm/frgwbNkzp1KmTcvHiRUVRFOXLL79UduzYYZN4PvzwQ6V58+aKr6+v4uvrq3Tq1En55Zdf9Muzs7OVZ555RgkKClK8vb2VIUOGKAkJCSYdIy4uTgHkJje5yU1ucnPYW1xcnKW/gkU52E2lbvXq1Tz22GMMGzaMgwcPkpubC0BqaipvvPEGv/zyS4XHVLNmTebOnUv9+vVRFIUvvviCe+65h4MHD9K0aVMmTJjAzz//zHfffYe/vz/jxo1jyJAh/Pnnn0Yfw9fXF4C4uDj8/PysdSpCCCGExaWlpREZGan/LhO2pVEURbF1EACtW7dmwoQJPP744/j6+vLPP/9Qp04dDh48SP/+/UlISLB1iAAEBQXx1ltvMXToUEJCQli+fDlDhw4F4Pjx4zRu3Jjdu3fTqVMno/aXlpaGv78/qampktQJIUQVlZ2djVarrZBjOTk54eXlZZF9yXeYfbGbSt2JEyfo2rVrscf9/f1JSUmp+IBuodVq+e6778jMzKRz587s37+f/Px8evfurV+nUaNGREVFlZnU5ebm6quQoL4hhBBCVF3nzp3jyy+/rLDj1ahRg9GjR1fY8UTFsZukLjw8nNOnT1OrVq0ij+/cuZM6derYJijg8OHDdO7cmZycHHx8fFi7di1NmjTh0KFDuLm5ERAQUGT9sLCwMquKc+bMYcaMGVaOWgghhKPIzMwEYOjQobi6ulr9eB4eHlY/hrANu0nqnnzySZ577jmWLl2KRqPh8uXL7N69m0mTJjF16lSbxdWwYUMOHTpEamoq33//PcOHD2fbtm3l3t/kyZOZOHGi/n5hfwQhhBBVW/369XFzc7N1GMKB2U1S9/LLL6PT6ejVqxdZWVl07doVd3d3Jk2axPjx420Wl5ubG/Xq1QOgbdu27N27l/fee48HH3yQvLw8UlJSilTrEhMTCQ8PL3V/7u7uuLu7WztsIYQQDsJOuraLSsAuJpXRarXs2LGDsWPHkpSUxJEjR9izZw9Xr17l9ddft3V4Reh0OnJzc2nbti2urq5s3rxZv+zEiRPExsbSuXNnG0YohBDCkRQmdTJXqTCXVSp1OTk5JrXZOzs707dvX44dO0ZAQABNmjSxRlgmmzx5Mv379ycqKor09HSWL1/O1q1b2bhxI/7+/jzxxBNMnDiRoKAg/Pz8GD9+PJ07dzZ65KsQQgghhKVYrFKn0+l4/fXXqVGjBj4+Ppw9exaAqVOn8tlnnxncvlmzZvpt7MWVK1d4/PHHadiwIb169WLv3r1s3LiRPn36APDOO+9w9913c99999G1a1fCw8NZs2aNjaMWQgjhSKRSJyzFYkndrFmzWLZsGfPmzSvS0bNZs2YsWbLEqO0nTZrETz/9RHx8PGlpaUVutvDZZ59x/vx5cnNzuXLlCr///rs+oQN1BNEHH3xAUlISmZmZrFmzpsz+dEIIIURpJKkT5rJY8+uXX37JJ598Qq9evfjf//6nf7xly5YcP37c4PYDBgwAYNCgQUVe2IqioNFoKmxSRiGEEKIiyUAJYSkWS+ouXbqkHyV6M51OR35+vsHtt2zZYqlQhBBCCIchza/CUiyW1DVp0oQdO3YQHR1d5PHvv/+e1q1bG9y+W7dulgpFCCGEEKLKsVhSN23aNIYPH86lS5fQ6XSsWbOGEydO8OWXX/LTTz8Z3H779u1lLi/pEmJCCCGEo5NKnbAUiyV199xzDz/++CMzZ87E29ubadOm0aZNG3788ccigwtK071792KP3fwClz51QgghhBCls+g8dV26dOG3334r17bJyclF7ufn53Pw4EGmTp3K7NmzLRGeEEIIYXekUicsxWJJ3ejRo3n00UdLrLgZw9/fv9hjffr0wc3NjYkTJ7J//34zIxRCCCHsj4x+FZZisXnqrl69yp133klkZCQvvvgihw4dssh+w8LCOHHihEX2JYQQQtgjqdIJS7BYpW79+vUkJyfz3XffsXz5chYsWECjRo0YNmwYjzzyCLVq1Spz+5iYmCL3FUUhPj6euXPn0qpVK0uFKYQQQtgVqdQJS7Fon7rAwEDGjBnDmDFjuHjxIitWrGDp0qVMmzaNgoKCMrdt1aoVGo2m2Iu7U6dOLF261JJhCiGEEHZFKnXCEiya1BXKz89n3759/PXXX5w/f56wsDCD25w7d67IfScnJ0JCQvDw8LBGiEIIIYRdKLxykhDmslifOlCvCvHkk08SFhbGiBEj8PPz46effuLixYsGt922bRvh4eFER0cTHR1NZGQkHh4e5OXl8eWXX1oyTCGEEMJuSPOrsBSLJXU1atRgwIABXLt2jU8++YTExESWLl1Kr169jPoFMnLkSFJTU4s9np6ezsiRIy0VphBCCGF3pFInLMFiSd1rr71GfHw8a9euZejQobi7u5u0fWnl54sXL5Y43UlFmDNnDu3bt8fX15fQ0FAGDx5cbCRu9+7d0Wg0RW7/+9//bBKvEEIIxyOVOmEpFutT9+STT5Zru9atW+uToV69euHi8l9IWq2Wc+fOceedd1oqTJNs27aNsWPH0r59ewoKCnjllVfo27cvR48exdvbW7/ek08+ycyZM/X3vby8bBGuEEIIByWVOmEJZiV1Q4YMYdmyZfj5+TFkyJAy112zZk2Jjw8ePBiAQ4cO0a9fP3x8fPTL3NzcqFWrFvfdd585YZbbhg0bitxftmwZoaGh7N+/v8i1aL28vAgPD6/o8IQQQlQCMlBCWIpZSZ2/v7/+hejn51euF+X06dMBqFWrFg8++KBFRrvm5+eTkJBAVlYWISEhBAUFmb1PQN/n79b9ffPNN3z99deEh4czcOBApk6dWmq1Ljc3l9zcXP39tLQ0i8QmhBDCMUnzq7AUs5K6zz//XP//ZcuWmRXI8OHDzdo+PT2dr7/+mpUrV/L333+Tl5en//VTs2ZN+vbty5gxY2jfvn259q/T6Xj++ee5/fbbadasmf7xRx55hOjoaCIiIoiJieGll17ixIkTpVYm58yZw4wZM8oVgxBCiMpJKnXCEjSKhX4i9OzZkzVr1hAQEFDk8bS0NAYPHswff/xR5vZarZZ33nmHb7/9ltjYWPLy8oosT0pKKnXbBQsWMHv2bOrWrcvAgQPp0KEDEREReHp6kpSUxJEjR9ixYwfr1q2jY8eOvP/++9SvX9+k83v66af59ddf2blzJzVr1ix1vT/++INevXpx+vRp6tatW2x5SZW6yMhIUlNT8fPzMykmIYQQFUhRIHE/XP4TMhNA0YFnMATUhZCW4F8bNKaPP/zzzz/ZsWMHL7/8shWCtq60tDT8/f3lO8xOWGygxNatW4slYgA5OTns2LHD4PYzZsxgyZIlvPDCC7z66qtMmTKF8+fPs27dOqZNm1bmtnv37mX79u00bdq0xOUdOnRg1KhRLF68mM8//5wdO3aYlNSNGzeOn376ie3bt5eZ0AF07NgRoNSkzt3d3eSRwUKUW0EO5KWDiwe4+do6GiHKpihw7Qikx4JOC14h4FcLvMPBlpWs7CQ4+iX8sxiST4CzO/jUUGPKugp5N7rRuPtDja4Q1QMie0JIc6OTPKnUCUswO6m7+ZqtR48eJSEhQX9fq9WyYcMGatSoYXA/33zzDZ9++il33XUXr732Gg8//DB169alRYsW7Nmzh2effbbUbVesWGFUrO7u7iZNN6IoCuPHj2ft2rVs3bqV2rVrG9zm0KFDAFSvXt3o41iKoijodLoKP66wE9o8tYqQuB+uHFRv14+BolWXewZDcDOo2RUiu0FYB3Ax8geGNh+cXGz7xSoqp8JE7uT3cOo7SD5dfB3PEAhtCRG3Q2RX01675sSVuBf++RhOfqcmmfUHQ4/3ocYd4Oz237qZCXAtBuL/hovbYfsU0Oaq77nI7hDVS7351yrxUDqdTk3qkk+r+7l645YeC+6B4BcNoa0gtDWEtbP+uQuHZXbzq5OTk/4XRkm78vT05P3332fUqFFl7sfb25tjx44RFRVF9erV+fnnn2nTpg1nz56ldevWJU5MbG3PPPMMy5cvZ/369TRs2FD/uL+/P56enpw5c4bly5czYMAAqlWrRkxMDBMmTKBmzZps27bNqGNYsnR97tw5ufqGEEI4IB9NBi94v63e8Qy50ZxbC3JTIOWMmvjq8tWqe8RtULM7RPaA6h2KJpjG0BWoP9IsQJpf7YvZf9Vz586hKAp16tTh77//JiQkRL/Mzc2N0NBQnJ2dDe6nZs2axMfHExUVRd26ddm0aRNt2rRh7969RjVX/vHHH4wbN449e/YUe2GlpqZy2223sXjxYrp06WL0uX300UeAOsHwzT7//HNGjBiBm5sbv//+O++++y6ZmZlERkZy33338eqrrxp9DEsKDg7WTxEjKiFtPqSehaTjkHQMUk6rj7l4QFBDCGwE1RqBbzQ4lfGe0+nUCkDSMbh+FJJOqFUFUJuV3HzU6oB7AHgEqP9381ObmDIuquvn3Ojj6l8bqjVVK4AB9cDZ1drPgnBkGZch4W+1opVxSX3thrWF8I4Q3NRwoqHTQXocJB+D68ch+TjkZ6uvu4AGUK0xBDVWX5dlvQdKkn0N4rZC3BbIy1Bf09F9IKRFufrJFZGfpb5vr/8L145C5mX1cWd38AoFv0iqhUVDg43q8bxLmCJLmwfXDkPcNjXO/fNh1zRw8VSTvMjuaqJXUpKnzYfEfRD7B5z9We2KMXSjeeck7JLFBkqY6+WXX8bPz49XXnmFVatW8eijj1KrVi1iY2OZMGECc+fOLXP7QYMG0aNHDyZMmFDi8oULF7JlyxbWrl1rjfDLTX7liFIV5ED8X3Dxxod4/G71scJ+O5Hd1A/ykFamf4HdTFHUCoCiU79kjZF6Xv2CuPAbxP6ufiG6eKpNu9F91Ftwc2murWwUBXJT1X6arl7gEVh2wqMocPUfOL0eTq1WkxI3X6h7DzR8AKL7mteUqNPC1UMQu0VNxi7tuBGbN4S3h/AO/918axZ/PWZdgzM/wLGv1PeYmy80HQGtxqo/lKwlL0Otlrn7l/89otOqXSwKPx8ubld/eBUmeYENoSALkk/ClUPq/1191Pdm/XuhyWMWORX5DrMvFk/qjh49WuLo1UGDBpm0nz179rBr1y7q16/PwIEDDa4fHR3Nhg0baNy4cYnLjx8/Tt++fYmNjTUpDmur1G+IzAT1y1+bC97VIbCe+b94K7OCHIjfo/4Sv7gVLu9Wnzv3gBv94LpDzW5qs4w5SZylKTq1/8+F3+DC73Bpu3ouXqEQ1Ruie6tfJL5lDzISdkqbD3F/wKk1agKU+V+/aVw8IbA+BNRXK7WB9dWqbsZFSNgHl3aqVWE3P6g9ABo+CLXvNP7Hg6l0BWqf0rhtakUw4W+1sgfqZ1B4hxtNmmlq37XEA+qyqJ7Q5HGoP0StVDsinVZNoOO2qglu+kU1YQ6oB8Et1B+BoW0sXk2v1N9hDshiSd3Zs2e59957OXz4MBqNRt+/rrC/nVarLXXb/Px8nnrqKaZOnWrUYISSeHh4cOTIEerVq1fi8tOnT9O8eXOys7PLtX9rqTRviNy0/5rzLu9Wfz0mnyy6jlcYNHoIWj5t3V/BjiI/+0YSt1V9vuL3qEmcR+CNSlx39Rbc3L6SOEMKcuDyrhtJ3m83vjgVCGqkJnnVO6pNboENHOu8KgOdVk1ycpLV+x6BalPfrUlWXrqakJ1eBydXQ8518K8D9e5VK2Du/pCfCWkXIOUUJN+4pccBitqsGNoKqndWk7nIbqb3+7KUjPj/Erz4vyEzXq0yBjVSR6jW6gc+FT+wrbKoNN9hlYTFkrqBAwfi7OzMkiVLqF27Nn///TfXr1/nhRde4O233zbYl83f359Dhw6VO6mrW7cu8+fPL7VP2Zo1a5g0aRJnz54t1/6txeHeEIqifnhf/lOtzlw/qt4yLv63TrUmakWpZjf1/85u6of/+U1w9Au1qa7WndD2ebX5xdGa6FLOqP1Szv6s/trPS1crET4R6i/i0FbqLaQVuN/4myo3+gIlHlQTnsu71JF12jzwCLqpEtfdpGkQHEL2dbWpNvZ39d+UG6MbXb1vjOZrq1YQQlqoX7SWruLkpqnJycXtkHZeTTo1GnVkolco+EapCWZgA7XS5GrDazennLnRZ/G6er+wz5VXGPhGlq+KlJsKp9aqlbbY36GghB+2HoHqMTyD1Sk6Uk6ro6b9aqnVtYYPqq9pQ+/Vghx1/+4Bjve+FuXicN9hlZzFkrrg4GD++OMPWrRogb+/P3///TcNGzbkjz/+4IUXXuDgwYNlbj98+HBatWpVap84Q8aPH8/WrVvZu3dvsUuNZWdn06FDB3r06MHChQvLtX9rcZg3ROo5OL4Cjn2jJnFo1ObUoCZq4lZ4C2qoflmXpiAXTn4L+9+FKwfUTs2Nh0H9+9RO/vZImwcXd8C5G4lc8kk1Ua3ZTe274u6vJg7psWrzx7XD6jagJgwA2VfULzxQ57eKuB1q3H6jEtesciVxhuQkq32BCqdfSdz/X6KncVKb8oKb/Xer1kx9rZkyWk+bB+d+hWPL4ewP6nMf2lqdJNbVW61Y5VyHzET1B0dhEgVq8hTUSO2TFNToxiCUhiX3ybKE7CQ4vhz+XaY+F2XxDlefn8AbzZ0B9dRz8q+jDmoplH4JYjerzaVnf1Kfjxq3Q52BavLsWQ3QqANeMuPValZWovqDyzP4vx9mgfUlORNlcpjvsCrCYkldYGAgBw4coHbt2tStW5clS5bQo0cPzpw5Q/PmzcnKyipz+1mzZjF//nx69epF27Zt8fYumhiUNU8dQGJiIm3atMHZ2Zlx48bppyA5fvw4H3zwAVqtlgMHDhAWFmbeiVqYXb8hsq7BiVXqF87lXeDiBfUGq02oNbuqyUx5KYraqfnQR+qXTn7Gf81zUT3VZMcj0FJnYrqsK2pScPYnOL9Rrcb5REDtu9TmpOjepVdNtPlqU/SVQ+q/Gmd1EtXCZMUvqkJPxSHkpsK1f+H6EXXqhsJb9lV1ubOb+gOgMMkLbqYmHn5R/yV7+dlqk+Gp1eq8YjlJagLTaBg0ehj8Iks/fvZ1NVlPPnljdPEJdZLZ5FPqIBJQk8HABmqC519bPbZv1H//upvw/tUVqE3TRz6HM+vVJLPOXdB0OIS1V6tzGo1a9cpMVBOutAtqPCmn1CQ4+ZQ63UUhjyA1jqxENVEDdU6zRg+rlTZfw/OFCmEqu/4Oq4IsltR16dKFF154gcGDB/PII4+QnJzMq6++yieffML+/fs5cuRImduX1eyq0WiMaja9cOECTz/9NBs3bizSp69fv3588MEH5W7atSa7fENcO6JW0o59rTbB1OoHjR6BeveUXYUrr/xsuLBJrYLFblan7dA4qZWVGneot4jbje/3kpuq9uuL36N26tbmqVUMz1C16hLUWK1uFHYYvrlp9NIOtX9bYQfq6h2gzt3qLaSlVC0qWtYVNdm7dqRowlc4g7/GWa2gOXtA2jn1b+0bBY0fUSvAwc3K3r8hugJ1sE/ScTXJK0z40mPVjujKTX2F3f2LJnk3/98vUk32rx1Rm0BPfKsmrMHNoOlINVZvE39wKoqauKaeVZttU86ocXmFqf0wI7urPyaEsCK7/A6rwiyW1G3cuJHMzEyGDBnC6dOnufvuuzl58iTVqlVj1apV9OzZ0xKHMUpycjKnT59GURTq169PYKANKz4G2M0bQtHBuQ2w/x31S8cnAlqNg+ajK/6LoXC6jIvb1MpL6o2E3r/OjSTvdvXfoEZq3Knn1P59l3aoo96u/gMo4FFNnYnd2U2taGRcvikZcALvCPXf7GvqcH9Q16/ZVe1AXbu/6V+0wvoURU2oko6qr5XCfnIB9dREplqTikm+dQXqj4a0WDWZuvXf9Nj/BiTczDcKGgxVE7nQ1vJDQTg0u/kOE4CV56lLSkoiMDDQpGva5eXlce7cOerWrYuLi8UuTWu3bP6GyM+Eo1+plbnkE2pzTdsJ0OB++5lINiNeHZhxaad6u3JIrZA4uahfrIX8aqmj7Gp0VZO+W/sDKYr6JVw4cW9arLrcI0hNEKs1K7uJTghT5aWrVeD0OHByg4A66g8HISoJm3+HiSIsktTl5+fj6enJoUOHaNasfM0dWVlZjB8/ni+++AKAkydPUqdOHcaPH0+NGjV4+eWXzQ3TLtnsDZEWB/98CDEfq82V9e6FNs+rVTB7rxzkZaiT8iYdV0cH+tdSqzM+EbaOTAghqhRJ6uyLRUphrq6uREVFlTkXnSGTJ0/mn3/+YevWrdx55536x3v37s1rr71WaZM6q1N06ui67Ktqn5vE/Wq1K3az2tG/+WhoPV7t+O0o3Hwgupd6E0IIIQRgoaQOYMqUKbzyyit89dVXBAUFmbz9unXrWLVqFZ06dSrSXNu0aVPOnDlT5rb5+fn6C9k/9thjuLnZaJJLW7t6GP6arXYuL7zlXFcTu0IeQWoTa59PoNGD6mVxhBBCCOHwLJbULVq0iNOnTxMREUF0dHSxKUkOHDhQ5vZXr14lNDS02OOZmZkG++RNmjSJPn36oCgKL774Iu+9957pJ1AJpKSmExPrCa5NwLU9+PtAsJc6YtXFSx2dVzgpaAqwp+y5A4UQQlSMkJCQUi9zKYSxLJbUlXYlB2O1a9eOn3/+mfHjxwP/XV5syZIldO7cucxtdTodOp0OrVaLTqcrc93KLM2tJnuzm0GxCeMzb9yuVnxQQgghypSXl4eLi4skdcJsVh39aoqdO3fSv39/Hn30UZYtW8ZTTz3F0aNH2bVrF9u2baNt27albpuXl8fy5ctRFIVHHnkEd3d3i8Q0Z84c1qxZw/Hjx/H09OS2227jzTff1E9sDJCTk8MLL7zAypUryc3NpV+/fnz44YdGT3IsnUyFEKJq27lzJ7t27eL//u//bB2KyeQ7zL5Y9NpEKSkpLFmyhMmTJ5OUlASoza6XLl0yuO0dd9zBoUOHKCgooHnz5mzatInQ0FB2795dZkIH4ObmxogRIxg5cqTFEjqAbdu2MXbsWPbs2cNvv/1Gfn4+ffv2JTMzU7/OhAkT+PHHH/nuu+/Ytm0bly9fZsiQIRaLQQghROWm0Wiwk/qKcHAWq9TFxMTQu3dv/P39OX/+PCdOnKBOnTq8+uqrxMbG6gcyOLLCfn/btm2ja9eupKamEhISwvLlyxk6dCigXpascePG7N69m06dOhncp/zKEUKIqu3PP/9kx44dDjnLg3yH2ReL9ambOHEiI0aMYN68efj6/jeicsCAATzyyCNG7UOr1bJ27VqOHTsGQJMmTbjnnnvsZhLi1NRUAP3o3v3795Ofn0/v3r316zRq1IioqKhSk7rc3Fxyc3P199PS0qwctRBCCHtmygT9QpTFYs2ve/fu5amnnir2eI0aNUhISDC4/b///kuDBg0YPnw4a9euZe3atQwfPpz69esbvG5sbGysSbEa0xx8K51Ox/PPP8/tt9+un2A5ISEBNzc3AgICiqwbFhZW6jnPmTMHf39//S0yUq5gIIQQVZk0vwpLsVhS5+7uXmLV6eTJk4SEGL526OjRo2natCkXL17kwIEDHDhwgLi4OFq0aMGYMWPK3LZ9+/Y89dRT7N27t9R1UlNT+fTTT2nWrBmrV682fEK3GDt2LEeOHGHlypUmb3uzyZMnk5qaqr/FxcWZtT8hhBBCCLBg8+ugQYOYOXMm3377LaD+8oiNjeWll17ivvvuM7j9oUOH2LdvH4GBgfrHAgMDmT17Nu3bty9z26NHjzJ79mz69OmDh4cHbdu2JSIiAg8PD5KTkzl69Cj//vsvbdq0Yd68eQwYMMCkcxs3bhw//fQT27dvp2bNmvrHw8PDycvLIyUlpUi1LjExkfDw8BL35e7ubtHBHEIIIRybVOqEpVisUjd//nwyMjIIDQ0lOzubbt26Ua9ePXx9fZk9e7bB7Rs0aEBiYmKxx69cuUK9evXK3LZatWosWLCA+Ph4Fi1aRP369bl27RqnTp0CYNiwYezfv5/du3eblNApisK4ceNYu3Ytf/zxB7VrF72UVtu2bXF1dWXz5s36x06cOEFsbKzBufWEEEKIQpLUCUuwWKXO39+f3377jZ07dxITE0NGRgZt2rQpMoigLHPmzOHZZ5/ltdde0w8w2LNnDzNnzuTNN98s0rRb2ggbT09Phg4dqh+Jaq6xY8eyfPly1q9fj6+vr76fnL+/P56envj7+/PEE08wceJEgoKC8PPzY/z48XTu3Nmoka9CCCGEDJQQlmI3kw87Of1XNCx8gReGdvN9jUaDVqutkJhKe6N9/vnnjBgxAvhv8uEVK1YUmXy4tObXW8lwcCGEqNr+/vtvNm3axKuvvmrrUEwm32H2xaJzhWzevJl33nlHPyVJ48aNef75542q1m3ZssWSoViEMfmuh4cHH3zwAR988EEFRCSEEEIIUTKLJXUffvghzz33HEOHDuW5554D1ObTAQMG8M477zB27Ngyt+/WrZulQhFCCCEchgyUEJZisaTujTfe4J133mHcuHH6x5599lluv/123njjDYNJHahNmTExMVy5cgWdTldk2aBBgywVqhBCCGFXJKkTlmCxpC4lJYU777yz2ON9+/blpZdeMrj9hg0bePzxx7l27VqxZRXZj04IIYSoSDJQQliKxaY0GTRoEGvXri32+Pr167n77rsNbj9+/Hjuv/9+4uPj0el0RW6S0AkhhKispPlVWIrFKnVNmjRh9uzZbN26VT9H2549e/jzzz954YUXWLhwoX7dZ599ttj2iYmJTJw4kbCwMEuFJIQQQghRZVgsqfvss88IDAzk6NGjHD16VP94QEAAn332mf6+RqMpMakbOnQoW7dupW7dupYKSQghhLB7t07bJUR5WSypO3fuHIC+T1xwcLBJ2y9atIj777+fHTt20Lx5c1xdXYssLykRFEIIIYQQKoskdSkpKUyZMoVVq1aRnJwMqNdtfeihh5g1a1aR66KWZsWKFWzatAkPDw+2bt1a5NdKadU9IYQQwtFJpU5YitlJXVJSEp07d+bSpUsMGzaMxo0bA3D06FGWLVvG5s2b2bVrF4GBgWXuZ8qUKcyYMYOXX365yNUlhBBCiMpMEjlhKWYndTNnzsTNzY0zZ84UG+Qwc+ZM+vbty8yZM3nnnXfK3E9eXh4PPvigJHRCCCGqJBkBK8xldga1bt063n777RJHrYaHhzNv3rwSpzq51fDhw1m1apW54QghhBAO5dbrnQtRXmZX6uLj42natGmpy5s1a0ZCQoLB/Wi1WubNm8fGjRtp0aJFsYESCxYsMDdUIYQQQohKy+ykLjg4mPPnz1OzZs0Sl587d46goCCD+zl8+DCtW7cG4MiRI0WWSX8DIYQQlZVU6oSlmJ3U9evXjylTpvDbb7/h5uZWZFlubi5Tp04t8fJht9qyZYu5oQghhBAORwoXwlLM7lM3c+ZMTpw4Qf369Zk3bx4//PAD69evZ+7cudSvX59jx44xY8YMo/d3+vRpNm7cSHZ2NmDbXy7bt29n4MCBREREoNFoWLduXZHlI0aMQKPRFLkZk8AKIYQQt5JKnTCX2ZW6mjVrsnv3bp555hkmT56sf1FqNBr69OnDokWLiIyMNLif69ev88ADD7BlyxY0Gg2nTp2iTp06PPHEEwQGBjJ//nxzQzVZZmYmLVu2ZNSoUQwZMqTEde68804+//xz/X13d/eKCk8IIUQlIM2vwlIsMvlw7dq1+fXXX0lOTubUqVMA1KtXz6i+dIUmTJiAq6srsbGx+rnuAB588EEmTpxok6Suf//+9O/fv8x13N3dCQ8PN3qfubm55Obm6u+npaWVOz4hhBCVx4EDB3BxsdiFnkrl7e1NkyZNrH4cUfEs+uoJDAykQ4cO5dp206ZNbNy4sdiAi/r163PhwgVLhGcVW7duJTQ0lMDAQHr27MmsWbOoVq1aqevPmTPHpOZoIYQQlVtgYCBubm5s3ry5Qo4XEREhSV0lZf2fBEbKzMzEy8ur2ONJSUl226R55513MmTIEGrXrs2ZM2d45ZVX6N+/P7t378bZ2bnEbSZPnszEiRP199PS0oxqnhZCCFE51ahRg8mTJ9s6DFEJ2E1S16VLF7788ktef/11QO1joNPpmDdvHj169LBxdCV76KGH9P9v3rw5LVq0oG7dumzdupVevXqVuI27u7vdJqlCCCGEcFx2k9TNmzePXr16sW/fPvLy8vi///s//v33X5KSkvjzzz9tHZ5R6tSpQ3BwMKdPny41qbtVYcdY6VsnhBDC0RR+d8kgD/tgN0mdn58fx44d46OPPsLX15eMjAyGDBnC2LFjyc/Pt3V4Rrl48SLXr1+nevXqRm+Tnp4OIE2wQgghHFZ6ejr+/v62DqPK0yh2kl47OzsTHx9PaGhokcevX79OaGgoWq22wmPKyMjg9OnTALRu3ZoFCxbQo0cPgoKCCAoKYsaMGdx3332Eh4dz5swZ/u///o/09HQOHz5sdBOrTqfj8uXL+Pr62nQCysK+fXFxcfj5+dksDnPIOdgHOQf7URnOQ87BPpR2DoqikJ6eTkREBE5OZk99K8xkN5W60nLLjIwMPDw8Kjga1b59+4r05ysc4DB8+HA++ugjYmJi+OKLL0hJSSEiIoK+ffvy+uuvm9RnzsnJqdRLrNmCn5+fw37oFJJzsA9yDvajMpyHnIN9KOkcpEJnP2ye1BUmShqNhmnTphUZAavVavnrr79o1aqVTWLr3r17mf0ENm7cWIHRCCGEEEKUzuZJ3cGDBwG1Unf48OEi1491c3OjZcuWTJo0yVbhCSGEEEI4BJsndVu2bAFg5MiRvPfeew5fmnZU7u7uTJ8+3aGnW5FzsA9yDvajMpyHnIN9qAznUBXYzUAJIYQQQghRfjJURQghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEpCkTgghhBCiEnCxdQBVnU6n4/Lly/j6+qLRaGwdjhBCCGE0RVFIT08nIiICJyepE9maJHVm2r59O2+99Rb79+8nPj6etWvXMnjwYKO3v3z5MpGRkdYLUAghhLCyuLg4atasaeswqjxJ6syUmZlJy5YtGTVqFEOGDDF5e19fX0B9Q/j5+Vk6PCGEEMJq0tLSiIyM1H+XCduSpM5M/fv3p3///kavn5ubS25urv5+eno6AH5+fhWT1J1eD6nnzNtHdB8IbmqZeMoj/i+4vLvsdaJ7Q3CzkpflpsLRr0GXX/4YXDygyXBw9Sx5ecJeuPRn+fdvCS4e0ORxcPUqeXnCPri0s+x9RPaA0JaWj81S0uLg9BpQlNLXqdYYavWruJhulX4RTq0uO8agRlD7zvIf4/wmuH609OUaDdS7F/yiSonxEpz6vuwYAxtAnQGlLz/2DWRdNS7e0tS/F/yizdtHWY4th6wr5u2j3mDwr1W+bQty4N8voCC79HXc/KDZCNCUsynz2r9w4bey1/EOh0YPlW//pZDuQ/ahXEmdVqtl7dq1HDt2DIDGjRszePBgXFwkRzRkzpw5zJgxw3YB/Hi/+gHv5Fa+7QuyoMH9cPdKy8Zlim2TIH4POHuUvLwgS/0CG/R9ycvP/gR/jANXb6AcH0SKTj2Gf+3Sk4VtL6oJk0spSZ+1FcboG1X6F/GOlyFua+kxFmRD7QFw7w/WitJ8MYvhrzfA1afk5dpc8AiEpxMrNq6bxXwCe14vPUZdHrj6wthr5T/GhhGQfQ2c3Utenp8JGfHQdW7Jyw8vgd2vlR2jiyeMSyl5eWYi/PKo+kNCU87vgfxMyLgE3d4q3/aGZF2FX4aZH2PaBejxTvm2v7QTfv8fuHiVnLQpWvV9F9EJqjUp3zF2z1QTdJdSfswBhLezeFIn7IPJr+x///2XQYMGkZCQQMOGDQF48803CQkJ4ccff6RZs1KqIwKAyZMnM3HiRP39wtJ1hdEVQJ+PocWT5dt+9Z3qB48t6QrUKlm/JSUvX3NX2THqCtR/x6WAUzk+3DPi4eMI0Bk4RpNH4c5lpu/fErKuwEdhhp+HRg/DgK9KXr5+iJoU2TNdAQTWh1EnS17+95uw10pJgrF0BeBfB0afKXn53rfhr9nmH+O2GdBxcsnLlzYq+7WgFKg/AMZcKHn5/nfhz6llHx9g0NryVxyXNbXuZ0vhvu/+DureXb59fNHCvBgLn6dRJ8G3RvHl8X/D8o5lf7YYohRAdF+479fy70M4LJO/0UaPHk3Tpk3Zt28fgYGBACQnJzNixAjGjBnDrl27jNrP5s2b2bx5M1euXEGn0xVZtnTpUlPDchju7u64u5fya7pClNG8YvQuLLAP8wIwYpUy1rFY/GXtx9bP0Q1lnqsxMdrJeZTGqL+lrc+hImKsiPdEBbzerfnZYql9m7UfY7c14xg2/3wWtmRyUnfo0KEiCR1AYGAgs2fPpn379kbtY8aMGcycOZN27dpRvXp1aYsXQgghhDCTyUldgwYNSExMpGnToh3lr1y5Qr169Yzax+LFi1m2bBmPPfaYqYcXFmFOEq3B5pUPRVH7BZZGYyjGwmXlfB4Kj22w8mHLHyuFxzYQo6Hn0e5/9Rt6nh3k9Wru82zo9WbMe8KcGAuXmfUD3dp/KwvEaPB5NBSCgRiM+WwxfBAz/w7CkZmc1M2ZM4dnn32W1157jU6dOgGwZ88eZs6cyZtvvklaWpp+3dJGc+bl5XHbbbeVM2T7kpGRwenTp/X3z507x6FDhwgKCiIqqpSRZkIIIYQQFmZyUnf33WoH0wceeEDfbKrc+FUxcOBA/X2NRoNWW3Jnz9GjR7N8+XKmTi2j462D2LdvHz169NDfLxwEMXz4cJYtW2ajqEphiV/TdlG9MaI6Y9WqghFVMIMxWpsxv/gdoMpliDEVKkd4vVqiT12Zr2cDxzBYWbZy9Rus/7dSLBCjoc8Ww0EYiMGYzxZDh7D1Z4+wJZOTui1btph90JycHD755BN+//13WrRogaura5HlCxYsMPsYFaV79+76pFYIIYQQwlZMTuq6detm9kFjYmJo1aoVAEeOHCmyTAZNWJOFfqnaunpjqI+StasKRvV7sXG/Fo0Rv/grS586s14LFaBCnmdzK5Y2fk8ZdQxz2UG/P+lTJ6ysXDMw5uTkEBMTU+J0JIMGDTK4vSWqfUIIIYQQ4j8mJ3UbNmzg8ccf59q14rOfl9WPrjQXL14EkAsBV4Sq0qdOo1GvqFDq5hXQp87m/Vos0e/PDqpcBkmfOvUQVu5TVxGjXx2hT525o1+lT52wMpMvLjd+/Hjuv/9+4uPj0el0RW7GJnQ6nY6ZM2fi7+9PdHQ00dHRBAQE8Prrrxer/AkhhBBCCMNMrtQlJiYyceJEwsLCyn3QKVOm8NlnnzF37lxuv/12AHbu3Mlrr71GTk4Os2ebeckcUQpH6PdiBGOqEmZfSaEsDtSnzqy59OyhymWA2f0rK0CFjNCVPnWGWahPnTl/K4MVTQtU6mw+8l7YkslJ3dChQ9m6dSt169Yt90G/+OILlixZUqT/XYsWLahRowbPPPOMJHVCCCGEECYyOalbtGgR999/Pzt27KB58+bFpiN59tlnDe4jKSmJRo0aFXu8UaNGJCUlmRqSMJalZn23i+qNodnzy2LmL1ljRpaqK5b/GGYzsk+dWVfmsAcO0qfO2tVEi1QszXhPOdRniyXe++VloKJpkdGvWCBO4ahMTupWrFjBpk2b8PDwYOvWrUWmINFoNEYldS1btmTRokUsXLiwyOOLFi2iZcuWpoYkqhxjPvDsoclNCAdhzOvV4V/Tlorf3p8He49PWJPJSd2UKVOYMWMGL7/8Mk5OJo+zAGDevHncdddd/P7773Tu3BmA3bt3ExcXxy+//FKufQpjWGrWd3sYzGKoKlEGg1UNY49t4MPT3vvUVYbRr47Qp66irihhbsXS4PNo6PjGrGfo+HZeqTO3+m5wBK4l+tSVtX9R2ZmcleXl5fHggw+WO6EDdQLjkydPcu+995KSkkJKSgpDhgzhxIkTdOnSpdz7FVWEQ1QVbH18IUzhANVvc1nqM8Hmny0G2Ht8wqpMrtQNHz6cVatW8corr5h14IiICBkQUdEs1e/FHj7cDfYFK4uF+tQZ/PC08z51leWKEvbep64inufK0qeuIj5bzJ1LzywVcUWJMvYvKj2TkzqtVsu8efPYuHGjSddtjYmJoVmzZjg5ORETE1PmMVq0aGFqWKJKcYCqgq0TCSFM4RDVb3NJnzpR+Zmc1B0+fJjWrVsDpl23tVWrViQkJBAaGkqrVq3QaDQoJXxIlHZVitatWxt9XdgDBw4YtV7VY4E3u0YDOnv/0JA+ddKnrpA9nIMxFTIr96kDK1eWLfTZYu+jX82uvkufOmFdJid15b1u67lz5wgJCdH/31SDBw8u13FFSRz9DW+JLzBrs/XxhX2x9XvO2IEOZW1vzGva1udZBulTJ6oAk5O68oqOjtb//8KFC9x22224uBQ9fEFBAbt27SqybqHp06dbPcZKzyJvdnuofBggfeow6hd/VehTZw+vV2NGnVq9Tx1YtbLsSJ8ttuxTZ6jvofSpE2YyOanr0aNHmc2gf/zxh1H7iI+PJzQ0tMjjqamp9OjRw+hryIpycvQ3vCW+wKzN7pMhUaFs/Z6zRMJnzGva1udZJulTJyo/k5O6Vq1aFbmfn5/PoUOHOHLkCMOHDzdqH4qilJgYXr9+HW9vb4Pba7Va3nnnHb799ltiY2PJy8srslyuSlEaR+r3Yg476VNny0pdVepTZ++jX+2lT51VX6+O9NkifepE5WVyUvfOO++U+Phrr71GRkZGmdsOGTIEUAdDjBgxAnd3d/0yrVZLTEwMt912m8EYZsyYwZIlS3jhhRd49dVXmTJlCufPn2fdunVMmzbNhLOpqhz9DS996oSjsfV7zhKDKKRPnUX3Yy32Hp+wqvLPIHyLRx99lKVLl5a5jr+/P/7+/iiKgq+vr/6+v78/4eHhjBkzhq+//trgsb755hs+/fRTXnjhBVxcXHj44YdZsmQJ06ZNY8+ePZY6pUrIgfq9mMVO+tTZtCmqCvWps/fRrxXRp84SP3SkT50xG5t3bOlTJ6zMYgMldu/ejYeHR5nrfP755wDUqlWLSZMmGdXUWpKEhASaN28OgI+PD6mpqQDcfffdTJ061ah9xMbGcuHCBbKysggJCaFp06ZFKoeVm4O/4R2hT53Njy/siq2/ZM1+vxgbvz1/tkifOlH5mZzUFTahFlIUhfj4ePbt22d0QmXuSNaaNWsSHx9PVFQUdevWZdOmTbRp04a9e/eWmZidP3+ejz76iJUrV3Lx4sUi8+S5ubnRpUsXxowZw3333WfWZdAqNVt/ORnDmBgr5Ne6PVTqzFnPAf7WgFlXQrALForR7HO198TTYgcyY1NLxWioT5219i8qO5Mzl5ubTP39/QkKCqJ79+788ssvJiVr33//PQ888ACdOnWiTZs2RW6G3HvvvWzevBmA8ePHM3XqVOrXr8/jjz/OqFGjStzm2WefpWXLlpw7d45Zs2Zx9OhRUlNTycvLIyEhgV9++YU77riDadOm0aJFC/bu3Wv0uTgMi1zKBztokjOzU7jF4q+IY5jL0EAJc7a3B45wJQRLDGIwtLmZfeKMbZotdT0H+GypiPe9xbY14xg2f70LWzK5UlfYhGqOhQsXMmXKFEaMGMH69esZOXIkZ86cYe/evYwdO9bg9nPnztX//8EHHyQqKordu3dTv359Bg4cWOI23t7enD17lmrVqhVbFhoaSs+ePenZsyfTp09nw4YNxMXF0b59+/KfZKXlCL8AjYmxAn6t28MVJQyvaIF92JjBPnV2zurVH2M3t/VzVUHHt2WfOoMx2MlrQTiscvep27dvH8eOHQOgSZMmtG3b1uhtP/zwQz755BMefvhhli1bxv/93/9Rp04dpk2bVq7pSDp37kznzp3LXGfOnDlG7+/OO+80OQbHYGg4van7sSGzPpiryCg4KDvGynC9T6Pis4eBEtYerW3twRjGVhorwWeLNRn7N7D3952wWyYndRcvXuThhx/mzz//JCAgAICUlBRuu+02Vq5cSc2aNQ3uIzY2Vj91iaenJ+np6QA89thjdOrUiUWLFhncx6lTp9iyZQtXrlxBp9MVWSbTmliTHfwCtMhoQmv/WjdzhK3ZLFFNtIO/tVEcvdpoD33qjH29WvN1be2/lQUST4uNCDcw+tUs5s7DKRyZyUnd6NGjyc/P59ixYzRs2BCAEydOMHLkSEaPHs2GDRsM7iM8PJykpCSio6OJiopiz549+v5uihFvmE8//ZSnn36a4OBgwsPDi0xkrNFoJKkrTaXpUwc2nyjVovuxJulTZ/vXawWM1rbYPHOlbW5kpbFSfLZYUwX0qRNVmslJ3bZt29i1a5c+oQNo2LAh77//Pl26dDFqHz179uSHH36gdevWjBw5kgkTJvD999+zb9++YqNrSzJr1ixmz57NSy+9ZGr4wlx28QvQAl9g1u5XY/ZVK8xkiT51jlKpc/hqox30ozL29aooVizUWflvpVigUmepufSs2qfO1q0EwpZMTuoiIyPJz88v9rhWqyUiIsKofXzyySf6JtOxY8dSrVo1du3axaBBg3jqqacMbp+cnMz9999vWuCCytXvxdYTpYJ9PA8GGLxMmDnb2wFH6FNXIV+y0qfOIRj9N6jkz4OwGpOnNHnrrbcYP348+/bt0z+2b98+nnvuOd5++23jDurkhIvLf/nkQw89xMKFCxk/fjxubm4Gt7///vvZtGmTqaGX6OLFi8X65Imy2MEvQFv3qTOqomDrX8sW6FNnF1VZYzj4OVgqxgrrU2ctFZT4mvM8OUKfOlu3EgibMrlSN2LECLKysujYsaM+MSsoKMDFxYVRo0YVmSfu5pGsMTExRh+jRYsWZS6vV68eU6dOZc+ePTRv3hxXV9ciy5999lmjj9WkSRMOHTpEnTp1jN7GYVmq34s9/IqU0a9GMncuPXs/RweoNprUtFnO13Vl6VNn9683c8noV2FdJid17777brkO1KpVKzQajcGBEBqNBq1WW+Y6n3zyCT4+Pmzbto1t27YV296UpM6YgRniJnbxC1D61BkkfeoKF1ZYGOVnJ33qjH1dW4vV3y921KfOqleUsHUrgbAlk5O64cOHl+tA586dK9d21t5X1VKVKlRlkD51hQuN2YGlIrEOh6g2mtKPqrxfxrauWMpni3GkT52wrnJPPmyq6OjoijqUSV555RWCgoJsHUYFc/RfcRUxmau5bH18YV9s/Z4zd0qVwu2tOf+jlVnqM8Hmny0G2Ht8wqoqLKm72Zdfflnm8scff7zM5RMnTizxcY1Gg4eHB/Xq1eOee+4xKlmbPHmywXUqDYt18LXzDw2j5gSzwEAJq05wbC4jvoQNNRFbrFO4NRn6W9pBkmHKdVXLXaizwFx45jR/OtJni7kDJcxhqO+h0Z8tBthFNxlhCzZJ6p577rki9/Pz88nKysLNzQ0vLy+DSd3Bgwc5cOAAWq1WP1/eyZMncXZ2plGjRnz44Ye88MIL7Ny5kyZNmljtPByWo7/hLfEFZm12nwyJCmXr95ylEj5jE1S7VFW6Xdh7fMKaTJ7SxBKSk5OL3DIyMjhx4gR33HEHK1asMLj9PffcQ+/evbl8+TL79+9n//79XLx4kT59+vDwww9z6dIlunbtyoQJEyrgbByJhSbNtPuExYimWYsMlLDjSp1RX8LGVLns/G9tqIO/pSofZrFU06ahY1iqibW8xzdThX22WHmQVJkMDdawxGuhrP2Lys7kpG7UqFH6a7XeLDMzs8h0JqaqX78+c+fOLVbFK8lbb73F66+/jp+fn/4xf39/XnvtNebNm4eXlxfTpk1j//795Y6ncnP0N7z0qROOxtbvOXPfLw7wQ8YQ6VMnqgCTk7ovvviC7OzsYo9nZ2cb7CtniIuLC5cvXza4XmpqKleuXCn2+NWrV0lLSwMgICCAvLw8s+KpdByp34s57KVPnU2boqpQnzqjpjSx4XmY0qfOLGYmbtKnzrrbgvSpE1ZndJ+6tLQ0FEVBURTS09Px8PDQL9Nqtfzyyy+EhoYata8ffvihyH1FUYiPj2fRokXcfvvtBre/5557GDVqFPPnz6d9+/YA7N27l0mTJjF48GAA/v77bxo0aKDfJjY2lqioKKPiA7h06RI1atQwen3H4uBveEfoU2fz4wu7YusvWelTh/SpE1WB0UldQEAAGo0GjUZTJFkqpNFomDFjhlH7Kky8bt42JCSEnj17Mn/+fIPbf/zxx0yYMIGHHnqIgoICQK3yDR8+nHfeeQeARo0asWTJEv027du3Z/DgwYwePVqfCN4qNTWVb7/9lvfee48xY8aYNImxY3Ckfi/mkD51VaZPnaFzsMTIUrNZuU+d0VdzkD51Nw5ko21B+tQJazM6qduyZQuKotCzZ09Wr15dZLoQNzc3oqOjiYiIMGpf5l5r1cfHh08//ZR33nmHs2fPAlCnTh18fHz067Rq1arINkePHmX27Nn06dMHDw8P2rZtS0REBB4eHiQnJ3P06FH+/fdf2rRpw7x58xgwYIBZMdo1u/41bQwH6FNn6+MLO2Pr95z0qZM+daIqMDqp69atG6BezSEyMhInJ5sMnC3Cx8fH4HViC1WrVo0FCxYwe/Zsfv75Z3bu3MmFCxfIzs4mODiYYcOG0a9fP5o1a2blqG3Jgfq9mEX61P3HQJ86Q1Uue/+CMFh1rQp96gxVf0yMw6wYzOEAferMTVqlT52wMpPnqSu8MkRWVhaxsbHFBiOUlmSVNmFwSRYsWFDssSFDhrBs2TL8/PwYMmRImduvWbOm1GWenp4MHTqUoUOHGh1P5VMJ3vDyoSUcir33qTO4AwuvJ4SwBpOTuqtXrzJy5Eh+/fXXEpdrtdoSHz948GCR+wcOHKCgoKDY5MFt27YtcXt/f380Nz6Y/P39TQ1bgGWqLvZQvTGq4mBGhcogY6o/5h7DEgxVPoypctl5pc6UPnU24wh96ox9vZayD4f4bDGyolkWQ58tZsdQUXMWisrK5KTu+eefJyUlhb/++ovu3buzdu1aEhMTmTVrVpmDHLZs2aL//4IFC/D19eWLL74gMDAQUCckHjlyJF26dClx+88//1z//w8//BCdToe3tzcA58+fZ926dTRu3Jh+/fqZekpVT6WoclWGcxBVhs3fc2Ye39j4bX6eQlRtJid1f/zxB+vXr6ddu3Y4OTkRHR1Nnz598PPzY86cOdx1110G9zF//nw2bdqkT+gAAgMDmTVrFn379uWFF14oc/t77rmHIUOG8L///Y+UlBQ6deqEq6sr165dY8GCBTz99NOmnlYV4UD9XspkTOXDnAqVod0bUf0xe4StBRisfBgx+tXWVVlDHKFPnaWmCzG4fzP61Bn7ei11Hw7w2WJ0RbMsZsZoMAYLvF7NbokQjszk0Q6ZmZn6+egCAwO5evUqAM2bN+fAgQNG7SMtLU2/3c2uXr1a4tUqbnXgwAF9Re/7778nLCyMCxcu8OWXX7Jw4UJjT6UKqwRveFsnTEKYxA4SfPN2YOH1hBDWYHJS17BhQ06cOAFAy5Yt+fjjj7l06RKLFy+mevXqRu3j3nvvZeTIkaxZs4aLFy9y8eJFVq9ezRNPPGFwEASogzR8fX0B2LRpE0OGDMHJyYlOnTpx4cIFU0+p6rDUrO+2rt4Y06fOrAqV0YFUwDHMYUTfQkNXlLB5VdYQB+hTZ7HpQgzs39H71Fn9s8VCferMitFADBZ5vdpBK4GwGZObX5977jni4+MBmD59OnfeeSfffPMNbm5uLFu2zKh9LF68mEmTJvHII4+Qn5+vBuLiwhNPPMFbb71lcPt69eqxbt067r33XjZu3MiECRMAuHLlSpHrwYpSVIo3fGU4B1Fl2Pw9J33qhKgKTE7qHn30Uf3/27Zty4ULFzh+/DhRUVEEBwcbtQ8vLy8+/PBD3nrrLc6cOQNA3bp19QMfDJk2bRqPPPIIEyZMoFevXnTu3BlQq3atW7c28YyqEguNULN59cbMPnUW6e9WEccwkyX61Nn8b21AZehTZ3aMFupTZ0ziZ80+dVb/bLFApc4R+tTZRSuBsBWTk7pCeXl5nDt3jrp169KmTZty7cPb21t/ZQpjEzqAoUOHcscddxAfH0/Lli31j/fq1Yt77723XLEIIYQQQjgyk/vUZWVl8cQTT+Dl5UXTpk2JjY0FYPz48cydO9eofeh0OmbOnIm/vz/R0dFER0cTEBDA66+/bvQlxMLDw2ndunWRK1t06NCBRo0amXpKVYdiB79ULcHsKyFY4JdshfXbM4cF+tTZuv+kQY7Sp86KMRrTp85gFczMiqcjfLZYYvSrxfrUlbF/MO8Y9tBKIGzG5KRu8uTJ/PPPP2zduhUPDw/9471792bVqlVG7WPKlCksWrSIuXPncvDgQQ4ePMgbb7zB+++/z9SpU00NSQghhBCiyjO5+XXdunWsWrWKTp066a/wANC0aVN9/zhDvvjiC5YsWcKgQYP0j7Vo0YIaNWrwzDPPMHv2bFPDEkaxh1+qlmDmlRCkT90N0qeuYlg7RmOqZGa+FgxWkBzhs8UeqonGvOcK17PWMURlZnKl7urVq/p56m6WmZlZJMkrS1JSUonNpI0aNSIpKcnUkIQQQgghqjyTk7p27drx888/6+8XJnJLlizRj0I1pGXLlixatKjY44sWLSoy8EFYi7m/VG3N3D51lL29MYyaE8zWz5URxzfYD8sRGFP5sLUKeJ6NqgaWwmIVTzv+bLGLPnUGjm+J14I9tBIImzG5+fWNN96gf//+HD16lIKCAt577z2OHj3Krl272LZtm1H7mDdvHnfddRe///67PhHcvXs3sbGx/Prrr6aGJIxlsaYNO2+SM8hC8du8GdoYhqaxMLS5vZ+jsfHZeqCEBdcr93YVMAjB/B1ZaD92ytqvBVHlmVypu+OOOzh06BAFBQU0b96cTZs2ERoayu7du2nbtq1R++jWrRsnTpxgyJAhpKSkkJKSwpAhQzh58qT+8l/Ciixx3VNbMqqqYOQos3IzZk4wO+hTZ3ilci6zJw5QbTSnimb8Qcp5fLDYKGK7/myxhz51ho5viefADj57hM2Ua566unXr8umnn5p14GrVqjFo0CA6deqkn8Zk3759AEUGUAhLqkoVqjJUpaqCwc7xBndgqUiswyEqH9aupNlBpU4+W4xkD38rUZkZndSlpaUZtZ4xl+nasGEDjz/+ONevX0e55U2s0WjQarXGhiXKxY77vRjN3F+71u5TZ+Q6VlVF+tRVSBXMXBXxPJv5PFjkeXSAzxaznm8LxFgR7zlHee8KizM6qQsICChzdKuiKEYnZOPHj+f+++9n2rRphIWFGRuCMFtlqVCZ2xesIqoKtn6OCkmfOtPWswKrVxMtsH9z++U5QvXbUjGaOzGwvRxDVEpGJ3VbtmzR/19RFAYMGMCSJUuoUaOGyQdNTExk4sSJlSah++CDD3jrrbdISEigZcuWvP/++3To0MHWYZXBnvu9GMnsX7tW7lNnkWOYS/rUOcTr1VLPs1XfExVQqauwv5Wt59Kzcf9KUakZndR169atyH1nZ2c6depEnTp1TD7o0KFD2bp1K3Xr1jV5W3uzatUqJk6cyOLFi+nYsSPvvvsu/fr148SJEyXO52dT9vBLtcKOb4mqhBnHcIRKXWXoU2eJKpXVWbmaaJHRr+Y+j47w2WIP1cSKqCzb+3tWWFO5BkqYa9GiRdx///3s2LGD5s2b4+rqWmT5s88+a4uwymXBggU8+eSTjBw5EoDFixfz888/s3TpUl5++WUbR1cKW/cpsQhr9h8y5vAO0Keuqox+lT51ho9RUf0rHeGzxdYx2noktKjUbJLUrVixgk2bNuHh4cHWrVuL9NXTaDQOk9Tl5eWxf/9+Jk+erH/MycmJ3r17s3v37hK3yc3NJTc3V3/f2AEoRkncD1ueL315QY5ljpN8AlbacOqZnOuG10k+VXqMaRcsE8f+d+BEKdc7zrpqmWOY68B7cPL7kpdlJRrePu28bf/Whlw/BuHtDK+3uh842eTjTo0xtJXh9db0BydXw+vdSmfkwLK4raX/LZOOQ3Azw/tYd3fJMeamGheDIZe2W+/1lmehz9rLf5Y/xozLxq23YzLse7t8x0g6AdWNuxCAqHzM+pQz9rJgt5oyZQozZszg5ZdfxsnJ5Kny7Ma1a9fQarXF+gaGhYVx/PjxEreZM2cOM2bMsE5ALp4QYKBJO7wdhBnxJViaxo9CQVb5t7eEgPpQ757Slzd+BPLTy9i+LoSb2eex82uQdKyMY9SHeveadwxzdZ4O1/8tfXlAfag/pPTljR68kUDbcXNOQF2oO7j05TW7QIsxoM0tfR1rC6gLdcuYpinidvNjDG6mnmtpWj4N3mX0YQ6oC3XuLiPG26DFU6At44dhVE/wN707jl6Lp8CzWvm3N0bN7hBQr/zbtxgDHgHl3z6gLgQWv0Smnmc1aP8SZCWYd4zGw8q/vXBoGuXWOUVKMWRI0Q//H3/8kZ49e+Lt7V3k8TVr1hjcV1BQEHv37nX4PnWXL1+mRo0a7Nq1q8gl0v7v//6Pbdu28ddffxXbpqRKXWRkJKmpqUZNByOEEELYi7S0NPz9/eU7zE4YXanz9/cvcv/RRx8t90GHDx/OqlWreOWVV8q9D3sQHByMs7MziYlFm7ESExMJDw8vcRt3d3fc3d0rIjwhhBBCVCFGJ3Wff/65xQ6q1WqZN28eGzdupEWLFsUGSixYsMBix7ImNzc32rZty+bNmxk8eDAAOp2OzZs3M27cOKP2UVgotWjfOiGEEKICFH53GdnoJ6zMJj2HDx8+TOvWrQE4cuRIkWXl7adnKxMnTmT48OG0a9eODh068O6775KZmakfDWtIerra9ysyMtKaYQohhBBWk56eXqxFT1Q8o/vUidItWrRIP/lwq1atWLhwIR07djRqW51Ox+XLl/H19bVpQlvYty8uLs5h+0XIOdgHOQf7URnOQ87BPpR2DoqikJ6eTkREhEMPfKwsbDTGv3IZN26c0c2tt3JycqJmzZoWjqj8/Pz8HPZDp5Ccg32Qc7AfleE85BzsQ0nnIBU6+yFptRBCCCFEJSBJnRBCCCFEJSBJnQDUqVamT5/u0NOtyDnYBzkH+1EZzkPOwT5UhnOoCmSghBBCCCFEJSCVOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISkCSOiGEEEKISsDF1gFUdTqdjsuXL+Pr64tGo7F1OEIIIYTRFEUhPT2diIgInJykTmRrktSZafv27bz11lvs37+f+Ph41q5dy+DBg43e/vLly0RGRlovQCGEEMLK4uLiqFmzpq3DqPIkqTNTZmYmLVu2ZNSoUQwZMsTk7X19fQH1DeHn52fp8IQQQgirSUtLIzIyUv9dJmxLkjoz9e/fn/79+5d7+8ImVz8/P0nqhBBCOCTpPmQfJKmrYLm5ueTm5urvp6WlWWzf+/bB2LEW252wcyEhsGYNuLmVb/v58+Hbb8te5/HHS39NHT8OTz4JeXnlO76oWtzc4JNPoHFj6+z//HkYPhxycsq/D1dXWLwYmjWzWFgmKSiA++6DhATrHqdpU1i61LrHELYhSV0FmzNnDjNmzLDKvn18oGVLq+xa2JnYWPj5Z7h2DSIiyrePH3+E69ehZ8+Sl2/aBL/8UnpSd/Ag7NwJI0aoX4ZClKagAD7/XH3NWCupO3IEtm+HYcPAy8v07XU6+OwzOHDAdkldSgr88AP06QO1alnvOLVrW2/fwrYkqatgkydPZuLEifr7hf0RLKFRI/WXsKj8fvkFNm40bx+KAp07l/6aGTLEuKrHokXg7W1eLKJyy85WkzpFsf6x3n4bwsNN3y4/X03qbKnw+Rk7Fu65x7axCMckSV0Fc3d3x93d3dZhiEqiIr4k7fHYQtzKUq9HeV0LRyaTygjhgCzVJ7ms/Rh7DOkfLQypyNdIeY9lT69je4pFOBap1JkpIyOD06dP6++fO3eOQ4cOERQURFRUlA0jE1WBOVUFY7Ytax2paAhTWfM1UxkqdfKeEuaSpM5M+/bto0ePHvr7hf3lhg8fzrJly2wUlajspFInHIlU6kxjT7EIxyJJnZm6d++OIj+vhI1IpU44EqnU2e+xReUgfeqEcEBSqROORCp1prGnWIRjkaROCAcmlTrhSKRSZ7/HFpWDJHVCOCCp1AlH4giVOntSGc5B2IYkdUI4MKnUCUdiz5W6wkRKKnXCkVWapC4tLY1169Zx7NgxW4cihNVJpU44EqnUmaYynIOwDYdN6h544AEWLVoEQHZ2Nu3ateOBBx6gRYsWrF692sbRCVExpFInHIlU6som7ylhLodN6rZv306XLl0AWLt2LYqikJKSwsKFC5k1a5aNoxPCuqRSJxyJVOpMUxnOQdiGwyZ1qampBAUFAbBhwwbuu+8+vLy8uOuuuzh16pSNoxOiYkilTjgSe67UWXo/jnZsUTk4bFIXGRnJ7t27yczMZMOGDfTt2xeA5ORkPDw8bBydENYllTrhSKRSZ5rKcA7CNhz2ihLPP/88w4YNw8fHh+joaLp37w6ozbLNmze3bXBCVBCp1AlHYu+VOo1GKnXCsTlsUvfMM8/QoUMH4uLi6NOnD05OatGxTp060qdOCCGEEFWOwyZ1AO3ataNdu3ZFHrvrrrtsFI0QFccSI/UUxXDzqzGVOmkqEoZUxMhSS7we7aVSJ+8pUV4OldRNnDjR6HUXLFhgxUiEEEIIIeyLQyV1Bw8eLHL/wIEDFBQU0LBhQwBOnjyJs7Mzbdu2tUV4QlQYS/ySN7dSZ8lYROVWkXPAmVupsyWp1AlzOVRSt2XLFv3/FyxYgK+vL1988QWBgYGAOvJ15MiR+vnrhKjs7KGpSAh7UBmmNBHCXA47pcn8+fOZM2eOPqEDCAwMZNasWcyfP9+GkQlhfVKpE45IKnVlk0qdMJfDJnVpaWlcvXq12ONXr14lPT3dBhEJUfGkUieESip1QjhwUnfvvfcycuRI1qxZw8WLF7l48SKrV6/miSeeYMiQIbYOTwirkkqdcDQVNbJUKnWiKnOoPnU3W7x4MZMmTeKRRx4hPz8fABcXF5544gneeustG0cnRMWQSp0QKqnUCeGgSZ1Wq2Xfvn3Mnj2bt956izNnzgBQt25dvL29bRydENZnT5U6IYzhCJU6W5NKnTCXQyZ1zs7O9O3bl2PHjlG7dm1atGhh65CEsAmp1AmhqgyXCRPCXA7bp65Zs2acPXvW1mEIYRP2UqmTioIwllTqDJNKnTCXwyZ1s2bNYtKkSfz000/Ex8eTlpZW5CZEVSCVOiFUUqkTwkGbXwEGDBgAwKBBg9Dc9LNGURQ0Gg1ardZWoQlhdVKpE45GKnWGSaVOmMthk7qbry4hRFUllTohVFKpE8KBk7pu3brZOgQhbEYqdcLRSKXOMKnUCXM5bFJXKCsri9jYWPLy8oo8bmhE7Llz59ixYwcXLlwgKyuLkJAQWrduTefOnfHw8LBmyEJYjFTqhFBJpU4IB07qrl69ysiRI/n1119LXF5an7pvvvmG9957j3379hEWFkZERASenp4kJSVx5swZPDw8GDZsGC+99BLR0dHWPAUhyk0qdcLRSKXOMKnUCXM57OjX559/npSUFP766y88PT3ZsGEDX3zxBfXr1+eHH34ocZvWrVuzcOFCRowYwYULF4iPj2f//v3s3LmTo0ePkpaWxvr169HpdLRr147vvvuugs9KCNNIpU4IlVTqhHDgSt0ff/zB+vXradeuHU5OTkRHR9OnTx/8/PyYM2cOd911V7Ft5s6dS79+/Urdp7u7O927d6d79+7Mnj2b8+fPW/EMhCg/qdQJRyOVOsOkUifM5bCVuszMTEJDQwEIDAzk6tWrADRv3pwDBw6UuE1ZCd2tqlWrRtu2bc0PVAghhEOQZEo4OodN6ho2bMiJEycAaNmyJR9//DGXLl1i8eLFVK9e3aR93XXXXcTHx1sjTCGsovDLx5zKh7mVOkPbC3Eza1fqLFXlsocuDfK+EuXlsM2vzz33nD4Rmz59OnfeeSfffPMNbm5uLFu2zKR9bd++nezsbCtEKYQQwlFIMiUcncMmdY8++qj+/23btuXChQscP36cqKgogoODbRiZENZniUrdzfsxdVnhseVLUBjL2q+VylCpKyTvK1FeDtv8evbs2SL3vby8aNOmTbkSuujoaFxdXS0VmhAOwZgvL3v4ghOVh72/nmydTNn78yPsn8MmdfXq1SMqKorHHnuMzz77jNOnTxvcJjY2tsTHjxw5QmRkZLHHL126ZHacQliDVOqEo5FKnfHkfSXKy2GTuri4OObMmYOnpyfz5s2jQYMG1KxZk2HDhrFkyZISt2nfvj1PPfUUe/fuLXW/qampfPrppzRr1ozVq1dbK3whbE4qdaKi2fvrydbJlL0/P8L+OWyfuho1ajBs2DCGDRsGwKlTp5g9ezbffPMNK1euZPTo0cW2OXr0KLNnz6ZPnz54eHjQtm1bIiIi8PDwIDk5maNHj/Lvv//Spk0b5s2bx4ABAyr6tIQwilTqhKORSp3x5H0lysthk7qsrCx27tzJ1q1b2bp1KwcPHqRRo0aMGzeO7t27l7hNtWrVWLBgAbNnz+bnn39m586dXLhwgezsbIKDgxk2bBj9+vWjWbNmFXsyQtiAVOpERZPXU9nk+RHmctikLiAggMDAQIYNG8bLL79Mly5dCAwMNGpbT09Phg4dytChQ60cpRDWIZU64WgcoVJnL5cJk/eVKC+HTeoGDBjAzp07WblyJQkJCSQkJNC9e3caNGhg69CEcAhSqRMVTV5PZZPnR5jLYQdKrFu3jmvXrrFhwwY6d+7Mpk2b6NKli76vnRCVmVTqhKORSp1pcQhRHg5bqSvUvHlzCgoKyMvLIycnh40bN7Jq1Sq++eYbW4cmhF2TSp2oaPJ6Kps8P8JcDlupW7BgAYMGDaJatWp07NiRFStW0KBBA1avXs3Vq1dtHZ4QViWVOuFopFJnWhxClIfDVupWrFhBt27dGDNmDF26dMHf39/WIQnhUKRSJyqavJ7KJs+PMJfDJnVlTSAsRFVhyy8B+QIS9sQSr0d7qdQJUV4O2/wKsGPHDh599FE6d+6sv6TXV199xc6dO20cmRDWZYnmGUPNp8Z8wUkzkTCWJEyGWWoCZVF1OWxSt3r1avr164enpycHDx4kNzcXUC/z9cYbb9g4OiEqhlTqhFBJpU4IB07qZs2axeLFi/n0009xdXXVP3777bdz4MABG0YmhPVJpU44mopImBz99SiVOmEuh03qTpw4QdeuXYs97u/vT0pKSsUHJIQQwqFJMiUcncMmdeHh4Zw+fbrY4zt37qROnTo2iEiIimOJKU3MrdTJlCbCFNau1Fnq9WgPXRrkfSXKy2GTuieffJLnnnuOv/76C41Gw+XLl/nmm2944YUXePrpp20dnhBCCAcjyZRwdA47pcnLL7+MTqejV69eZGVl0bVrV9zd3XnxxRcZPXq0rcMTwqqkUiccjVTqjD+2vK9EeTlspU6j0TBlyhSSkpI4cuQIe/bs4erVq/j7+1O7dm1bhyeEEMLBSDIlHJ3DJXW5ublMnjyZdu3acfvtt/PLL7/QpEkT/v33Xxo2bMh7773HhAkTbB2mEFYllTrhaKRSZ/yx5X0lysvhml+nTZvGxx9/TO/evdm1axf3338/I0eOZM+ePcyfP5/7778fZ2dnW4cphBDCwUgyJRydwyV13333HV9++SWDBg3iyJEjtGjRgoKCAv755x808o4UVYRU6oSjkUqd8ceW95UoL4drfr148SJt27YFoFmzZri7uzNhwgRJ6IQQQghRpTlcUqfVanFzc9Pfd3FxwcfHx4YRCVHxpFInHI0jVOpsfZkwqdQJczlc86uiKIwYMQJ3d3cAcnJy+N///oe3t3eR9dasWWOL8IQQQgghbMLhkrrhw4cXuf/oo4/aKBIhbEcqdcLRSKXOMKnUCXM5XFL3+eef2zoEIYQQQgi743B96uzRBx98QK1atfDw8KBjx478/ffftg5JVHJSqROORip1hkmlTphLkjozrVq1iokTJzJ9+nQOHDhAy5Yt6devH1euXLF1aEIIIYSoQhyu+dXeLFiwgCeffJKRI0cCsHjxYn7++WeWLl3Kyy+/bOPoRGVV+Ev+yhW4eLF8+ygoMFypy88vff+pqVJREMbTaNTXTHlfr4YkJ1umUmfNGA1JTPwvDiHKQ5I6M+Tl5bF//34mT56sf8zJyYnevXuze/fuErfJzc0lNzdXfz8tLc3qcYrKx9NT/XfIEMvsp7RlZ85AZGTp60RHm3d8UXV4esLbb6s3awkKMm97T0945x31ZktlvS+FKIskdWa4du0aWq2WsLCwIo+HhYVx/PjxEreZM2cOM2bMqIjwRCUWHQ27d0NKSvn34eQEt99e+vIXX4Q77ii7j1Ht2uU/vqhaNmyAs2ete4yyfoAY45df1B8ytuTnB3Xr2jYG4bgkqatgkydPZuLEifr7aWlpRJr7SSSqpE6drLt/Pz/o18+6xxBVR4MG6s2e1a+v3oRwVJLUmSE4OBhnZ2cSCztC3JCYmEh4eHiJ27i7u+snTgZ1MmWQZlghhBCOp/C7S7HlsGGhJ0mdGdzc3Gjbti2bN29m8ODBAOh0OjZv3sy4ceOM2kd6ejqAVOuEEEI4rPT0dPz9/W0dRpUnSZ2ZJk6cyPDhw2nXrh0dOnTg3XffJTMzUz8a1pCIiAji4uLw9fVFY8MhT4XNwHFxcfj5+dksDnPIOdgHOQf7URnOQ87BPpR2DoqikJ6eTkREhA2jE4UkqTPTgw8+yNWrV5k2bRoJCQm0atWKDRs2FBs8URonJydq1qxp5SiN5+fn57AfOoXkHOyDnIP9qAznIedgH0o6B6nQ2Q9J6ixg3LhxRje3CiGEEEJYg1xRQgghhBCiEpCkTgDqqNzp06cXGZnraOQc7IOcg/2oDOch52AfKsM5VAUaRcYhCyGEEEI4PKnUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUApLUCSGEEEJUAi62DqCqKygo4ODBg4SFheHkJDm2EEIIx6HT6UhMTKR169a4uEhKYWt29RfYsWMHH3/8MWfOnOH777+nRo0afPXVV9SuXZs77rjD1uFZxcGDB+nQoYOtwxBCCCHK7e+//6Z9+/a2DqPKs5ukbvXq1Tz22GMMGzaMgwcPkpubC0BqaipvvPEGv/zyi40jtI6wsDBAfUNUr17dxtEIIYQQxouPj6dDhw767zJhW3aT1M2aNYvFixfz+OOPs3LlSv3jt99+O7NmzbJhZNZV2ORavXp1atasaeNoLEjJU//VuNk2DiGEKIeMbB3ZOTqcnTUE+jqh0WhsHZJdk+5D9sFukroTJ07QtWvXYo/7+/uTkpJS8QGJ8tPlwvmWoHGFWodA42zriIQQolQFWoXj5/M4eCKHgydyOHMpn/QsnX65l4eG21t4Mu6BIHy9JHkR9stuXp3h4eGcPn262OM7d+6kTp06NogItm/fzsCBA4mIiECj0bBu3boiyxVFYdq0aVSvXh1PT0969+7NqVOnbBKrXcnaAnknIPcIZO2wdTRCCFHMtZQCfv4zg6mLrzL4xYs8Oz+Rb39Pw8vTiQd6+/LqqGrMHRvCzDHBPNjHj7/+zeH5BYlkZOsM71wIG7GbSt2TTz7Jc889x9KlS9FoNFy+fJndu3czadIkpk6dapOYMjMzadmyJaNGjWLIkCHFls+bN4+FCxfyxRdfULt2baZOnUq/fv04evQoHh4eNojYTmRuBJdI0KVB9k7w7m7riIQQgqRULb/vzeSPfVmcjM3DSQNN67jzUB8/2jbyoEGUG87OxZtZ72gFXVp5Mf6tBN784joznwqW5lhhl+wmqXv55ZfR6XT06tWLrKwsunbtiru7O5MmTWL8+PE2ial///7079+/xGWKovDuu+/y6quvcs899wDw5ZdfEhYWxrp163jooYcqMlT7khsDnh1AmwY5e20djRCiijtzMY+vN6Sx41AWzk7QubknQ3tWo30TD/x9jOseUqu6Ky8Pr8bUj6/x21+Z9O3kY+WohTCdXSR1Wq2WP//8k7Fjx/Liiy9y+vRpMjIyaNKkCT4+9vnGOXfuHAkJCfTu3Vv/mL+/Px07dmT37t2lJnW5ubn6kb0A6enpVo+1wuUegYCnQZcMmZtsHY0QoorKyNbx4ffJbNidSfVqzoy7P5Be7b3L3S/u9pZe9G7vxQffp9CusSdB/tJfWNgXu+hT5+zsTN++fUlOTsbNzY0mTZrQoUMHu03oABISEgCKDeMOCwvTLyvJnDlz8Pf319+aNGli1TgrXMEV0F4B92bg1gDyzoCitXVUQogq5tLVfJ6ak8D2g1k8/1AgX7wWweBuvmYPdBh7fyDOTrDw2yQLRSqE5dhFpQ6gWbNmnD17ltq1a9s6FKuaPHkyEydO1N+/dOlS5Urs8m4MdnFrCFo/IB/yL4CbbQa7COHI8vIVFEXBxVlTYl8vUbLTcXm89MEVvD2c+PSV6lQPttxXnb+PM+MfCOT1pdfZfjCLrq29LLZv4Vg2b97M5s2buXLlCjpd0QE0S5cutUlMdpPUzZo1i0mTJvH666/Ttm1bvL29iyz38/OzUWQlCw8PByAxMbHIpMGJiYm0atWq1O3c3d1xd3fX309LS7NajDZREKv+6xoNTjc+7PLPSlInhAFarcLxC3nsPZrNvmM5xCbkk5Gt6JcH+DgRHOhMrXBX6kW6Ua+mG03quOHhZr0GF0VRUBRwcnKchPKfkzm8uvgqNUNdmTM2hABfyzeRdm/rxeZ9WSxclUTrhh4yzUkVNGPGDGbOnEm7du2oXr263QycsZukbsCAAQAMGjSoyJOjKAoajQat1r6a8GrXrk14eDibN2/WJ3FpaWn89ddfPP3007YNzpbyY8HJH5z9/pt4uOCSbWMSwgbyCxSuJBeQnqnDyUmDqwu4uznh4abBw02DsxPEX9fy79lc9h7N5uCJXNKzdPh6OdGmkQe3t/Simr8zThq1Ync9TcuVpALOXc5n5z/Z5OQpuLpA09ru9OnkTa923ri5mvfFotMpnLmUz6GT6nxth0/nkleg8Fh/f4bd6Wc3X1yl2Xkoi9eXXqN5XXdmPhWCl4d1ki2NRsNzDwUyamY8761MYsrIanb/3AjLWrx4McuWLeOxxx6zdShF2E1St2XLFluHUExGRkaRufPOnTvHoUOHCAoKIioqiueff55Zs2ZRv359/ZQmERERDB482HZB21p+LLhGqf938gDnYMi/aNuYhKggefkKW/erU2YcPJlDfoHhbZw00KiWG/d296F9E08aRZc8rcbNtDqFuMQCDhzP4a9/s3nrqyQ+XZvCwC4+3NPV1+gO/EmpWk5dzON0XB7Hz+cRc1pNLN1cNTSt48YDffxIz9Sx9MdUfL2duKerr1H7tYUftqezcFUyXVp7MXl4NbMTXENCAlyY8EgQs5ZeJ8jPmafuDZAm8iokLy+P2267zdZhFGORpC42NpbIyMhiv1QURSEuLo6oqCiD++jWrZslQrGoffv20aNHD/39wr5ww4cPZ9myZfzf//0fmZmZjBkzhpSUFO644w42bNhQteeoK4gFl5v+3i41oUCSOlG5FWgV1m9LZ/mmNJLTdLSq784TgwKoV9MNfx8ntDq1cpebp5CTpyM3TyG/QCGsmgt1a7jhY2LznbOThlrVXalV3ZUhPXy5eCWfNVvS+e6PdFZsSqNnO2+6tPakbg03XF00ZOboSEnXculKAXFXCjh3KY/TF/O5nqq2gHh7aKgfqSaWrRt40Li2e5GkKC9fYfHqFG5r7klIoN3UAgDIztGxeE0KP+7M4N7uPowdGlhhzcU923mTkq6OsP3r32we7utH7w7euEhyV+mNHj2a5cuX22we3dJoFEVRDK9WNmdnZ+Lj4wkNDS3y+PXr1wkNDTWq6XT79u1lLi/pEmKVwcWLF4mMjCQuLq5yXPv1XCvwvA3CP1Tvxw1U/4380WYhCWEtiqLw15EcPlqTzKUrBdzZ2ZsHevsRFe5qk3gysnT8/GcGP+zIIP5a8TKhRgMhN/rl1Y90U/vmRbpSvZpLmYlQZraOx6ZfpmMzT156vJo1T8Ekh0/n8OZXSSSlavnfkAAGdvGxSTPoydg8vvwllV0x2YQFOfNQHz/63+Zj9WqhPah032FGeu655/jyyy9p0aIFLVq0wNW16Ht+wYIFNonLIj+5Cvu93SojI8PoqlX37t2LPXbzPu2tT50oRX4s+N00R59rTcjeY7t4hLCS8/H5fLQ6mb1Hc2jd0J1pTwRTt6abTWPy8XLiwT5+PNDbl4TrWi5dzSe/QL12qb+PMxHBLuVKNLw9nXj8Ln/e/zaZ+3v5UqeGbc8zNUPLJ2tT+HV3Jk1quzF3bAg1Q22TSAM0iHJj1v9COHspj282pvH+t8l89Wsq9/fyY1AXHzyt1LdP2E5MTIy+P/2RI0eKLLNl/0qzkrrC5kiNRsPUqVPx8vpvaLdWq+Wvv/4qcyTozZKTk4vcz8/P5+DBg0ydOpXZs2ebE6aoKLoMdcJhl8j/HnOpIQMlRKWzdms6H36fTFg1F2aOCeb2lp521VFeo9FQPdjFolN53HW7D99tTmfpj6nM+l+IxfZriuxcHeu3Z7ByUxo6ncLzDwVy1x0+ONvJ6Nw6NdyYOiqYkXfns2JjGkvWp7BiUxpDevhyb3fz58gT9sMexwGAmUndwYMHAbVSd/jwYdzc/vv15ubmRsuWLZk0aZJR+/L39y/2WJ8+fXBzc2PixIns37/fnFBFRSi4MemyS8R/j7mEgfaaOgGxRmZfF47v+z/S+PD7FIb08GXM4IAq0cQG4OqiYeTd/ryx7Dr/ns2laR13wxuZQVEUUjN0xF8vIPF6Af+czmXLviyycnT07+zDiIH+BFphuhJLqBnqyouPVePxu/xZ9Vsayzem8d3vadzbw5f7evgafWky4RguXlT7jdtD87NZSV1hpjpy5Ejee+89q8wlFxYWxokTJyy+X2EFBVfUf11u6lvpHAooamLnElbiZkJYUl6+QlKaFkWBQD8ni87jtv1gFh+tTuGhPr6MuTfQYvt1FD3bebFyUxqfrU9h/vOhZlUnFUUhI1sh/loBCddvvWlJSCogJ/e/Lt/BAc707+zNPd18Ca9mX4M1ShMW5MKzDwbx6J3+rPo9je83p7P6j3Tu7e7L/b0kuXNkOp2OWbNmMX/+fDIyMgDw9fXlhRdeYMqUKTg52aYqa5F3xueff272PmJiYorcVxSF+Ph45s6da3QTrrAxbaL6r/NNyVthIldwRZI6YRV5+Qr/nMrhryPqpL0XrxSgu5ELaDRQPdiFVg3cadfYk7aNyj9R7JEzubyx7Drd23ox+p4Ay52AA3Fy0jBqkD+vLr7GvmM5tG/iWeb6uXk6Ll8r4PK1AhKuFZCQpL3xr3o/M+e/pM3DTUN4NRfCqznTqoE74dW8b9xXm5F9PDV21cRtiiB/Z56+L5CH+vrx3e9prNmazpqt6Qzu6sP9vf3stuIoSjdlyhQ+++wz5s6dy+233w7Azp07ee2118jJybFZtzGLjH7NzMxk7ty5pV4u4+zZswb34eTkhEaj4dZwOnXqxNKlS2nUqJG5YdqlSjVyKHkxJI6DhnmgufHFmXcWztaFyN/Bu5dt4xOVQoFW4WRsHodO5nDoZC6HT+eSm68QGuhMh6aeNIx2IyxI/ZK8mqLldFweB47ncCGhACcndbLejk096NDUk7o1XY1KFGIT83n27URqVXdl3vjQKtPkWhJFURj/diJ5BQof/l94kek78vIVjp7LZf/xHA4cz+HEhTx9gu3qgpqgVXMhrDBZq+ZM2I3H/H2cHDZpM1VqhpbvN6ezdls6Oh0M7OLDg739jJ5f0J5Uqu8wE0RERLB48WIGDRpU5PH169fzzDPPcOmSbfqSW6RSN3r0aLZt28Zjjz1W7stlnDt3rsh9JycnQkJCqvacb45Gm6g2vWpuqoQUNsUWJNomJuHwtFqFk3F5/HMyl0Mnczh8JpfsXAUPdw0t6roz/G5/Ojb1oFb1UhK0zuo/iUkF7D2qVvS+3pjGkh9SCfR1omkdd/2tQZRbsYQt4XoBL71/hUA/Z17/X0iVTuhAHYQxdmggzy1I5M0vr9O3ozfnLudz4EQOMadyyclT8PN2ok1DD/p18qZWhCsRwS4E+Tk71OXGrMnfx5kn7gng/t6+rP4jnTVb0lm/PYMh3X0YOTAAVxd5nuxdUlJSicWmRo0akZSUZIOIVBap1AUEBPDzzz/rS5Dl8eWXX/Lggw8WuS4qqLM2r1y5kscff9zcMO2SzX7l5MepV3twKrv5xCQJYyH7T6h9qOjjJ7whZDYEPW+5Y4lKS1EUzlzMZ//xHH0Sl5WjJnHN67rTsr47rRp40CDKrdyTvOblKxw5k8uBEzkcPZvL8Qt5+stu1Y900yd52bk6PlmXgoerhncmhBEa5Bh9uSrC5r2ZvP11Ern5Cm6u6t+mbSMP2jb2oG4NV0ngTJCepWP1H+qAimZ13Jn5vxB8PB1jpGxVrdR17NiRjh07snDhwiKPjx8/nr1797Jnj22m8rJIUle7dm1++eUXGjduXO59WGICY0dkkzdE2rdw+SFwbw7Rf6mX87KES0NBlw6RG4s+fqY2+D0MIW9Y5jii0kq4XsCcZdc5fCYXDzcNTeu407qBOy0beNAwuvxJnCEFWoWzl/I5ciaXo+dy+fdsLolJ6mdOp2YevPhYNen3VIL0LB2pGVqqB7vYzbQijuzw6RymfHSVqHBX3hwXircDJHZVNanbtm0bd911F1FRUXTurDYH7N69m7i4OH755Re6dOlik7gsktR9/fXXrF+/ni+++KLIXHWmcHJyIjExkZCQovMf/fPPP/To0cOm5UxrqvA3hJIHZ+qBcyDkHobwTyBgtGX2faELuNaGiC+LPn6+o5pAVl9imeOISin+WgET3knEyQnGPxBEu8YeNm2GuppSgKJDqnOiQp24kMukhVeoVd2VuWPtP7GrqkkdwOXLl/nggw84fvw4AI0bN+aZZ54hIiLCwJbWY5FPq/nz53PmzBnCwsKoVatWsctlHDhwoNRtW7dujUajjmrq1asXLi7/haTVajl37hx33nmnJcIUAJm/Q0Ec1PwBrk6B1K8sl9QVJIJnx+KPu4RJnzpRptw8HdM+uYqzs4b3JoYSHGD7RCrEDmIQVU/DaHfmjQ/l/96/wv+9f4U3x4WafG1gUbqCggLeeOMNRo0aZXYSGhERYXcXR7DIp9bgwYPN3vbQoUP069cPHx8f/TI3Nzdq1arFfffdZ2aEQi99Nbg1APeW4HsvJDwFBdfAJdj8fWuvFJ3OpJBzqFoVFKIUH65O4WJiAYteDLOLhE4IW2pcy535z4Xx4sIrvPBeInOeCXXIkbH2yMXFhbfeeqtc/fRjYmJo1qwZTk5OxaZhu1WLFi3KG6JZLPLpOX36dLO3rVWrFg8++KBFRrvm5+eTkJBAVlYWISEhBAUFmb3PSkFRIONn8B+uTuDl3R/QQdY28DMzcdblgC615LnoXEIh64p5+xeV1uHTOfy4I4NnHwy0+bVThbAXDaLcWPB8KC8tusLo2fE8OTiAXu29q/zoa0vo2bMn27Zto1atWiZt16pVKxISEggNDaVVq1YlTsMG6ghxW40DsNhP4pSUFL7//nvOnDnDiy++SFBQEAcOHCAsLIwaNWoY3H748OFmHT89PZ2vv/6alStX8vfff5OXl4eiKGg0GmrWrEnfvn0ZM2YM7du3N+s4Di3/lDrtiFcP9b5rDXCtpY5YNTep095I2kqs1IVAwVXz9i8qpfwChQUrkmlcy41BXXwMbyBEFVK3phufTqnOeyuTeOvrJBZ9l0zHpp50auZBs3oeVK/mXGXm9rOk/v378/LLL3P48GHatm2Lt7d3keW3zj1X6Ny5c/p+/7dOw2YvLJLUxcTE0Lt3b/z9/Tl//jxPPvkkQUFBrFmzhtjYWL788kuD+9Bqtbzzzjt8++23xMbGkpeXV2R5WQMlFixYwOzZs6lbty4DBw7klVdeISIiAk9PT5KSkjhy5Ag7duygb9++dOzYkffff5/69eubfd4OJ2sH4ASet/33mOdtkL3L/H0X9plzCS2+zDkElEzQZYFT+QbSiMrp29/TiEvM5+OXw2UKDCFKEOjrzGtPhhCXmM+2A1nsOJTF3C+z1GV+TtSv6UZ0dVdqVXclurorNUJc8POuOhM5l8czzzwDqLnDrcqqskVHR+v/f+HCBW677bYi4wBA7bO3a9euIuuWxdItixZJ6iZOnMiIESOYN28evr6++scHDBjAI488YtQ+ZsyYwZIlS3jhhRd49dVXmTJlCufPn2fdunVMmzatzG337t3L9u3badq0aYnLO3TowKhRo1i8eDGff/45O3bsqJpJXfZOcG8Bzjddo9fzNkj7DnS54GTGBboNVeoAtFfBybgXuqj84q8V8NWvaQzt6SvNrkIYEBnmyqP9/Xm0vz9pmVqOnsvj37O5nL2Uz58x2Xz/RzrKTVfvCA5wITjAmZAAZ0KDXKgV7kLdmm7UqWHcVVQqs1uvelUePXr0KHEattTUVHr06FFm86s1WxYtktTt3buXjz/+uNjjNWrUICEhwah9fPPNN3z66afcddddvPbaazz88MPUrVuXFi1asGfPHp599tlSt12xYoVRx3B3d+d///ufUetWStm7wat30cfcWwP5kHcMPFqVf9/6Sl1I8WX6q0pcBVdJ6oRq0XfJ+Hs7MXyAv61DEcKh+Hk706mZJ52a/Td5fG6ejtjEAuKvFXAtRcu1FPXfqylajpzN5cqNeRcbRrvxv3sDaNlArtZkjsIk7FbXr18v1px7M2u3LFokqXN3dyctLa3Y4ydPniw271xpEhISaN68OQA+Pj6kpqYCcPfddzN16lRLhFm16TIh7yQEvVT0cXf1OSf3H/OSOm0iOFcDjWvxZTdX6oQA/ozJYvfhbF57MhhPD5muQQhzubs5UT/SjfqRJVe9s3J0xJzK5esNqXy8LoUPXgyr8IrdnDlzWLNmDcePH8fT05PbbruNN998k4YNG+rX6d69O9u2bSuy3VNPPcXixYv192NjY3n66afZsmULPj4+DB8+nDlz5hRrCi1LZmYm27ZtK7G7V1lFpCFDhgBqM+2IESOKXAVLq9USExPDbbfdVtrmVm9ZtEhSN2jQIGbOnMm3334LqCcbGxvLSy+9ZPR0JDVr1iQ+Pp6oqCjq1q3Lpk2baNOmDXv37i126bCS/PHHH4wbN449e/bg5+dXZFlqaiq33XYbixcvttkszzaXexhQwKNl0cedfcG1HuQcUkfFlldBojp1SUkkqRM3ycnTsejbZNo38aBLKwtepk4IUSovDyc6NfekQ1MP0jJ1NmmC3bZtG2PHjqV9+/YUFBTwyiuv0LdvX44ePVqkuvXkk08yc+bM/2K/6aIGWq2Wu+66i/DwcHbt2kV8fDyPP/44rq6uvPGGcVctOnjwIAMGDCArK4vMzEyCgoK4du0aXl5ehIaGlpnU+furLQuKouDr64un53+fYW5ubnTq1Iknn3yy1O1XrFhBQUGBwRjL3bKoWEBKSorSu3dvJSAgQHF2dlYiIyMVV1dXpWvXrkpGRoZR+3jppZeU2bNnK4qiKCtXrlRcXFyUevXqKW5ubspLL71kcPuBAwcqCxYsKHX5e++9pwwePNi4E6pAcXFxCqDExcVZ90BJHynKMRdF0WYXX3bxPkW50MO8/V96WFEudC99+QlfRbn2tnnHEJXCZz8kK33HX1DiEvNsHYoQwkyF32FHjx5VUlNT9becnByD2165ckUBlG3btukf69atm/Lcc8+Vus0vv/yiODk5KQkJCfrHPvroI8XPz0/Jzc01KuZu3bopTz75pKLVahUfHx/lzJkzSmxsrNK1a1dl9erVRu3jtddeMzq/uVVERIQyd+5cJTk5uVzbl8UiSV2hHTt2KB988IHy5ptvKr/99ptZ+9q9e7cyf/585YcffjBq/aioKOXo0aOlLj927JgSGRlpVkzWUGFJXfxTinK2ecnLrs5UlBOBiqLTlX//F3oqysUHS19+uo6iJP5f+fcvKoWLV/KUvuMvKEvWJ9s6FCGEBRR+h916mz59usFtT506pQDK4cOH9Y9169ZNCQ4OVqpVq6Y0bdpUefnll5XMzEz98qlTpyotW7Yssp+zZ88qgHLgwAGjYvb391eOHz+u/39h7rBnzx6lYcOGRu3DHPPmzVOioqIUHx8f5dlnn1XOnj1rsX1bdOr2O+64gzvuuMPk7fLz83nqqaeYOnUqtWvXBqBTp0506tTJ6H0kJiYWuzzZzVxcXLh6tQo3/+UcAvdWJS9zbwm6ZCi4CK6R5du/9gq4l9xHAFCbYKX5tcr78PsUAn2dGXann+GVhRAO4+jRo0XmpDXUbUqn0/H8889z++2306xZM/3jjzzyCNHR0URERBATE8NLL73EiRMnWLNmDaD2vw8LKzrLQuF9Ywdmurq64uSk9uUNDQ0lNjaWxo0b4+/vT1xcnFH7APj+++9LnYatrMujvvjii0ycOJFvv/2Wd999lwYNGnDPPffw4osv0rFjCZfaNEG5k7qFCxcyZswYPDw8WLhwYZnrltU+DeoTvHr1arMGRNSoUYMjR45Qr169EpfHxMRQvXr1cu/foSlatU+d3wMlL3e/cTmT3MPlT+oKEkuezqSQTEBc5e05ks3uw9lMHx2Mh5sMjhCiMvH19S3Wn70sY8eO5ciRI+zcubPI42PGjNH/v3nz5lSvXp1evXpx5swZ6tata5FYW7duzd69e6lfvz7dunVj2rRpXLt2ja+++qpIglmWhQsXMmXKFEaMGMH69esZOXIkZ86cYe/evYwdO9bg9s7Ozjz88MM8/PDD7NixgwULFnD77bfTsWNHJk2axODBg8vX77G8Jb5atWop165d0/+/tFvt2rWN2t/jjz9eZp84Q8aNG6c0a9ZMyc4u3mcsKytLadasmTJ+/Phy799aKqT5Nee4ohxDUTI2l7xcp7vR521u+favK1CUYxpFSf6k9HUuj1SUcx3Lt3/h8HLzdMqj0y4pL7yboOjMaeYXQtiV8nyHjR07VqlZs6ZRzY4ZGRkKoGzYsEFRFMs0v+7du1f5448/FEVRlMTERKVfv36Kr6+v0qZNG+XQoUNG7aNhw4bK8uXLFUVR9P3yCuMbO3asUfu41ZkzZ5Tx48crfn5+Sr169cq1j3JX6m6+RIYlLpdRv359Zs6cyZ9//lniZTsMVfteffVV1qxZQ4MGDRg3bpx+iPTx48f54IMP0Gq1TJkyxew4HVLuIfXfW0e+FtJo1KlNcg+Xb//a64BS+uhXUJdpt5W+XFRq3/+RRsL1Al7/X0iVn/hUiKpKURTGjx/P2rVr2bp1q767VVkOHToEoG9p69y5M7Nnz+bKlSv6iX9/++03/Pz8aNKkiVFxtGvXTv//0NBQNmzYYOKZqNOqFE5d4unpSXp6OgCPPfYYnTp1YtGiRaVuO336dFJTU0u8paSkkJWVxdmzZ02OCSx47ddCyo0prU394P7ss88ICAhg//797N+/v8gyjUZjMKkLCwtj165dPP3000yePLlIHP369eODDz4o1g5fZeQcApdIdR650rg3L//lwgqvJlHSJcIKSZ+6KutqcgFfb0hjSA9falUvvd+rEKJyGzt2LMuXL2f9+vX4+vrq+8D5+/vj6enJmTNnWL58OQMGDKBatWrExMQwYcIEunbtSosWajehvn370qRJEx577DHmzZtHQkICr776KmPHjjVq+jNLCQ8PJykpiejoaKKiotizZw8tW7bk3Llz+vyjNK+//joeHh6MGDGCNm3a4O/vj5+fH35+fvr/F06dYiqLJXWfffYZ77zzDqdOnQLUytvzzz/P6NGjjdreEtW+6OhofvnlF5KTkzl9+jSKolC/fn0CAwPN3rdDyz2kDoYoi3tzSFkKSn7JEwiXpaDwEmFlJHUuIaBLB10OOMlM5lXJJ+tS8HTT8JhcOUKIKu2jjz4C1AmGb/b5558zYsQI3Nzc+P3333n33XfJzMwkMjKS++67j1dffVW/rrOzMz/99BNPP/00nTt3xtvbm+HDhxeZ164krVu3NrrYVNYgh0I9e/bkhx9+oHXr1owcOZIJEybw/fffs2/fPv0ExaXZvHkz8+fPZ+nSpTz00ENMmjTJ6L58hlgkqZs2bRoLFixg/PjxdO7cGYDdu3czYcIEYmNjDT7ZN8vLy+PcuXPUrVvXpNmhbxYYGFiua6ZVWjmHIOCJstdxb4F6ubAT4G7ii0tbeIkwAwMl4Mb1X8s5GEM4nCNnctm8N4sXHw3Cx1MGRwhRlRmqYEVGRha7mkRJCgs4phg8eLBJ6xvyySef6K8hO3bsWKpVq8auXbsYNGgQTz31VJnb9ujRgx49enDixAkWLFhAx44d6dKlCy+++CK9evUyL7By9cS7RXBwsL7D4M2WL1+uVKtWzah9ZGZmKqNGjVKcnZ0VZ2dnfafDcePGKXPmzLFEmHbJ6gMl8hPUQRKp35e9XkGSul7KN6Yf4/q7inLcs+x57rL2qvvP3m/6/oVDKtDqlDFvXFb+Nyde0WplcIQQlVGFzbVaiV25ckWZPn26EhYWprRq1Ur5+uuvlYKCgnLtyyI/nfPz84t0PCzUtm1boy6HATB58mT++ecftm7diofHf81zvXv3ZtWqVZYIs2rK/Uf9t7RBEoWcA9V+d+UZLKG9oja9llXadrlRqZNpTaqMDbszORWXz7gHAnFyksERQgj7sn//fr7++mu+/vprDh48aHD9mJgYfXUuJiamzJspQkJCeO211zh+/DhDhgzh2WefpU6dOuU6J4s0vz722GN89NFHLFiwoMjjn3zyCcOGDTNqH+vWrWPVqlV06tSpSLt306ZNOXPmTJnb5ufn8+WXX+pjcXMr+YLGVVLOP+DkA65GvEDKOwK24ErZgyRArv9axWRk6fhsfQq9O3jRtE7FdV4WQghDrly5wkMPPcTWrVsJCAgAICUlhR49erBy5UpCQkJK3K5Vq1YkJCQQGhpKq1at0Gg0JTYpazQatFptqce/7777Shz5mp+fr99fSkpKuc6t3EndxIkT9f/XaDQsWbKETZs26a8C8ddffxEbG8vjjz9u1P6uXr2qH558s8zMTIOdGydNmkSfPn1QFIUXX3yR9957z4QzqeRyD6nJmsaIoqx7c0hbbvoxCit1ZXHyAo2X0UldaoaW2IR8Ll8rIC1TR1qmjtw89cV+61tIg1okdHZSX4tOGvV0nTQ33dfcuO+kISrMhXaNPXCXCXCt5uO1yeQVKDw5OMDWoQghRBHjx48nPT2df//9l8aNGwPqFTGGDx/Os88+y4oVK0rc7ty5c/qEz5zBnV5eXkRERBAQEFDmrTzKndTdWqps27YtgL6qFhwcTHBwMP/++69R+2vXrh0///wz48ePB/6bEmXJkiX6wRel0el06HQ6tFqtvjQqbsj5B7y6GreuewsoeBO0KeAcYPwxChLLvkRYIZfQEpO6rBwdu2KyiTmdy4X4fC4k5JOW+d/f0dNdg5+3Ex5umhKbeBWdgk4BReHGvzfu6/67X7isQKuQma0Q4OPE0J6+3NPNF2/pwG9Rh07m8POfmTz3YCAhARafNUkIIcyyYcMGfv/9d31CB9CkSRM++OAD+vbtW+p20dHRJf7fVF999VW5tzWk3J+4W7ZssWQcvPHGG/Tv35+jR49SUFDAe++9x9GjR9m1a5fB0TDz589n+fLlKIrC22+/bdG4HJouB/KOQ+B449Z3b67+m3sYvLoYfxztFXDuYXi9Wy4Vdi2lgDVbM/hxRzqZ2Qq1qrtSt6Yr7Rp7EBnuSnS4CzVCXHFztWx/rAvx+azZms4Xv6Sy6vd07uvpy5Duvvh4WT65S83QcvFKAZeu5HPpagGZOWr26ePlhJ+3M8EBzkSGWec8bSE3T8eC5Uk0r+vOwC4+tg5HCCGK0el0JV4r3tXVtczC0A8//GD0MQYNGlTi47GxsURFRRm9n0uXLhW5pq4hGqWkBmEbOXPmDHPnzuWff/4hIyODNm3a8NJLL9G8eXNbh2Y1Fy9eJDIykri4OGrWrGnZnefsh/PtIHo3eHYyvL6SBye8Iew9CHzG+OOc8IGQ1yFoQtnrxd0FGhfO8B3f/5HO5r2ZuLtquPsOH+7t7ktoUMVWda6mFLDqt3R+2pmBkxO0rOdOkJ8zGs1/VT8AnU5BQX3s5pvuxltHp+OW5QqZOQqxCfmkZvz3AVHN3xkfLyc0QEa2jtQMLfk3xhE5aSCsmgv1arrSMMqN+lFuNKrljq8VEk1r+mRdCqv/SOPTV6oTFS4TDQtR2Vn1O8xK7rnnHlJSUlixYgURERGAmjwNGzaMwMBA1q5dW+J2Tk5FP49v7VN3c1ex0vrUhYWFMXjwYEaPHl3q1Gupqal8++23vPfee4wZM8bgxRduZpFv0ZycHN5//322bNnClStXimW6xkzkB1C3bl0+/fRTS4QkQG16RfNfBc4QjRu4Ny4yWCI7R8fWA1mcjM0jN18pktAoCrhoMnmpUyZrdngSc+Xqf8mQUrTpU1FgaBNf/N1O89TyBEIDnXliUAB33eFjs/nLQgJcGHd/II/09ePX3RkcPp3L+fh89anQFN406r+oiRea//rq/bfOjWWAk5MGJ42G0EAn2jbyICrMhcgwVyJCXPB0L3qeiqKQnK7j4pV84hILiE3I51RcHis2pZGZo6DRQJ0arrSs70HL+u60qOeOv49zsfPQ6hSuJGn1+7l0NR+tFrw8NHh6OBEc4Ez1ai7UDHUh2IrNoYdO5rDqtzSeGOQvCZ0Qwm4tWrSIQYMGUatWLSIj1XlTY2Njad68OV9//XWp292c2/z++++89NJLvPHGG0Xm53311Vd54403St3H0aNHmT17Nn369MHDw4O2bdsSERGBh4cHycnJHD16lH///Zc2bdowb948BgwYYNK5WeQT/oknnmDTpk0MHTqUDh06lPvajlqtlrVr13Ls2DFAbeO+5557yj0JcZWX8ze4NQEnb8PrFnJvDrnqcOw//8ninRVJJKfrqFXdFU93DU5Omv8GHQDVvNWJh6+lVyM7V9EnQIUJkbPTf8lPrq4awT57mTkmmE7NPXFxto/mxiB/Z4bdWfFXO9BoNAT5ORPk50yLev89rtMpXLpawL9nc/nnVC67Y7JYs0W9rmCdCFfqRbrh4gJJqVrir2u5fDVfX/FzdYGIEFdcXSA7RyErR0dy+n8fRCEBzjSu7cZtzT25o5UXXh6WSajTMrW8sew6Leu582AfP4vsUwghrCEyMpIDBw7w+++/c/z4cUDNN0yZ+Pf5559n8eLF3HHHHfrH+vXrh5eXF2PGjNHnMbeqVq0aCxYsYPbs2fz888/s3LmTCxcukJ2dTXBwMMOGDaNfv37lvsKERZpf/f39+eWXX7j99tvLvY9///2XQYMGkZCQQMOGDQE4efIkISEh/Pjjj2WeoLXbqK3JqqXrc63Aox1UX2L8NtffhOtv8M3ZC3z2Qxq3tfBk3P2BhFcrJbHO+hNi74Dahw1fieL6PLj+BjRIMT4eAUBiUgH/nMol5lQOFxLyKdBCkJ8zYUHO1Ax1JTLMhZqhroQGOeN8y5xwefkKCUkFXIjP59i5XGJO53L0XB7eHhru7eHL0J6++HkXrwAaS6tVmPLRVY5fyOPTV8IJCZQfYUJUFY7U/Lp7926uX7/O3XffrX/siy++YPr06WRlZTF48GDef/99o64h6+npyd69e4vlJjExMXTs2JHs7GyLx28Us6dCVhSlcePGyj///GPWPjp16qQMHDhQSUpK0j+WlJSkDBo0SOncuXOZ24aGhipjxoxR/v7771LXSUlJUT755BOladOmynvvvWdWrIWmT59+o6Hxv1vDhg1N2ofVZuPWpivKMSdFSf7UtO3Sf1aUYygP/d+fyrKfUhRdWVeJUBRFSf1WvVJEQVLZ6ymKoiQvVdfV5ZoWk7C4hOv5yoffJyl3Phur3D0hVln2U4qSkaU1eT86nU55f9V1pdfYC8q+Y9lWiFQIYc8c6YoSd955pzJ37lz9/ZiYGMXV1VUZPXq0Mn/+fCU8PFyZPn26Ufvq0qWL0qdPHyUhIUH/WEJCgtK3b1+la9eulg7daBb5ST1//nxeeuklFi9eXO5hvocOHWLfvn0EBgbqHwsMDGT27NkGr+Nq7TbqsjRt2pTff/9df99umoqz9wE64wZI3GTjwdr0C4Ex/WPp0f02wxsUXAKNBzgFGF63cILigmvgGmFSXMKywoJcePq+QB7q68fKTWks35jK2q3pPNjHj8HdfIr1/yuJoigsWZ/Kmq0ZPPdQIG0beRjcRgghbOXQoUO8/vrr+vsrV66kQ4cO+r78kZGRTJ8+nddee83gvpYuXcq9995LVFSUvl9eXFwc9evXZ926ddYI3ygWyUDatWtHTk4OderUwcvLq9hQ4aSkJIP7aNCgAYmJiTRtWnS+sytXrlCvXr1StlJZu426LC4uLoSHh1t8v2bL2gROQeDW2PC6N+w9ms1bK73o+nQ1ujc5DDxkeKOCy+BSo+xLhBW6+aoSktTZhUBfZ56+L5ChvXz55tc0Pv8xhe83p3FPN196tfOiRmjJAx7Ss3S8tzKJP/Zl8fR9AdzT1beCIxdCCNMkJycTFhamv79t2zb69++vv9++fXvi4uKM2le9evWIiYnht99+0/fLa9y4Mb179y73uAJLsEhS9/DDD3Pp0iXeeOMNwsLCynVCc+bM4dlnn+W1117TX5Viz549zJw5kzfffJO0tDT9un5+JXfE9vT0ZOjQoQwdOrR8J1IOp06d0lcFO3fuzJw5c8rs35ebm0tubq7+fnp6uuWDUhRI+xZ8h4DGuL5SsQn5zPzsGu2beOIe0AtN1mZgtuENCy6Bi5EJmj6pu2Lc+qLChAS48PzDQTzYx49vNqay6rc0lv2USr2arjSv507tCDd8PDVkZCscv5DLtgNZKApMHVWNHu1MGIgjhBA2EhYWxrlz54iMjCQvL48DBw4wY8YM/fL09PQS568rjUajoW/fvmVOWFzRLDJQwsvLi927d9OypYGLxpfh5vlfCpPCwtBuvm/ommoV6ddffyUjI4OGDRsSHx/PjBkzuHTpEkeOHMHXt+TKxWuvvVbkRVTIop1Mcw7A+bYQuQm8+xhcPT1LxzNvJuDiomHRpDC8c5dCwlNQ/7rhK0vE9gDncKhR8mVVitBlwElfqP4N+D9i3LkIm8jJ07HnSA67Y7I4ei6Py9cKUBR11HPNUBdua+HJ0J5+BPmXf4CFEMLxOdJAiaeffpp//vmHN998k3Xr1vHFF19w+fJl/fXiv/nmG95991327t1r1P4yMzPZtm0bsbGx5OXlFVlmytxylmSRSl2jRo3MHulh6StUVISby7YtWrSgY8eOREdH8+233/LEE0+UuM3kyZOLXDf30qVLNGnSxLKBpX0LztXAy4irPAAffJdMSoaWjydXVy+Z5dIX0EHGj+D/WNkb518C9zbGxaXxVvvfGXn9V2E7Hm5OdG/jRfc2XoA6F152joKnh6bY6FohhHAEr7/+OkOGDKFbt274+PjwxRdf6BM6UPvJGVt1O3jwIAMGDCArK4vMzEyCgoK4du0aXl5ehIaGOnZSN3fuXF544QVmz55N8+bNi5UvS2suvVm3bt0sEYpNBQQE0KBBA06fPl3qOu7u7kWGS9/crGwRigLp34HPENAY/vMeOJHDpr8y+b/HgogIvrG+axR4doPUZWUndYpiWvOrRqM2wUpS53CcnTT4eEkyJ4RwXMHBwWzfvp3U1FR8fHxwdi7a0vDdd9/h42Pc5Q0nTJjAwIEDWbx4Mf7+/uzZswdXV1ceffRRnnvuOWuEbxSLJHV33nknQLGJ+0xtLs3JySEmJqbEq1KUdh01e5KRkcGZM2d47DED1S1ryj0A+WfB72OjVv/y51QaRLnRr9Mt/aICRkL8CMg7B261S95YexWUrNKXl0SSOiGEEDbk71/yZPNBQUFG7+PQoUN8/PHHODk54ezsTG5uLnXq1GHevHkMHz6cIUOGWCpck1gkqSur6fTw4cOlLrvZhg0bePzxx7l27VqxZfbUj+5mkyZNYuDAgURHR3P58mWmT5+Os7MzDz/8sO2CSvsWnIPBq7vBVf85lUPM6Vxefyq4+OAW36GQOAGSF0LYOyXvIP+s+q9rHePjcwmBAknqhBBCOC5XV1f9WIDQ0FBiY2Np3Lgx/v7+Ro+gtQaLJHW3Np2mp6ezYsUKlixZwv79+xk3bpzBfYwfP57777+fadOmFRlybM8uXrzIww8/zPXr1wkJCeGOO+5gz549hISE2CYgRYH0b8H3PqOaXr/+NY06Ea50bu5ZfKGTNwSOhaR3IHgqOJfwCyavHEmdc+h/yaAQQgjhgFq3bs3evXupX78+3bp1Y9q0aVy7do2vvvrKKtOnGcuiV1Lfvn07w4cPp3r16rz99tv07NmTPXv2GLVtYmIiEydOdJiEDtSJCy9fvkxubi4XL15k5cqV1K1b13YB5eyD/PPg+4DBVY+dy2X/8Rwe7e+HU2kd3wPHA1pIml/y8vyzalXQ2YRrfUrzqxBCCAf3xhtvUL16dQBmz55NYGAgTz/9NFevXuWTTz6xWVxmV+oSEhJYtmwZn332GWlpaTzwwAPk5uaybt06k0Z1Dh06lK1bt9o2KXJ0ziFQbRp4dTW46le/phIV5kKX1l6lr+QSCkEvQNLb4P8EuN1Skcs7Aa5lTwxdfJ/S/CqEEMJxKYpCaGioviIXGhrKhg0bbByVyqykbuDAgWzfvp277rqLd999lzvvvBNnZ2cWL15s8r4WLVrE/fffz44dO0ocQWur4cEOxa0WhBSfA+9Wp+Ly2HMkh8nDqxmenqLaZEj9AuJHQdTvRZt1cw6CVxfTYnQOAV0yKPmgMX6SRyGEEMIeKIpCvXr1+Pfff6lfv76twynCrKTu119/5dlnn+Xpp582+8RWrFjBpk2b8PDwYOvWrUU67ms0GknqLOibDalEhLjQs10ZVbpCTt4Q8TXE9oTE5yDsfdA4gS4L8o7faKI1gf6qEtfBxQ4vryaEEEKUwcnJifr163P9+nW7S+rM6lO3c+dO0tPTadu2LR07dmTRokUljl41xpQpU5gxYwapqamcP3+ec+fO6W9nz0rHeks5czGP7QezeaSvH87ORs475tUNwhdDyodwcRBk74bUrwCtUaNsi3BR+yBQcMm07YQQQgg7MXfuXF588UWOHDli61CKMCup69SpE59++inx8fE89dRTrFy5koiICHQ6Hb/99ptJ1zXNy8vjwQcfLHK5MGF5X/2aSvVgF/reOi+dIQFPQo0fIDcGLtwGif8DzzvArYFp+3G9Madd3jnTthNCCCHsxOOPP87ff/9Ny5Yt8fT0JCgoqMjNViwypYm3tzejRo1i1KhRnDhxgs8++4y5c+fy8ssv06dPH3744QeD+xg+fDirVq3ilVdesURIogQnY9Uq3YuPBeFibJXuZr4DwWcA5OwHXSZ4dlKvEmEK52rg5KOO0hVCCCEc0LvvvmvrEEpkkaTuZg0bNmTevHnMmTOHH3/8kaVLlxq1nVarZd68eWzcuJEWLVoUGyixYMECS4dapeh0CgtXJVGruit9OphYpbuZxhk8O5ixvUat1uVLpU4IIYTjyc/PZ9u2bUydOpXatf+/vXsPaupO3wD+JCGBgKBcCgEBRZSqoyBeu1pQBmbE7dRlV61rddHKuv4cqK1MS9GO2tXuVq07WuutSnXXtdNqF1c7jC0qq3atXWu5SREUHFwVuYk3Lmog5/z+cMiaIpKEwEkOz2cmMyQ5Oed9Upu8+Z7L14IZlXqAzZu6NiqVCgkJCUhISDBr+aKiIkRGRgJAu33U7WY7IIsdPNGAixV6bFrma90onS2xqSMiIgelVquRmZmJlStXSl1KO93W1FnqWVONUdccOtmA3UfuYu5UD0QMcZG6HEAdCjR2vkueiIjIHiUkJODw4cNYtmyZ1KWYsJumrk15eTmuXLmC6OhoaLVaiKLIkTorNT8UsP0fd3D0bBNmxbrjtZefPolxj3OJBO5sAgx3AVU/qashIqIetG3bNnz44Yeorq5GREQEPv74Y4wf34XDeiQwZMgQrFmzBt999x3GjBkDNzfTw5qkugybQhRFUZIt/0x9fT1eeeUVnDx5EgqFAmVlZRg0aBAWLlwIT09P/OUvHUxV5eBu3LiBoKAgXL9+HYGBgTZZp0EQkfNDE/Zm3cO9JgEpMz3xy0l9bLJum3hUAlQMB4JOAG6xUldDRERWsvQ77MCBA0hMTMTOnTsxYcIEbN68GV9++SUuXboEX1/fHqjYNp51LJ1CoZDsUmx209QlJiaitrYWGRkZGDZsGAoLCzFo0CBkZ2cjNTUVxcXFUpfYLWzZ1NXebsWRbxtx4ocm1N01YFKEFktmeCLAx84GZEUDUB4AOA8H3H4pdTVERL2Lkz/Qd55NVtX2HXbx4kX079/f+LizszOcnZ3bLT9hwgSMGzcOW7duBQAIgoCgoCC8/vrrSE9Pt0lNvZndfNsfO3YM2dnZ7RqbIUOG4L///a9EVTmW+vsGZJ1pxJQxrnhpUh+EBWukLunpFCrAZw1w613gYYHU1RAR9S4uY2zW1LX5+Vzvq1evxnvvvWfymF6vR25uLpYvX258TKlUIi4uDt9//71N6+kOqampWLt2Ldzc3JCamtrhcgqFQrK9i3bT1DU1NcHVtf20Vbdv335qt0/tDR2gwZcf9IdG7QDHIHoufnwjIiKH97SRup+7desWDAYD/Pz8TB738/NDaWlpt9fYVfn5+WhpaTH+3REpzwOwm6YuKioK+/btw9q1awE8flMEQcCGDRsQExMjcXWOQaFQQKPufDkiIiJbcnd3h4eHh9RldKsnr9Jhr1fssJumbsOGDYiNjcWPP/4IvV6PtLQ0FBcX4/bt2/juu++kLq/bCIIAAKiqqpK4EiIiIsu0fXe1fZc9i4+PD1QqFWpqakwer6mpgU6n65b6ehu7aeo8PDxQUlKCHTt2wN3dHY2NjfjNb36D5ORk43CnHLX943a007mJiIja1NTUIDg4+JnLaDQajBkzBjk5OcaJCQRBQE5ODlJSUnqgSvmzm7NfVSoVqqqq2p3SXF9fD19fXxgMBokq616tra3Iz8+Hn58flEqlZHU0NDRg+PDhuHjxItzd3SWroyuYwT4wg/2QQw5msA8dZRAEATU1NYiMjISTU+fjRAcOHMD8+fPxySefYPz48di8eTMOHjyI0tLSdsfakeXsZqSuo96ysbERLi52MAtCN3FycsK4ceOkLgP3798HAPTv399hj4tgBvvADPZDDjmYwT48K0NnI3RPmj17Nurq6rBq1SpUV1dj1KhR+Oabb9jQ2YjkTV3bacEKhQKrVq0yOQPWYDDg3LlzGDVqlETVERERkS2lpKRwd2s3kbypazstWBRFFBUVQaP537XVNBoNIiIi8NZbb0lVHhEREZFDkLypazst+LXXXsNHH33ksEPTjs7Z2RmrV6926GsCMoN9YAb7IYcczGAf5JChN7CbEyWIiIiIyHrSnW5JRERERDbDpo6IiIhIBtjUEREREckAmzoiIiIiGWBT56AWLFhgnGalu+zatQtTpkyBh4cHFAoF7t69a/L81atXkZSUhJCQEGi1WoSGhmL16tXQ6/VmrV8OGQD7yAEA06dPR3BwMFxcXODv74/f/e53uHnzplnrZwbzMEPn7CVDm0ePHmHUqFFQKBQoKCgwextyyCGHDGQZNnXUoebmZsTHx2PFihVPfb60tBSCIOCTTz5BcXExNm3ahJ07d3a4vBTkkAHoPAcAxMTE4ODBg7h06RIyMzNx5coVzJw5swerfDZmsA+9JUObtLQ0BAQE9EBVlpNDDjlkkBWRHNL8+fPFX/3qV6IoiuLXX38tTpo0Sezbt6/o5eUlvvTSS2J5eblx2YqKChGAmJmZKU6ZMkXUarVieHi4ePbsWbO2dfLkSRGAeOfOnU6X3bBhgxgSEtJrMthzjiNHjogKhULU6/XMwAy9MsPRo0fFoUOHisXFxSIAMT8/36z1yiWHHDKQZThSJwNNTU1ITU3Fjz/+iJycHCiVSvz617+GIAgmy7377rt46623UFBQgLCwMMyZMwetra02reXevXvw8vKy+HVyyADYT47bt2/js88+w8SJE6FWq5mBGXpdhpqaGixatAh///vfTaaftIYccsghA5lB6q6SrPPkL7Cfq6urEwGIRUVFoij+7xdYRkaGcZm2X0slJSWdbsvcX/RlZWWih4eHuGvXrl6Twd5ypKWlia6uriIA8YUXXhBv3brFDMzQ6zIIgiDGx8eLa9euNdmOtSN1jppDDhnIMhypk4GysjLMmTMHgwYNgoeHBwYOHAgAuHbtmsly4eHhxr/9/f0BALW1tTapobKyEvHx8Zg1axYWLVpk8evlkAGQPsfbb7+N/Px8HDt2DCqVComJiRAtnDSGGZgBcOwMH3/8MRoaGrB8+XKr1/EkOeSQQwbqnORzv1LXvfzyyxgwYAB2796NgIAACIKAESNGtDuD88ldJwqFAgDaDb1b4+bNm4iJicHEiROxa9cuq9YhhwyA9Dl8fHzg4+ODsLAwDBs2DEFBQfjPf/6DX/ziF8zADL0mw7/+9S98//337eYpHTt2LObOnYu//e1vFq1PDjnkkIE6x6bOwdXX1+PSpUvYvXs3oqKiAABnzpzpse1XVlYiJiYGY8aMwd69e6FUWj74K4cMgPQ5fq7tg/jRo0dmv4YZbI8Zej7Dli1b8P777xvv37x5E1OnTsWBAwcwYcIEi9YlhxxyyEDmYVPn4Dw9PeHt7Y1du3bB398f165dQ3p6uk3WXV1djerqapSXlwMAioqK4O7ujuDgYHh5eaGyshJTpkzBgAEDsHHjRtTV1Rlfq9PpelUGqXOcO3cO58+fx4svvghPT09cuXIFK1euRGhoqEUjK8zADHLIEBwcbLJ8nz59AAChoaEIDAy0aFtyyCGHDGQeHlPnoARBgJOTE5RKJb744gvk5uZixIgRWLZsGT788EObbGPnzp2IjIw0Hl8WHR2NyMhIfPXVVwCA48ePo7y8HDk5OQgMDIS/v7/x1lsy2EsOV1dXHDp0CLGxsXj++eeRlJSE8PBwnD59ut1uD2ZgBrlnsAU55JBDBrKMQrT0yFeyC/Hx8Rg8eDC2bt0qdSlWk0MGQB45mME+MIP9kEMOOWQgy3CkzsHcuXMHWVlZOHXqFOLi4qQuxypyyADIIwcz2AdmsB9yyCGHDGQlKa+nQpZLSEgQ+/fvL65YsUIUBEHqcqwihwyiKI8czGAfmMF+yCGHHDKQdbj7lYiIiEgGuPuViIiISAbY1BERERHJAJs6IiIiIhlgU0dEREQkA2zqiIiIiGSA04QRWUAQhHYTYBMRUdeo1WqoVCqpy3B4bOqIzKTX61FRUWGc3JyIiGynX79+0Ol0UCgUUpfisNjUEZlBFEVUVVVBpVIhKCgISiWPXCAisgVRFNHc3Iza2loAsGjubTLFpo7IDK2trWhubkZAQABcXV2lLoeISFa0Wi0AoLa2Fr6+vtwVayUONxCZwWAwAAA0Go3ElRARyVPbD+aWlhaJK3FcbOqILMBjPYiIugc/X7uOTR0RERGRDLCpIyIiIhPXrl1Dnz59UFRUJHUpZAGeKEFEREQmAgICUFBQgODgYKlLIQuwqSMiIiITTk5OGDx4sNRlkIW4+5WIiIhIBtjUEREREckAmzoiIiIyKigowG9/+1vodDpoNBqEhoZizZo1aG1tlbo06gSbOiKibjRw4EBs3rxZ6jJkKzs7GwqF4pm3Y8eOSV2mw9izZw/Gjx8PPz8/ZGVloaSkBCtXrsTmzZuRlJQkdXnUCTZ1RDK2YMEC4xebWq1GSEgI0tLS8PDhQ+MyCoUChw8flq5IB/Lk+6nRaDB48OBORzDOnz+PP/zhDz1YZe8SHR2Nqqoq483b2xsrV640eSw2NlbqMh3CqVOnsGjRImRkZOCjjz7C2LFjERoaigULFmD9+vXYt28fysvLpS6TnoFnvxLJXHx8PPbu3YuWlhbk5uZi/vz5UCgUWL9+vdSlOaS29/PRo0c4evQokpOToVarsXz5cpPl9Ho9NBoNnnvuOYkq7R20Wq1x3tDKykrU19cjKioKOp1O4soczxtvvIFp06YhMTGx3XOTJ08GABQWFvKsWDvGkToimXN2doZOp0NQUBASEhIQFxeH48ePS12Ww2p7PwcMGIAlS5YgLi4OX331FRYsWICEhAT86U9/QkBAAJ5//nkA7Xe/3r17F4sXL4afnx9cXFwwYsQIZGVlGZ8/c+YMoqKioNVqERQUhKVLl6KpqamnYzqk/Px8AMDo0aMlrsTx5Ofn48KFC0hOTn7q8w8ePADw+FInZL/4X4fIWi3NwO3Snt+u11BA7WrVS3/66SecPXsWAwYMsHFRtnHv3j00Njaif//+AICqqio4OzvDy8sLLS0tqKqqgq+vL1xcXHD//n3cv38fgYGBAICamho4OTnB29sbBoMBlZWVeO6556DVatHQ0IC7d+8iKCgIAFBbWwulUgkfH58u16zValFfXw8AyMnJgYeHR4dNsyAImDZtGhoaGrB//36Ehobi4sWLUKlUAIArV64gPj4e77//Pvbs2YO6ujqkpKQgJSUFe/fu7XKtlnqoF3CtumcPjg/WOcFFY914Q15eHoKCguDt7W3jqrpIaAb0EnxWaIYCSvM+KwoKCgAAo0aNeurzeXl5AIDw8PAulRQXF4cdO3ZgyJAhXVoPPR2bOiJr3S4F9o/p+e3OywX8zB+JyMrKQp8+fdDa2opHjx5BqVRi69at3Vig9f7973/j7Nmz+POf/wwA+PTTTxEWFoZXXnkFd+7cwbp16/Dmm28iLCwM586dQ3Z2NjZu3AgA2LdvH/z9/TFv3jw0NjZi3bp1WLJkCUaOHIm8vDxkZmZiy5YtAIDPP/8c7u7uWLhwodW1iqKInJwcZGdn4/XXX0ddXR3c3NyQkZEBjUbz1NecOHECP/zwA0pKShAWFgYAGDRokPH5Dz74AHPnzsWbb74JABgyZAi2bNmCyZMnY8eOHXBxcbG6Xmtcq27F/62r7tFt7kzXISz46e9fZ/Ly8uxzlE5fClyV4LNiYC7gYt77odfrAaDDf2Pbt29HdHQ0QkJC2j1nMBiMP0w6U1ZWhtDQULOWJcuxqSOyltfQxw2WFNu1QExMDHbs2IGmpiZs2rQJTk5OmDFjRjcV1zVRUVGIjIw03k9KSoKzszMAwNPTE+np6fD19QUATJgwAcOGDTMum5iYaNw11KdPH6SnpxuPZxs9erRJ8zRnzhwoldaNBrU1yS0tLRAEAa+++iree+89JCcnY+TIkR02dMDj0ZDAwEBjQ/dzhYWFuHDhAj777DPjY6IoQhAEVFRUmOTtCcE6J+xM79lj04J11n8t5eXl4fe//327x/fv348tW7bgwYMHCA4OxqFDh4z/rnqEZujjBqunacz/rIiIiAAAnD59GgkJCSbPbdy4ESUlJThz5ozxsenTpyMwMBDnz5/H4sWL4eLi8tT3uLi4GAsXLsSDBw8we/Zs6HQ6q//fo86xqSOyltrVohEzqbi5uRkPbN6zZw8iIiLw6aef2uXlCfr27Yu+ffsa7/v7+xv/VqvVJvNQenh4wMPDw3jfz8/P+LdKpTJZ1t3dHe7u7sb7bY2hNdqaZI1Gg4CAAJNjjNzc3J752rYD+jvS2NiIxYsXY+nSpe2ek2IOTheN0upRs55269YtXL9+/akjddOmTcO8efMAAIsWLcKpU6cwderUnitO6Wr2iJlUXnjhBUydOhXJycloaWnB2LFjUVNTg4yMDHzxxRf45z//abJrtqioCLGxsdi+fTsAoL6+vt17PGXKFMyePRuff/45Ro4ciYSEhC7vvqVnY1NH1IsolUqsWLECqampePXVVzttMqi9J5tkS4WHh+PGjRu4fPnyU0frRo8ejYsXL/LsQiu0HfP186ZOFEXs3r0bmZmZ0Ov1uH79urH5IFOHDh3CH//4R7z99tu4ceMGDAYD4uPjcfnyZQQEBBiXa2hogMFgwBtvvAGg4/f48OHDmDx5MkaOHAkAGDZsmMl6yPY4BkrUy8yaNQsqlQrbtm0zPlZRUYGCggKTG8+4tL3JkycjOjoaM2bMwPHjx1FRUYGvv/4a33zzDQDgnXfewdmzZ5GSkoKCggKUlZXhyJEjSElJkbhy+5efnw8/P792TcNf//pXlJaW4ttvv0VhYSE8PT0xfPhwiaq0b66urli/fj2uXr2K1tZWJCUl4fLly3B1NT3Zori4GBMnTjTe7+g9LioqMhndy83N5UhdN2NTR9TLODk5ISUlBRs2bDA2bqmpqYiMjDS5tV0egmwrMzMT48aNw5w5czB8+HCkpaXBYDAAeDySd/r0aVy+fNl4fOGqVas4umGGd955B9XV7U/qKC4uxqRJk6DVarFt2zY0Nzfz2oFm2rZtGxYuXNjus6CoqMg4+gZ0/B57eXnhp59+AgAcO3YMJ06cYFPXzRSiKIpSF0Fk7x4+fIiKigqEhIT0+BmIRGS9wsJCzJw5E97e3njxxRdx4cIFThvWRUuXLkVcXBymT58OoOP3uLa2FtOmTYMgCBgxYgTOnDmDioqKDtfLz9muY1NHZAZ+2BARdS9+znYdd78SERERyQCbOiIiIiIZYFNHREREJANs6oiIiIhkgE0dkQV4XhERUffg52vXsakjMkPbZNVtk14TEZFtNTc3A3g8JSBZh9OEEZnByckJrq6uqKurg1qt5oTUREQ2IooimpubUVtbi379+hl/RJPleJ06IjPp9XpUVFRAEASpSyEikp1+/fpBp9NBoVBIXYrDYlNHZAFBELgLlojIxtRqNUfobIBNHREREZEM8MAgIiIiIhlgU0dEREQkA2zqiIiIiGSATR0RERGRDLCpIyIiIpIBNnVEREREMsCmjoiIiEgG/h/dBscUe9szmAAAAABJRU5ErkJggg==",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "q(s,a) = \n",
+ "[[0. 0. ]\n",
+ " [1.936 1.398]\n",
+ " [0.594 0. ]]\n",
+ "-------------------------------------------------------------------\n",
+ "Episode number: 2, starting day: 67.8 (from beginning of the year)\n"
+ ]
+ },
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "q(s,a) = \n",
+ "[[0. 0.17 ]\n",
+ " [2.414 2.116]\n",
+ " [1.491 0.594]]\n",
+ "-------------------------------------------------------------------\n",
+ "Episode number: 3, starting day: 0.9 (from beginning of the year)\n"
+ ]
+ },
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "q(s,a) = \n",
+ "[[1.594 1.141]\n",
+ " [2.411 2.221]\n",
+ " [1.491 0.594]]\n",
+ "-------------------------------------------------------------------\n",
+ "Episode number: 4, starting day: 29.9 (from beginning of the year)\n"
+ ]
+ },
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "q(s,a) = \n",
+ "[[2.411 2.219]\n",
+ " [2.382 2.409]\n",
+ " [1.491 0.594]]\n",
+ "-------------------------------------------------------------------\n",
+ "Episode number: 5, starting day: 19.8 (from beginning of the year)\n"
+ ]
+ },
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "q(s,a) = \n",
+ "[[2.411 2.219]\n",
+ " [2.833 2.967]\n",
+ " [2.287 1.918]]\n",
+ "-------------------------------------------------------------------\n",
+ "Episode number: 6, starting day: 11.0 (from beginning of the year)\n"
+ ]
+ },
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "q(s,a) = \n",
+ "[[2.411 2.219]\n",
+ " [4.677 4.24 ]\n",
+ " [2.367 1.918]]\n",
+ "-------------------------------------------------------------------\n",
+ "Episode number: 7, starting day: 45.2 (from beginning of the year)\n"
+ ]
+ },
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "q(s,a) = \n",
+ "[[2.452 3.17 ]\n",
+ " [4.847 4.609]\n",
+ " [2.432 1.997]]\n",
+ "-------------------------------------------------------------------\n",
+ "Episode number: 8, starting day: 362.2 (from beginning of the year)\n"
+ ]
+ },
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "q(s,a) = \n",
+ "[[2.452 3.271]\n",
+ " [5.66 5.162]\n",
+ " [2.737 2.014]]\n",
+ "-------------------------------------------------------------------\n",
+ "Episode number: 9, starting day: 72.3 (from beginning of the year)\n"
+ ]
+ },
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "q(s,a) = \n",
+ "[[2.452 3.917]\n",
+ " [6.03 6.114]\n",
+ " [2.737 2.014]]\n",
+ "-------------------------------------------------------------------\n",
+ "Episode number: 10, starting day: 357.8 (from beginning of the year)\n"
+ ]
+ },
+ {
+ "data": {
+ "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnUAAAHVCAYAAACXAw0nAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAACt+0lEQVR4nOzdd3QUZRfA4d+mJ6QDSQiE0HvvSO+CgogoKgqiiJ/SiwoqVZoooIiKUkVpKsWGgCJVilQR6TUBklDSe7I73x8DCyFtN9lsy33O2UN2ZmfmnWF35s6dt2gURVEQQgghhBA2zcHSBRBCCCGEEIUnQZ0QQgghhB2QoE4IIYQQwg5IUCeEEEIIYQckqBNCCCGEsAMS1AkhhBBC2AEJ6oQQQggh7IAEdUIIIYQQdkCCOiGEEEIIO+Bk6QIUdzqdjhs3buDl5YVGo7F0cYQQQgiTUhSFhIQEgoODcXCQXFKRUqzI7t27lf79+ystWrRQrl27piiKoqxcuVLZs2ePRcrz+eefK3Xr1lW8vLwULy8vpUWLFsrmzZv181NSUpQ33nhD8ff3V0qUKKH06dNHiYyMNGob4eHhCiAveclLXvKSl12/wsPDTX2ZFg+xmkzd+vXrefHFF+nfvz/Hjh0jLS0NgLi4OGbOnMnmzZvNXqZy5coxe/ZsqlatiqIofP311zzxxBMcO3aM2rVrM3r0aH799Ve+//57fHx8GDZsGH369OGvv/4yeBteXl4AhIeH4+3tXVS7IoQQQlhEfHw8ISEh+uudKDoaRVEUSxcCoGHDhowePZoBAwbg5eXFP//8Q6VKlTh27Bjdu3cnMjLS0kUEwN/fnw8//JC+fftSunRpVq9eTd++fQE4c+YMNWvWZP/+/bRo0cKg9cXHx+Pj40NcXJxJgrrk5GR0Ol2u8zUaDR4eHvKo10wURSEpKcnSxRBGcHBwwMPDw9LFKDa0Wi0pKSmWLoZ4iIeHh8kelZr6OidyZzWZurNnz9K2bdts0318fIiNjTV/gR6i1Wr5/vvvSUpKomXLlhw5coSMjAw6d+6s/0yNGjUoX758nkFdWlqaPgsJ6pfdlJYsWUJMTEyen+ncuTOtWrUy6XZFzrZs2cLff/9t6WIIIz377LNUr17d0sUoFlavXs2lS5csXQzxkOHDh+Pv72/pYggjWU1QFxQUxIULF6hQoUKW6Xv37qVSpUqWKRTw77//0rJlS1JTU/H09GTjxo3UqlWL48eP4+Ligq+vb5bPBwYG5plVnDVrFlOnTi2y8vbs2ZOMjIxc5//yyy8kJiYW2fZFVklJSQQGBtKxY0dLF0UYaM2aNZJdNaPExESqVq1KkyZNLF0U8QBPT09LF0EUgNUEda+++iojR45k2bJlaDQabty4wf79+xk3bhwTJ060WLmqV6/O8ePHiYuL44cffmDgwIHs2rWrwOubMGECY8aM0b+/V9fAVCpWrJjnfFdXV5NtS+RPURRKlChBtWrVLF0UYQQrqZVSbPj5+clvRAgTsJqgbvz48eh0Ojp16kRycjJt27bF1dWVcePGMXz4cIuVy8XFhSpVqgDQuHFjDh06xCeffEK/fv1IT08nNjY2S7YuKiqKoKCgXNfn6upq0cBKo9HIBcvMpP6iELlTFEV+I0KYiFV0GKPVatmzZw9Dhw4lOjqakydPcuDAAW7dusX7779v6eJlodPpSEtLo3Hjxjg7O7N9+3b9vLNnzxIWFkbLli0tWML8SVBnPnKsbY/c+JiXHGshTKdIMnWpqam4ubkZ/HlHR0e6du3K6dOn8fX1pVatWkVRLKNNmDCB7t27U758eRISEli9ejU7d+5k69at+Pj48MorrzBmzBj8/f3x9vZm+PDhtGzZ0uCWr5Ygd8TmJVkI2ySBhnnJb0QI0zBZpk6n0/H+++9TtmxZPD099a2ZJk6cyNKlS/Ndvk6dOlbXAurmzZsMGDCA6tWr06lTJw4dOsTWrVvp0qULAPPnz+fxxx/nqaeeom3btgQFBbFhwwYLlzpvkoUwP7lg2Rb5/zIvufERwnRMFtRNnz6dFStWMGfOHFxcXPTT69Spw5IlSwxafty4cfzyyy9EREQQHx+f5WUJS5cu5cqVK6SlpXHz5k3++OMPfUAH4ObmxmeffUZ0dDRJSUls2LAhz/p0oviRANr2yI2PEMJWmezx68qVK/nqq6/o1KkT//vf//TT69evz5kzZ/JdvkePHgD06tUry13bvbs4rVZrqqIWa3LBMj/JQtge+Y2Yj2TqhDAdkwV1169f17cSfZBOp8uz37R7duzYYaqiiHzIBct85FjbHgkwzEt+I0KYjsmCulq1arFnzx5CQ0OzTP/hhx9o2LBhvsu3a9fOVEUReZALlnkpimKyoXaE+UigYV5yXhLCNEwW1E2aNImBAwdy/fp1dDodGzZs4OzZs6xcuZJffvkl3+V3796d5/ychhATxpPHr+YnFyzbIv9f5iWPX4UwHZMFdU888QQ///wz06ZNo0SJEkyaNIlGjRrx888/Z2lckJv27dtnm/bgD13q1JmOBHXmI8fa9siNj3nJsRbCdEzaT12bNm34/fffC7Tsw4PQZ2RkcOzYMSZOnMiMGTNMUTyBZCEsQY657ZFAw7zkNyKEaZgsqBs8eDAvvPBCjhk3Q/j4+GSb1qVLF1xcXBgzZgxHjhwpZAnFPXLBMh851rZHAgzzkt+IEKZjshrct27d4tFHHyUkJIQ333yT48ePm2S9gYGBnD171iTrEnLBMjepL2R75PGr+clvRAjTMFmm7scffyQmJobvv/+e1atXM2/ePGrUqEH//v15/vnnqVChQp7LnzhxIst7RVGIiIhg9uzZNGjQwFTFLPbk5Gl+csyFyJ3c+AhhOiatU+fn58eQIUMYMmQI165dY82aNSxbtoxJkyaRmZmZ57INGjTI8Q65RYsWLFu2zJTFLPYkC2E+cqxtj2TqzEuOtRCmY9Kg7p6MjAwOHz7MwYMHuXLlCoGBgfkuc/ny5SzvHRwcKF26NG5ubkVRxGJLLljmJ1kI2yO/EfOS34gQpmHSXlF37NjBq6++SmBgIC+99BLe3t788ssvXLt2Ld9ld+3aRVBQEKGhoYSGhhISEoKbmxvp6emsXLnSlMUs1uTkaV7yaMn2yP+XeclvRAjTMVlQV7ZsWXr06MHt27f56quviIqKYtmyZXTq1MmgH+ygQYOIi4vLNj0hIYFBgwaZqpgCyUKYkxxr2yPZbCGErTJZUDdlyhQiIiLYuHEjffv2xdXV1ajlc7tbu3btWo7dnZjDrFmzaNq0KV5eXgQEBNC7d+9sLXHbt2+PRqPJ8vrf//5nkfIaQi5Y5idZCNsjvxHzkUydEKZjsjp1r776aoGWa9iwoT4Y6tSpE05O94uk1Wq5fPkyjz76qKmKaZRdu3YxdOhQmjZtSmZmJu+88w5du3bl1KlTlChRQv+5V199lWnTpunfe3h4WKK4wgpJcGB7JMAwL/mNCGE6hQrq+vTpw4oVK/D29qZPnz55fnbDhg05Tu/duzcAx48fp1u3bnh6eurnubi4UKFCBZ566qnCFLPAtmzZkuX9ihUrCAgI4MiRI1nGovXw8CAoKMjcxSsQydSZnwQJtkV+I+YnvxEhTKNQQZ2Pj4/+x+jt7V2gH+bkyZMBqFChAv369TNJa9eMjAwiIyNJTk6mdOnS+Pv7F3qdgL7O38PrW7VqFd9++y1BQUH07NmTiRMn5pqtS0tLIy0tTf8+Pj7eJGUzlJw8zUseLQmRN/mNCGE6hQrqli9frv97xYoVhSrIwIEDC7V8QkIC3377LWvXruXvv/8mPT1df7IoV64cXbt2ZciQITRt2rRA69fpdIwaNYpWrVpRp04d/fTnn3+e0NBQgoODOXHiBG+//TZnz57NNTM5a9Yspk6dWqAymIpkIcxHjrXtkUydecmxFsJ0TNZQomPHjsTGxmabHh8fT8eOHfNdXqvV8tFHH9GsWTOCgoLw9/fP8srLvHnzqFChAsuXL6dz585s2rSJ48ePc+7cOfbv38/kyZPJzMyka9euPProo5w/f97o/Rs6dCgnT55k7dq1WaYPGTKEbt26UbduXfr378/KlSvZuHEjFy9ezHE9EyZMIC4uTv8KDw83uiyFIRcsIfInvxHzkkydEKZhsoYSO3fuJD09Pdv01NRU9uzZk+/yU6dOZcmSJYwdO5b33nuPd999lytXrrBp0yYmTZqU57KHDh1i9+7d1K5dO8f5zZo14+WXX2bRokUsX76cPXv2ULVqVcN2DBg2bBi//PILu3fvply5cnl+tnnz5gBcuHCBypUrZ5vv6upqdMtgYbts6tGSokDKLUiLAxcvcC8FDkacIlLuwI19cPM4pMWCo4u6Dt8qUKou+FQAjUm7xiwSNvP/ZSckgBbCdAod1D04ZuupU6eIjIzUv9dqtWzZsoWyZcvmu55Vq1axePFiHnvsMaZMmcJzzz1H5cqVqVevHgcOHGDEiBG5LrtmzRqDyurq6mpUdyOKojB8+HA2btzIzp07qVixYr7LHD9+HIAyZcoYvB1z0mg06HQ6tFqtpYtSLFh9UKfTqoHYhR/h4o8Qe+H+PAdnKFkLAhpAQKO7rwbg4gmKDuKvqgHctV0Q9ifc/lddzr0UuJUEXTok34KMRHW6qw+UbQ3l2kG5tur6HJ3NvMP5k9+I+Vn1b0QIG1LooO7emK0ajSbHx6zu7u58+umn+a4nMjKSunXrAuDp6alvlPD4448zceLEwhazQIYOHcrq1av58ccf8fLy0gesPj4+uLu7c/HiRVavXk2PHj0oWbIkJ06cYPTo0bRt25Z69epZpMz5cXJy4syZM0yfPt3SRSk2DLmpMauMZLiyTQ3iLv0CKbfBIxAq94I2s9SALD0BEsLh1j9w8xicWQvaNEAD7iUhPR60dzPz3hWgfEdo+pYatHmHwr2LtKJA4g24cxIi/obre2DfFMhMBmdP9fMhHaB8BwhoaFxmsIg4Ojry119/8ddff1m6KMWGo6OjpYsghF0o9Bn08uXLKIpCpUqV+PvvvyldurR+nouLCwEBAQb9YMuVK0dERATly5encuXKbNu2jUaNGnHo0CGDHlf++eefDBs2jAMHDuDt7Z1lXlxcHI888giLFi2iTZs2Bu/bF198AagdDD9o+fLlvPTSS7i4uPDHH3/w8ccfk5SUREhICE899RTvvfeewdswty5dulCjRg1LF6NYMSTDm6f0RDWwunlMDaTc/MC/JpSqA67e+S8PkHAdLv8GF3+CsN8hM1VdR93BUPkJKNMs70ej2gy4cwpuHoWkCHD2Ar+qahm88qiSoNGAV1n1VaHb/XVFHYHwnRC+A/ZPhT1vg4u3msEL6QAh7aF0fXAw/8W+T58+3Lp1y+zbLa40Gg3VqlWzdDGEsAsaxUoqNIwfPx5vb2/eeecd1q1bxwsvvECFChUICwtj9OjRzJ49O8/le/XqRYcOHRg9enSO8xcsWMCOHTvYuHFjURS/wOLj4/Hx8SEuLi5bMCpsQFq8GqDEXgBFqz569KkIJeuAUwHqTt4L4KKO3H0dhuizgAKOruDkpm4TRQ3CAhqqQVC5dhDURN1+ZoqaZbv5D0T+rQZOt/5RP1+mJVR5Qg3k/K3kQqpNh8hDajnDd8KNv9Sg080PyrZVs3ghHdQA0gbq5AkhspLrnPmYPKg7deoUYWFh2RpN9OrVy6j1HDhwgH379lG1alV69uyZ7+dDQ0PZsmULNWvWzHH+mTNn6Nq1K2FhYUaVo6jJl93GpCeqwceVbRC2HaLPAHd/QhpHNbADtZFAQGMI7QTVnrkbkDxUbyglGm4dv5+Fizp6f31ObmqmKrAJBDZWX/411TpomalqoBd1+H62K/F6zuX1rQzBj0DFHhDaFdxN02djkcpMg8iDELZD3beI/XczlCXVDN69TF7JWtmPqRDWTlHU32vsBbUqhMYBPALAs6z6rx1+p+U6Zz4mC+ouXbrEk08+yb///pul24x7FWDzqnSckZHBa6+9xsSJEwv8qMrNzY2TJ09SpUqVHOdfuHCBunXrkpKSUqD1FxX5sls5RacGXFe2qoHcjX2gy1CzcaFd1MxXUFPwr6GenFOj1ZN1xEH1s1e2qi1BvcqrrT9BfR9/VW1lCuDkcTeAa3Q/gCtZy/D6ZYoC8VfUcqbFqxk9r7LgXws8Spn+mJhbRooa2IXvUAO9yIOgy1QvgOXa38/k+VWzywuisBPxYXDiKzizBuIu5fwZt5LqDWBAA7W+aXAr8LTORnfGkOuc+ZgsqOvZsyeOjo4sWbKEihUr8vfff3Pnzh3Gjh3LRx99lG9dNh8fH44fP17goK5y5crMnTtXP+zYwzZs2MC4ceO4dCmXH5OFyJfdCiXeUAO4q9vg6u9qQwJnT7UxQGhXtW6Yb2XDAghtOlz9Q82oJV5XAz8XL7VxgU8FtasPv2oWqTtmszKS4Ppfdx/X7oDIw2qGtESZu4+i20BQM/XYFmXr2jun4OIvEHMO0mLUlsRO7uojcI/S4FNJrXfoW9U2MqTC9BRF/a4e+wTOb1Rbjld7Gio9rmbeXTzVG5Tkm5AQBrf/U1uRRx2GuMvqOnwrq9UrQrtC+U42eaMm1znzMVlQV6pUKf7880/q1auHj48Pf//9N9WrV+fPP/9k7NixHDt2LM/lBw4cSIMGDXKtE5ef4cOHs3PnTg4dOpRtqLGUlBSaNWtGhw4dWLBgQYHWX1Tky24FMlLUVplXtqqB3O2TgEbNmFW4G8SVaaE+UhXWJz0Bru+9/7j25jE1yHNyg9IN1ACvTDMIbAp+VQpXLy8xAs6uhVPfqo1GnD3Bv/r9Pv0ykiH1DiRFqhfqe9xKQsmaagbWv6b6t39N8AqR7KI9ykyDc9/B0U/UurF+1aHRCKg1QA3kDJFwXa1fen2v2mXQnf9Qz0uN1AAvtItataIgdXdBzRzePKbeJDm5qd9hr/LgXd7kdVflOmc+Jgvq/Pz8OHr0KBUrVqRy5cosWbKEDh06cPHiRerWrUtycnKey0+fPp25c+fSqVMnGjduTIkSJbLMz6ufOoCoqCgaNWqEo6Mjw4YNo3r16oBal+6zzz5Dq9Vy9OhRAgMDC7ejJiZfdgtQFPUEee+R6vXdaj01z+D7mbjynW3yjligBlY3j6uNRCIPqf/e63/PxUvN4JWuB6Xq3f23bt6tiNPi4cJGOL0awv5Qg7dKj0PNF9S6irldVNMTIOYCxJ5X60reOQ3RpyHmrPp9A3AukTXIu/e3b2Wr6N5FGCkpCk58Ccc/h+Qo9VzSaJR6c1jYQCnhuvr9u3LvCcIttepGSHs1wAvpAKXr5r6d5FvqE4Ow7errwT4pH+Tkpj496LURfCsVrsx3yXXOfEwW1LVp04axY8fSu3dvnn/+eWJiYnjvvff46quvOHLkCCdPnsxz+bweu2o0GoMem169epXXX3+drVu3ZqnT161bNz777LPCdy1RBOTLbibJt9UT4dW7j1UTb6gnr3uPNSp0k4r39iwlWn2kdfM43D4Bt06oAZYuU53vXUGtF+lTSa0v6eKpPi6/sV/NlGjT1O5WavZXH5+5+RW8LDqtWqcy+rT6CPdesBd9+n49Swdn9cL6YLDnX00tX2G2LYpG5GE4tgDOrgONE9QeCA2Hq/9/RUHRqd/hewHe9T3qd9TND8q2Uc9rJcqoVUeiT6nf4dt3r8F+1dXHuOU7QnBLtSshbZoa9MVfUW9Aos9Auw/VmyATkOuc+ZgsqNu6dStJSUn06dOHCxcu8Pjjj3Pu3DlKlizJunXrDBr/1VRiYmK4cOECiqJQtWpV/Pys9yQoX/Z8KDr1MYFGAy4+4OZr2HIP1ru6+rvashRFzcrcC+LKtgZn96IsvbBm2nT14nXrH/UCGXNOrcAed1nNpJUIVLuMKd8Jqj6lPpYqSoqiPrbNKdhLuj9SD66+auDpU0l9+Va6/94rRL1ZEUVPmwHnN6jB3I196o1Bw2FQ52XzB94ZKWoDovBdcG2nejOiTVOrjPhVg6Dmd1uOt8+7X8kiItc58ynSfuqio6Px8/MzagiY9PR0Ll++TOXKlXFysv/HD3b7ZU+LU08w8VfUC6STu3rB8a2svpw9si+jKJBwDaIOqaMPRB1S74DT4+9/xtVXHUvUr5qaWfGvrl7Q0hMg7op6gY46oj5y02WoLSTLd1KDuNAu6iNWIfKiKOj7AbQWqTEQe1ENOmMvQfxl9d+4S2rWT3mgdwGPQDUAvVc/6uF/3UtLRrowUqLVVqzHF6rZ3JD20HAkVO5pPQ2etBnqMH2OblZRJru9zlkhkwR1GRkZuLu7c/z4cerUqVOgdSQnJzN8+HC+/vprAM6dO0elSpUYPnw4ZcuWZfz48YUtplWyqy97/FW4+LM6akH4TjWocnRVA7rM5PvDSoE6lJRfdTVA06aq2bhbx9XHBaD22RTUVK3kHtBAfaSRdvfCFntB7act5uz9zwOgUVsblm6gPior30F9bCUXMGHPdJnqzVDsRbUFZUK4+ntKCLv/b+YDXTk5ud3tYqei+vKu+MDfFdRh4OQ3k130ObXhw38r1CC6Zn9oNFKtlynyZFfXOStnklSYs7Mz5cuXL9QA2BMmTOCff/5h586dPProo/rpnTt3ZsqUKXYb1Nm8xAi1HsmZ1WqldAdntcJu+/lQ6bH744AqilpxOPai+pgr+qz66Ov6HjVrVyIYGgxTH3cFNTE8o5YSrQaTrt5qHZKcMoBC2DMHJ7V7nHv9ID5MUSDlzgNB3lU1qx13WX1Md3p11my4k8fdzF7IA1m+u397hagvS1RbSE9U68PGXlT7etQ4qK2PPYPV84dflbutiU2YYVUU9Qb1yHx1nGSP0uoYxw1eV58CCGFlTPb4denSpWzYsIFvvvkGf3/j+2QKDQ1l3bp1tGjRAi8vL/755x8qVarEhQsXaNSoEfHx8bkum5GRwcqVKwF48cUXcXGxna4nbPIOJj0Rzv0Ap1dB+J/qSAoVe0CNZ6Fid3D1sXQJhRCGUhT18W78ZTXQi7+b7Xsw65cclXWZe91feIVkD/q8y6s3WKZ47KdNV1upn16tPgHITFYr9rv5q/Vt0+PVAO8eJ3f1CYB/DfXlV1Wt7uFT2bgM5L0uSQ7PU58glKoDjUZDzeelzmIB2OR1zkaZrNLawoULuXDhAsHBwYSGhmbrkuTo0aN5Ln/r1i0CArLf+SQlJeVbJ2/cuHF06dIFRVF48803+eSTT4zfgWLi+vXrXLx40fgFFUU9yUcdVTvHvDeqQrkPoVRt9WR6E3W8UZFFzZo1KV26tKWLIQx0/vx5IiIiLF0MCwlQX45NwBf1Beoj3vR4SI2F9Di1zmxaHETEwZUISD+dtXoFmrvBl6/6r6uverN37+XiqwZHGg0ajYYGDRrg5XW3paVOq2bHzq6D8+vVUVpK1YEWE6FGP/W886CMlPvDbkWfudu45IzabUfKrfufc/G+X6fQvZQa5Ln6qeVwdFO7ptFmqH23Xf5V3b+K3aHtHAjtLI+khU0wWVCX20gOhmrSpAm//vorw4cPB+4PL7ZkyRJatmyZ57I6nQ6dTodWq0Wn0xWqHPZu9+7dXLx4EXd3Ax+fKFq1Pk5GCiiZoHEBpzZqEJfiCOEKhOfdXU1xlpycTHx8PI8//riliyIM9Ntvv5GYmIirawE7dS0WStx93a0m4Qg46NTzhaJTA7N0LaRp1b+VDFBuAg9m/DSgcSRR547LxfU0D0UNzG7sUzOHPhWh3hCo8bza/1punN3VR69+VaDio1nnpcXfbWBy8W6dw3D1UXTqHXVbqTFqK1Ftqpqd02jUro0ajoQaz0HJGqY8aEIUOZMFdZMnTy7U8jNnzqR79+6cOnWKzMxMPvnkE06dOsW+ffvYtWtXnsvOnTuX1atXoygKH330UaHK8aBZs2axYcMGzpw5g7u7O4888ggffPCBvmNjgNTUVMaOHcvatWtJS0ujW7dufP7551bXyfE9Op2OqlWr0q9fv9w/lBYHF35Ue86/sg3cnaHeU2pT/ZD21tUq0MotWbKEImxgLoqAoig0b96cTp06Wboo9kXRqZ3zZmnMEc6svS7oEm/AlZNqJq3RKKjwqNpQqrDZMVdvCGyovoQoBkzaZ0hsbCw//PADFy9e5M0338Tf318/ikPZsmXzXLZ169YcP36c2bNnU7duXbZt20ajRo3Yv38/devmcZcGuLi48NJLL5lwT1S7du1i6NChNG3alMzMTN555x26du3KqVOn9I+XR48eza+//sr333+Pj48Pw4YNo0+fPvz1118mL49BEm+oJ0+No5pNc/XJcmJUFCXnx9mJN9T+3C5sgsu/qXevZVtDp4VQ/VnD+4cT2UhQZ1ty/Y2IwtE4qIPTe5aBMs3vTz8wC6XRKHjkEYsVTQh7YbKg7sSJE3Tu3BkfHx+uXLnCq6++ir+/Pxs2bCAsLEzfkCEvlStXZvHixaYqUqFt2bIly/sVK1YQEBDAkSNHaNu2LXFxcSxdupTVq1frO1devnw5NWvW5MCBA7Ro0cL8hV7bRu276h4nN7Vl2L0WYtFVwdVNbc2VFq+2RL15TK2HgkYdI7P1TLXXfO8Q85ffzmg0GgnqbIz8f5mXBNBCmI7JgroxY8bw0ksvMWfOnPsVXoEePXrw/PPPG7QOrVbLxo0bOX36NAC1atXiiSeesJpOiOPi1CF87rXuPXLkCBkZGXTu3Fn/mRo1alC+fHn279+fY1CXlpZGWlqa/n1erXoL5NEV6mgKik79NylCzcIl3YDE6ygp/miS4uCvn9QhYHwqQ0g7aDlJxjsV4i4JNMxHbnyEMB2TRUuHDh3iyy+/zDa9bNmyREZG5rBEVv/99x+9evUiMjJSX2ftgw8+oHTp0vz88895dmocFhZG+fKGD+Fz/fr1fB8HP0yn0zFq1ChatWqlL0tkZCQuLi74+vpm+WxgYGCu+zxr1iymTp1q1LaNUq5N3vO/+QaNuzv0XV10ZRB6EhzYHnn8KoSwVSar8e7q6ppj1uncuXMGdecwePBgateuzbVr1zh69ChHjx4lPDycevXqMWTIkDyXbdq0Ka+99hqHDh3K9TNxcXEsXryYOnXqsH79+vx36CFDhw7l5MmTrF271uhlHzRhwgTi4uL0r/Dw8EKtz1hyR2xekoWwPfL/ZV7yGxHCdEyWqevVqxfTpk3ju+++A9QfalhYGG+//TZPPfVUvssfP36cw4cP4+d3fyBkPz8/ZsyYQdOmTfNc9tSpU8yYMYMuXbrg5uZG48aNCQ4Oxs3NjZiYGE6dOsV///1Ho0aNmDNnDj169DBq34YNG8Yvv/zC7t27KVfu/mDIQUFBpKenExsbmyVbFxUVRVBQUI7rcnV1tWhXCZKFMD+5YNke+Y2Yl/xGhDANk2Xq5s6dS2JiIgEBAaSkpNCuXTuqVKmCl5cXM2bMyHf5atWqERUVlW36zZs3qVKlSp7LlixZknnz5hEREcHChQupWrUqt2/f5vz58wD079+fI0eOsH//fqMCOkVRGDZsGBs3buTPP/+kYsWsnV42btwYZ2dntm/frp929uxZwsLC8u1bTxQPEhzYHgkwzEt+I0KYjskydT4+Pvz+++/s3buXEydOkJiYSKNGjbI0IsjLrFmzGDFiBFOmTNE3MDhw4ADTpk3jgw8+yPJoN7dhRtzd3enbty99+/Yt/A6hPnJdvXo1P/74I15eXvp6cj4+Pri7u+Pj48Mrr7zCmDFj8Pf3x9vbm+HDh9OyZUvLtHw1gGTqzE+CBNsjvxHzkcevQpiOyZuVtm7dmtatWxu93L0e95955hn9CfXeD71nz5769xqNBq1Wa6LS5u2LL74AoH379lmmL1++XN8v3vz583FwcOCpp57K0vmwtZKgzrzkgmV75DdifvIbEcI0TBrUbd++nfnz5+u7JKlZsyajRo0yKFu3Y8cOUxbFJAw50bi5ufHZZ5/x2WefmaFEwtZIcGB7JMAwL/mNCGE6JgvqPv/8c0aOHEnfvn0ZOXIkoD4+7dGjB/Pnz2fo0KF5Lt+uXTtTFUXkQbIQ5idBgu2R34h5yW9ECNMwWVA3c+ZM5s+fz7Bhw/TTRowYQatWrZg5c2a+QR2o46ieOHGCmzdvotPpsszr1auXqYpa7MkFy3zkWNseufExLznWQpiOyYK62NhYHn300WzTu3btyttvv53v8lu2bGHAgAHcvn072zxz1qOzd3JHbH5yzG2L/H+Zl9Q7FcJ0TNalSa9evdi4cWO26T/++KO+EURehg8fztNPP01ERAQ6nS7LSwI605GTp3nJBcs2SfbIvOQ3IoRpmCxTV6tWLWbMmMHOnTv1fbQdOHCAv/76i7Fjx7JgwQL9Z0eMGJFt+aioKMaMGUNgYKCpiiRyIRcs85FjbXskwDAv+Y0IYTomC+qWLl2Kn58fp06d4tSpU/rpvr6+LF26VP9eo9HkGNT17duXnTt3UrlyZVMVSeRA6guZnwQJtkd+I+Yj2WwhTMdkQd3ly5cB9HXiSpUqZdTyCxcu5Omnn2bPnj3UrVsXZ2fnLPNzCgSFsHYSHNgeufERQtgqkwR1sbGxvPvuu6xbt46YmBhAHbf12WefZfr06VnGRc3NmjVr2LZtG25ubuzcuTPLSTW37J4wnlywzE+yELZF/r/MSzJ1QphOoYO66OhoWrZsyfXr1+nfvz81a9YE4NSpU6xYsYLt27ezb98+/Pz88lzPu+++y9SpUxk/fjwODiZrvyEeIidP85ILlm2SGx/zkt+IEKZR6KBu2rRpuLi4cPHixWyNHKZNm0bXrl2ZNm0a8+fPz3M96enp9OvXTwI6M5ALlvlIUGd7JJttXnKshTCdQkdQmzZt4qOPPsqx1WpQUBBz5szJsauThw0cOJB169YVtjgiH3LBMj8J6oTIndz4CGE6hc7URUREULt27Vzn16lTh8jIyHzXo9VqmTNnDlu3bqVevXrZGkrMmzevsEUVSIBhbnLBsj1y42N+8hsRwjQKHdSVKlWKK1euUK5cuRznX758GX9//3zX8++//9KwYUMATp48mWWenGBNS46nELmTAMO85HwkhOkUOqjr1q0b7777Lr///jsuLi5Z5qWlpTFx4sQchw972I4dOwpbFGEAuWCZl2TqbJMEGuYlvxEhTKPQdeqmTZvG2bNnqVq1KnPmzOGnn37ixx9/ZPbs2VStWpXTp08zdepUg9d34cIFtm7dSkpKCmDZH/vu3bvp2bMnwcHBaDQaNm3alGX+Sy+9hEajyfIyJIC1NLlgmY8ca9sjj1/NS461EKZT6ExduXLl2L9/P2+88QYTJkzQB2EajYYuXbqwcOFCQkJC8l3PnTt3eOaZZ9ixYwcajYbz589TqVIlXnnlFfz8/Jg7d25hi2q0pKQk6tevz8svv0yfPn1y/Myjjz7K8uXL9e9dXV3NVbwCkQuW+UkWQojcSTZbCNMxSefDFStW5LfffiMmJobz588DUKVKFYPq0t0zevRonJ2dCQsL0/d1B9CvXz/GjBljkaCue/fudO/ePc/PuLq6EhQUZPA609LSSEtL07+Pj48vcPlycuLEiSzrf9i9DKgwD41GQ1xcHIcOHbJ0UYQR5MbHvKKiouQ3YmXq1q2Lm5ubpYshjGSyYcJAHUWiWbNmBVp227ZtbN26NVuDi6pVq3L16lVTFK9I7Ny5k4CAAPz8/OjYsSPTp0+nZMmSuX5+1qxZRj2ONtaePXv0o3rkxtgh3ETBlSpVijNnzrB161ZLF0UYyNnZOd/O0oXpBAQEcPr0aa5du2bpoogHVK5cWYI6G6RRrCTv7eXlxdGjR6latSpeXl78888/VKpUicOHD9OtWzfu3Llj0fJpNBo2btxI79699dPWrl2Lh4cHFStW5OLFi7zzzjt4enqyf/9+HB0dc1xPTpm6kJAQ4uLi8Pb2LurdEEIIIcwqPj4eHx8fuc6ZgUkzdYXRpk0bVq5cyfvvvw+oQZROp2POnDl06NDBwqXL2bPPPqv/u27dutSrV4/KlSuzc+dOOnXqlOMyrq6uVl/vTgghhBC2x2qCujlz5tCpUycOHz5Meno6b731Fv/99x/R0dH89ddfli6eQSpVqkSpUqW4cOFCrkHdw+4lSk1dt04IIYSwBveub1byYNCuWU1Q5+3tzenTp/niiy/w8vIiMTGRPn36MHToUDIyMixdPINcu3aNO3fuUKZMGYOXSUhIADCohbAQQghhqxISEvDx8bF0Meya1dSpc3R0JCIigoCAgCzT79y5Q0BAAFqt1uxlSkxM5MKFCwA0bNiQefPm0aFDB/z9/fH392fq1Kk89dRTBAUFcfHiRd566y0SEhL4999/DX7EqtPpuHHjBl5eXvm2uLtX/y48PNym6yXYy36A/eyL7Id1kf2wPvayL5bYD0VRSEhIIDg4GAeHQnePK/JgNZm63GLLxMREi7XAOXz4cJb6fGPGjAFg4MCBfPHFF5w4cYKvv/6a2NhYgoOD6dq1K++//75RdeYcHBxyHWItN97e3jZ9UrnHXvYD7GdfZD+si+yH9bGXfTH3fkiGzjwsHtTdC5Q0Gg2TJk3Cw8NDP0+r1XLw4EEaNGhgkbK1b98+zzoA0k2FEEIIIayFxYO6Y8eOAWqm7t9//80yfqyLiwv169dn3LhxliqeEEIIIYRNsHhQt2PHDgAGDRrEJ598Yhdp7aLi6urK5MmTbb5LFHvZD7CffZH9sC6yH9bHXvbFXvZD5MxqGkoIIYQQQoiCk2YoQgghhBB2QII6IYQQQgg7IEGdEEIIIYQdkKBOCCGEEMIOSFAnhBBCCGEHJKgTQgghhLADEtQJIYQQQtgBCeqEEEIIIeyABHVCCCGEEHZAgjohhBBCCDsgQZ0QQgghhB2QoE4IIYQQwg5IUCeEEEIIYQckqBNCCCGEsANOli5AcafT6bhx4wZeXl5oNBpLF0cIIYQwKUVRSEhIIDg4GAcHySUVJQnqCmn37t18+OGHHDlyhIiICDZu3Ejv3r0NXv7GjRuEhIQUXQGFEEIIKxAeHk65cuUsXQy7JkFdISUlJVG/fn1efvll+vTpY/TyXl5egPpl9/b2NnXxhBBCCIuKj48nJCREf70TRUeCukLq3r073bt3N/jzaWlppKWl6d8nJCQA4O3tbZ6gLuUOnF4NirZot+NfAyo+mvv8S5sh5lze6wjtAqVq5zxPlwknV0BGYoGLaBIaDVTtC15lc56fFgenvgVdRuG2U+kx8Kua87zMNPhvBWSmFG4bIe0hoEHO8xQd/Pe1uj+50ThAjWfBI6Bw5bC02Etw8ae8P+MVAtWeMk957F18OFzYAIqS+2c8AqHmc0VbjrPfQeKNot2GkxvUGgjO7kW3jfBdcPNY3p8JfgTKNCu6MuRCqhgVvQIFdVqtlo0bN3L69GkAatasSe/evXFykhgxP7NmzWLq1KmWK8DZ72DHCHD2LLpt6NLBxRveuJX7ZzY/rwYhDi45z89MhmpPw+Nrc55/8zj8/io4uYPGsdBFLrCMREiNhUcm5zz/4s/w5zBwLgEU8ISWkQR3TkHXr3Kef2Mf/PE/cPJQA6uCyEyBCt2gz685z485D1tfzvt4ZySqwXaTMQUrg7U4Mh+Of3b3/ywHugzQpsHIFPUiLQrnxCI4ODP3c5KSCZmpUKEruJcsmjKkJ8Av/cDRFRyci2Ybik49r3lXyPuGt7D++B/EXlT3JSeZKWpQ9+zuoiuDsBijo7D//vuPXr16ERkZSfXq1QH44IMPKF26ND///DN16tQxeSHtyYQJExgz5v5F715a2mx0meqFeURC0W3j0Ifw96z8y9F2DjQamfP89Y/mnU3UZar/9v8bSlnwO/dVqGHlHBYHDgUMPlc/Ytg2Bp0G7/IF28bPT6sXtvy28fR2CG6Z82cW+hV9BtgclEwIaAgvHsl5/ulVsPmFvDNLwnC6TDUL/XIumfuLP8OmXkX73dLdXXePVUWXgU2KhEVliv43osuExqOh7Qc5z98yKP+nJMJmGR3UDR48mNq1a3P48GH8/PwAiImJ4aWXXmLIkCHs27fPoPVs376d7du3c/PmTXQ6XZZ5y5YtM7ZYNsPV1RVX11zuoMzCTBei/C54hlwQ8/yMNV1QzVBOcxwLU2zDHgIdg/fBDvbVGhh6vIv0u2XG/8si/40U9twqbJnRQd3x48ezBHQAfn5+zJgxg6ZNmxq0jqlTpzJt2jSaNGlCmTJl5Dm73THn/6eFvztW9d01R1ny2IZVHYtCynNf7Gg/rUZxOKbWcl4sDse6+DI6qKtWrRpRUVHUrp21AvvNmzepUqWKQetYtGgRK1as4MUXXzR286LQFMzzo87vTjC/cmjyXoc13WnmVZZ78woT8GjyORYmyTKY6nhb0f9LgRWnfbUG+R3He7+dIjzepvidGr6xIl69IeuX7669MjqomzVrFiNGjGDKlCm0aNECgAMHDjBt2jQ++OAD4uPj9Z/NrTVneno6jzzySAGLbF0SExO5cOGC/v3ly5c5fvw4/v7+lC9fwPpNts6cGRuLZ4csvf0HmONYFJsMVnHZT2tRDI6ptZwXLX7OFEXJ6KDu8ccfB+CZZ57RPzZV7t4Z9OzZU/9eo9Gg1eZcIXTw4MGsXr2aiRMnFqjQ1uTw4cN06NBB//5eI4iBAweyYsUKC5UqD4pinh91vneL+WTqNBo7q1NXmGNujqyliY63NWVQC8oq6ngVI/kdx3vnK7PUqbOGc2OhN2CizwhbZHRQt2PHjkJvNDU1la+++oo//viDevXq4eyctQn5vHnzCr0Nc2nfvr0+qBX3WEvdEXNs3pruei1cp87S/xemJJkO8yoWx9RazovF4VgXX0YHde3atSv0Rk+cOEGDBg0AOHnyZJZ50miiqFlJnbp8M4ZSp07PHFnL/OrtSZ26QnxO5M2K6tRZw7mx0KuX1q/FWYF6C05NTeXEiRM5dkfSq1evfJc3RbZPWDGzBuaWvgmw9PYfYOk6dXZ1QyaZDvMqDsdU6tSJomd0ULdlyxYGDBjA7du3s83Lqx5dbq5duwYgg/yai9SpKwK20jK1ENsoTtkrqVNnZlZUp84qzo2F3oCJPiNskdFjCg0fPpynn36aiIgIdDpdlpehAZ1Op2PatGn4+PgQGhpKaGgovr6+vP/++9kyf8IWWckdaXHYfhZSp85kJNNhXsXhmFrNE4xicKyLMaMzdVFRUYwZM4bAwMACb/Tdd99l6dKlzJ49m1atWgGwd+9epkyZQmpqKjNmzCjwukV+7KROnTXdaeabUSzk8TZXnTpD6gbmxy6yV8UoK2kN8v3OmKFOnTlbv0qdOlGEjA7q+vbty86dO6lcuXKBN/r111+zZMmSLPXv6tWrR9myZXnjjTckqLN1VnNHWhy2/wCpU2dCkukwr+JwTK3kCYZd/U7Fw4wO6hYuXMjTTz/Nnj17qFu3brbuSEaMGJHvOqKjo6lRo0a26TVq1CA6OtrYIglj2Eqduvz6TbOqO818ylno4y116syqWGUlrYEV1KmzpxElpE5dsWZ0ULdmzRq2bduGm5sbO3fuzNIFiUajMSioq1+/PgsXLmTBggVZpi9cuJD69esbWyRhdazkjrQ4bD8LqVNnOpLpMKvicEyt5glGMTjWxZjRQd27777L1KlTGT9+PA4ORrezAGDOnDk89thj/PHHH7Rs2RKA/fv3Ex4ezubNmwu0TmEoG6lTp9GAklejGSu60ywOdeqK04gSxSkraQ2KW526ov6NSJ26Ys3oqCw9PZ1+/foVOKADtQPjc+fO8eSTTxIbG0tsbCx9+vTh7NmztGnTpsDrFdbCWu5Ii8P2H2CWTEAxyWAVmzFurUVxOKZWcl60p9+pyMboTN3AgQNZt24d77zzTqE2HBwcLA0iLMFcderyvas2oE6drYwoURzq1MmIEjl8zB721RpYQaZO6tQJO2F0UKfVapkzZw5bt241atzWEydOUKdOHRwcHDhx4kSe26hXr56xxRLWxJx3gha/67T09h9k4davVnUsCksyHWZVHI6p1ZwXi8GxLsaMDur+/fdfGjZsCBg3bmuDBg2IjIwkICCABg0aoNFoUHK4081tVIqGDRsaPC7s0aNHDfpc8WSmOnX5ZTEMqVOnk37qAKlTZ27FKitpBfI73uYcUcIazo2F34AVlEFYitFBXUHHbb18+TKlS5fW/22s3r17F2i7whKspO6IWTZvTXe9UqfOZCTTYWbF4Zhay3mxOBzr4svooK6gQkND9X9fvXqVRx55BCenrJvPzMxk3759WT57z+TJk4u8jMWCYiWtX6VOnRGkTp15FaOspFWQOnWmXb3UqSvOjA7qOnTokOdj0D///NOgdURERBAQEJBlelxcHB06dDB4DFlhpaym7ohZCmDh7T/A0iNKWNOxKLTisp/WohgcU2s5L1r8nCmKktFBXYMGDbK8z8jI4Pjx45w8eZKBAwcatA5FUXIMDO/cuUOJEiXyXV6r1TJ//ny+++47wsLCSE9PzzJfRqXIi7WMKAH5Psor6npkJlPEderyG11D6tSZVrHKSloBqVNn6g2Y6DPCFhkd1M2fPz/H6VOmTCExMTHPZfv06QOojSFeeuklXF1d9fO0Wi0nTpzgkUceybcMU6dOZcmSJYwdO5b33nuPd999lytXrrBp0yYmTZpkxN6IomEtdUfMsXlruuuVESVMRjId5lUsjqm1nBeLw7Euvgreg/BDXnjhBZYtW5bnZ3x8fPDx8UFRFLy8vPTvfXx8CAoKYsiQIXz77bf5bmvVqlUsXryYsWPH4uTkxHPPPceSJUuYNGkSBw4cMNUu2SkrqFNnUP0VG6pTl98YtYW9YGlMNS5rnhvJez3FKntVnPbVGkidOtOuXlq/Fmcmayixf/9+3Nzc8vzM8uXLAahQoQLjxo0z6FFrTiIjI6lbty4Anp6exMXFAfD4448zceJEg9YRFhbG1atXSU5OpnTp0tSuXTtL5lAUgrXUHTFPASy8/QdYuk6dxf8vTEkyHeZVDI6ptZwX7ep3Kh5mdFB37xHqPYqiEBERweHDhw0OqArbkrVcuXJERERQvnx5KleuzLZt22jUqBGHDh3KMzC7cuUKX3zxBWvXruXatWtZ+slzcXGhTZs2DBkyhKeeeqpQw6BZNXONKGFQ/Sx7OfGYoU6dWVq/CmEJBtapM0sZLH1uFKJwjI5cHnxk6uPjg7+/P+3bt2fz5s1GBWs//PADzzzzDC1atKBRo0ZZXvl58skn2b59OwDDhw9n4sSJVK1alQEDBvDyyy/nuMyIESOoX78+ly9fZvr06Zw6dYq4uDjS09OJjIxk8+bNtG7dmkmTJlGvXj0OHTpk8L6Ih5noxFjYwNAcrCr4LGRZCn2883tMbCMM6Rj73ueEaRjyO7L5422uc0Uhu4sSNs3oTN29R6iFsWDBAt59911eeuklfvzxRwYNGsTFixc5dOgQQ4cOzXf52bNn6//u168f5cuXZ//+/VStWpWePXvmuEyJEiW4dOkSJUuWzDYvICCAjh070rFjRyZPnsyWLVsIDw+nadOmBd9Jq2VLdepshL2MKCGEJeQbrNlZpk4CKlGEClyn7vDhw5w+fRqAWrVq0bhxY4OX/fzzz/nqq6947rnnWLFiBW+99RaVKlVi0qRJBeqOpGXLlrRs2TLPz8yaNcvg9T366KNGl0E8wGQBgg1k6iy+/QcU+rgXMhDPN/i0FYZkOu59TpiGId9dWz/eZjpXGJJptovfqciJ0UHdtWvXeO655/jrr7/w9fUFIDY2lkceeYS1a9dSrly5fNcRFham77rE3d2dhIQEAF588UVatGjBwoUL813H+fPn2bFjBzdv3kSn02WZJ92a5MFW6tRZU7CUL3sZUUIIS7CCOnX21PpVFGtGB3WDBw8mIyOD06dPU716dQDOnj3LoEGDGDx4MFu2bMl3HUFBQURHRxMaGkr58uU5cOCAvr6bYsAFavHixbz++uuUKlWKoKCgLB0ZazQaCeoszgx16sx6Es6DpbefRRFm6qRO3X1Sp870ikOdOrOdK6ROXXFmdFC3a9cu9u3bpw/oAKpXr86nn35KmzZtDFpHx44d+emnn2jYsCGDBg1i9OjR/PDDDxw+fDhb69qcTJ8+nRkzZvD2228bW3xhK3XqrCpYyofUqROi4IpbnTpbD06FVTM6qAsJCSEjIyPbdK1WS3BwsEHr+Oqrr/SPTIcOHUrJkiXZt28fvXr14rXXXst3+ZiYGJ5++mnjCi7MSOrUWURhAzNDMqOFGtrNVkidOvOTOnUmo+Tz/bWb36nIidFdmnz44YcMHz6cw4cP66cdPnyYkSNH8tFHHxm2UQcHnJzux5PPPvssCxYsYPjw4bi4uOS7/NNPP822bduMLXqOrl27lq1Onl0zV526vAtx91+pU2cYqVMn7JkVZOqkTp2wE0Zn6l566SWSk5Np3ry5PjDLzMzEycmJl19+OUs/cQ+2ZD1x4oTB26hXr16e86tUqcLEiRM5cOAAdevWxdnZOcv8ESNGGLytWrVqcfz4cSpVqmTwMiIf5mj9ai116qwqGDJDnbrCDO1mKwzJdOg/J0xC6tSZUH43knbyOxU5Mjqo+/jjjwu0oQYNGqDRaPJtCKHRaNBqtXl+5quvvsLT05Ndu3axa9eubMsbE9QZ0jDDvpirTh25Z6mkTp1xpE6dsGf5nYNlRAkhDGZ0UDdw4MACbejy5csFWq6o1yWKgowoYRFFWafO0KHd7OKCZUim4+7nhIlInTrTMaBOqF38TkVOCtz5sLFCQ0PNtSmjvPPOO/j7+1u6GOZj1h9zbicXezuhFHV9N3NtQwhLsILvr9nPi0IUDbMFdQ9auXJlnvMHDBiQ5/wxY8bkOF2j0eDm5kaVKlV44oknDArWJkyYkO9n7E8R3zEanDEqRL0PW6pTZ5KGEqb8XG7LmmJoN3u4YEmdOvOzksxoUZ5PzHWuMqifRfnu2iuLBHUjR47M8j4jI4Pk5GRcXFzw8PDIN6g7duwYR48eRavV6vvLO3fuHI6OjtSoUYPPP/+csWPHsnfvXmrVqlVk+2GbzPhjVpRcEnX2dkIp4vpuUPR16oSwFKs4H5j5vChEETG6SxNTiImJyfJKTEzk7NmztG7dmjVr1uS7/BNPPEHnzp25ceMGR44c4ciRI1y7do0uXbrw3HPPcf36ddq2bcvo0aPNsDc2qMjvGA1cf6HGJ7SlOnU2kKkzxfG2lzp1+XZDI3XqTM5qRvAoyvOJFdWpk++u3TI6qHv55Zf1Y7U+KCkpKUt3JsaqWrUqs2fPzpbFy8mHH37I+++/j7e3t36aj48PU6ZMYc6cOXh4eDBp0iSOHDlS4PLYL2uoO2JnJxSDOu0t9EYKOE8Ia2cF31+pUyfshNFB3ddff01KSkq26SkpKfnWlcuPk5MTN27cyPdzcXFx3Lx5M9v0W7duER8fD4Cvry/p6emFKo/9kjp1pmOGOnWGLl+o7UiduvtkRAnzs5LjXVzq1NlDRl3kyOA6dfHx8SiKgqIoJCQk4Obmpp+n1WrZvHkzAQEBBq3rp59+yvJeURQiIiJYuHAhrVq1ynf5J554gpdffpm5c+fStGlTAA4dOsS4cePo3bs3AH///TfVqlXTLxMWFkb58uUNKh/A9evXKVu2rMGftxnm/DHnui17O6GYI4smrV+FvbKG7681nBeFKDyDgzpfX180Gg0ajSZLsHSPRqNh6tSpBq3rXuD14LKlS5emY8eOzJ07N9/lv/zyS0aPHs2zzz5LZmYmoGb5Bg4cyPz58wGoUaMGS5Ys0S/TtGlTevfuzeDBg/WB4MPi4uL47rvv+OSTTxgyZIhRnRjbFGupU1eo+llSp8602zHF8baTDIBBrQexj321FlZzvKVOnbBtBgd1O3bsQFEUOnbsyPr167N0F+Li4kJoaCjBwcEGrauwY616enqyePFi5s+fz6VLlwCoVKkSnp6e+s80aNAgyzKnTp1ixowZdOnSBTc3Nxo3bkxwcDBubm7ExMRw6tQp/vvvPxo1asScOXPo0aNHocpovayg7oi9XQzNUadOWr8Ke2UN5wOpUyfshMFBXbt27QB1NIeQkBAcHCzScDYLT0/PfMeJvadkyZLMmzePGTNm8Ouvv7J3716uXr1KSkoKpUqVon///nTr1o06deoUcamtgZXUqStM/SzFSjJ1Bm3fBurU5dd3lcFDu9nDBUvq1JmftRxvqVMnbJvR/dTdGxkiOTmZsLCwbI0RcguycuswOCfz5s3LNq1Pnz6sWLECb29v+vTpk+fyGzZsyHWeu7s7ffv2pW/fvgaXx65InboiIHXqhCg4a/j+WsN5UYjCMzqou3XrFoMGDeK3337Lcb5Wq81x+rFjx7K8P3r0KJmZmdk6D27cuHGOy/v4+KC5e/fh4+NjbLHFg+ypTp2lW78asn17GFHC0Dp1VnGBLiSpU2d+1nK8LX0+MQmpU1ecGR3UjRo1itjYWA4ePEj79u3ZuHEjUVFRTJ8+Pc9GDjt27ND/PW/ePLy8vPj666/x8/MD1A6JBw0aRJs2bXJcfvny5fq/P//8c3Q6HSVKlADgypUrbNq0iZo1a9KtWzdjd6mYsYK6I/Z2MTRHfTepUyfslVWcD6zgvCiECRgd1P3555/8+OOPNGnSBAcHB0JDQ+nSpQve3t7MmjWLxx57LN91zJ07l23btukDOgA/Pz+mT59O165dGTt2bJ7LP/HEE/Tp04f//e9/xMbG0qJFC5ydnbl9+zbz5s3j9ddfN3a3ihk7qFNnLa1f7alOnSENPopFXR2pU2d+1nK8reUpRiFInbpizejWDklJSfr+6Pz8/Lh16xYAdevW5ejRowatIz4+Xr/cg27dupXjaBUPO3r0qD6j98MPPxAYGMjVq1dZuXIlCxYsMHRXiiepU1cELDyihJyghU2zgu+vtH4VdsLooK569eqcPXsWgPr16/Pll19y/fp1Fi1aRJkyZQxax5NPPsmgQYPYsGED165d49q1a6xfv55XXnkl30YQoDbS8PLyAmDbtm306dMHBwcHWrRowdWrV43dpeLHFurU5dfnWXEaUULq1JmXkk+mTurUmV5xqVNnlvOV1Kkrzox+/Dpy5EgiIiIAmDx5Mo8++iirVq3CxcWFFStWGLSORYsWMW7cOJ5//nkyMjLUgjg58corr/Dhhx/mu3yVKlXYtGkTTz75JFu3bmX06NEA3Lx5M8t4sCInVnBHam8XQ6lTJ0TBWcX5wBqeYAhReEYHdS+88IL+78aNG3P16lXOnDlD+fLlKVWqlEHr8PDw4PPPP+fDDz/k4sWLAFSuXFnf8CE/kyZN4vnnn2f06NF06tSJli1bAmrWrmHDhkbuUXFkJZm6QvV5ZiV16swxooQ11Kkz5HjbTV0dBTR5PcSQOnWmJ3XqTMeATLNd/E5FTowO6u5JT0/n8uXLVK5cmUaNGhVoHSVKlNCPTGFoQAfQt29fWrduTUREBPXr19dP79SpE08++WSBylJsWEXdEXs7oUidOiEKzgq+v1ZxXhSi8IyuU5ecnMwrr7yCh4cHtWvXJiwsDIDhw4cze/Zsg9ah0+mYNm0aPj4+hIaGEhoaiq+vL++//77BQ4gFBQXRsGHDLCNbNGvWjBo1ahi7S8WP1dQbMcGIElKnrgCfy23Zwh5ve6mrI3XqzC7f7xWY5btlNefGQsiv9avd/E5FTowO6iZMmMA///zDzp07cXNz00/v3Lkz69atM2gd7777LgsXLmT27NkcO3aMY8eOMXPmTD799FMmTpxobJGEUayg7ojdXQylTp0QBWYV5wMrOC8KYQJGP37dtGkT69ato0WLFvoRHgBq166trx+Xn6+//polS5bQq1cv/bR69epRtmxZ3njjDWbMmGFssYRRrKTeSKH6UipGdepMcTwNWVbq1KkUBRysJHNUbFhLZtRKzo2FYkDrV3v4nYocGZ2pu3Xrlr6fugclJSVlCfLyEh0dneNj0ho1ahAdHW1skYQxrKLuiJ2dUAzpeqXwGzHDNoSwBCv4/lrFeVGIwjM6qGvSpAm//vqr/v29QG7JkiX6Vqj5qV+/PgsXLsw2feHChVkaPoiiYi31RgzJhphqW0XFDHXqzLKPphoFxF4Ul/20FlZyvO2hTl1+27H4OVMUJaMfv86cOZPu3btz6tQpMjMz+eSTTzh16hT79u1j165dBq1jzpw5PPbYY/zxxx/6QHD//v2EhYXx22+/GVskYRQrqDti8F2xrWSnzFHfzcLbMMX/mc0wcB+s6jtow6ziu2UF50Wzrl++u/bK6Exd69atOX78OJmZmdStW5dt27YREBDA/v37ady4sUHraNeuHWfPnqVPnz7ExsYSGxtLnz59OHfunH74L1GEbGFECZNk+8zAaurUWUMLWzvKAEimw7ys5nhby7mxKLcj3197VqB+6ipXrszixYsLteGSJUvSq1cvWrRooe/G5PDhwwBZGlAIU7OGuiOmyIRY0Z2mOerU2cqoFfaQvbKKzFFxYgXfLbuqUyeZuuLM4KAuPj7eoM8ZMkzXli1bGDBgAHfu3EF56Mek0WjQarWGFksUiJXUGzFF/SyLZ02sZEQJc9Xbs5qMSlGTTId5Wcnxljp1wsYZHNT5+vrm2bpVURSDA7Lhw4fz9NNPM2nSJAIDAw0tgjAFc96R5rotqVNn0m1YVQtba/p/KajitK/WwBqOtzWcF824fqs6fwpTMjio27Fjh/5vRVHo0aMHS5YsoWzZskZvNCoqijFjxthNQPfZZ5/x4YcfEhkZSf369fn0009p1qyZpYuVO6lTZzrmyKJJnTrzk0yHeVnN8baWc2NRbke+v/bM4KCuXbt2Wd47OjrSokULKlWqZPRG+/bty86dO6lcubLRy1qbdevWMWbMGBYtWkTz5s35+OOP6datG2fPns2xPz/Ls4K6I4beJdpDnTprqu+W/0ZMsA0r+n8pKJN8P4XBrOJ4W8F50azrl++uvSpQQ4nCWrhwIU8//TR79uyhbt26ODs7Z5k/YsQISxSrQObNm8err77KoEGDAFi0aBG//vory5YtY/z48RYuXW6spN6ISfo8s/Rdp9SpM1kZrIpkOszLWo63nWTqis3vVDzMIkHdmjVr2LZtG25ubuzcuTNLXT2NRmMzQV16ejpHjhxhwoQJ+mkODg507tyZ/fv357hMWloaaWlp+veGNkAx2C/PQuL13OfHXgQ3f9NuMzcbuoODc/bp2nTDlo85C2tz6eIm+WbBy2VqV7flXs64K6Axuueg7G4ey30bSRGFXz9AfFju20i5Y9g6zm+E2ydNUx5Luf0vhHbN/3O/vQjOJYq+PPbuzmkIapr/5/54DVzyb4hXIOkJRbPenBz9GM59X3TrV3T5fybldu6/9XseWwtexlevEpZVqKDO0GHBHvbuu+8ydepUxo8fj4ODCS54FnL79m20Wm22uoGBgYGcOXMmx2VmzZrF1KlTi65QXuXByS33+b6VoVy73OebQnArqDcEtGm5fyagIZRpkfv8mi9AZnLu830rQ/V+ee+rOTQYCpd+zn2+b2UIal64bdR5GRzy+Kn6VoY6rxRuG9WfuRso5/JYxrcyVO4J7qVzX0ejURD2R+HKYQ18K0ON53KfX7qu+v+ekWi+Mtkz38pQ5cnc5/tVhYYjID2uaMtRtg2UrFm023hkCtz5r2i34VcdQrvkPr9Kb4i9BOQT/DnmcEMurJ5GebhPkVz06dMny/uff/6Zjh07UqJE1jvVDRs25Lsuf39/Dh06ZPN16m7cuEHZsmXZt29fliHS3nrrLXbt2sXBgwezLZNTpi4kJIS4uDiDuoMRQgghbEl8fDw+Pj5ynTMDgzN1Pj4+Wd6/8MILBd7owIEDWbduHe+8806B12ENSpUqhaOjI1FRUVmmR0VFERQUlOMyrq6uuLq6mqN4QgghhChGDA7qli9fbrKNarVa5syZw9atW6lXr162hhLz5s0z2baKkouLC40bN2b79u307t0bAJ1Ox/bt2xk2bJhB67iXKDV53TohhBDCCty7vhn4YFAUgkUaSvz77780bNgQgJMns1aqLmg9PUsZM2YMAwcOpEmTJjRr1oyPP/6YpKQkfWvY/CQkqBV0Q0JCirKYQgghhEUlJCRke+onTMvgOnUidwsXLtR3PtygQQMWLFhA8+aGVY7X6XTcuHEDLy+vfAPae/XvwsPDbbpegr3sB9jPvsh+WBfZD+tjL/tiif1QFIWEhASCg4NtunGkLbBIps7eDBs2zODHrQ9zcHCgXLlyRi3j7e1t0yeVe+xlP8B+9kX2w7rIflgfe9kXc++HZOjMQ0JmIYQQQgg7IEGdEEIIIYQdkKDOhri6ujJ58mSb7xLFXvYD7GdfZD+si+yH9bGXfbGX/RA5k4YSQgghhBB2QDJ1QgghhBB2QII6IYQQQgg7IEGdEEIIIYQdkKBOCCGEEMIOSFAnhBBCCGEHJKgTQgghhLADEtQJIYQQQtgBCeqEEEIIIeyABHVCCCGEEHZAgjohhBBCCDsgQZ0QQgghhB2QoE4IIYQQwg5IUCeEEEIIYQecLF2A4k6n03Hjxg28vLzQaDSWLo4QQghhUoqikJCQQHBwMA4OkksqShLUFdLu3bv58MMPOXLkCBEREWzcuJHevXsbvPyNGzcICQkpugIKIYQQViA8PJxy5cpZuhh2TYK6QkpKSqJ+/fq8/PLL9OnTx+jlvby8APXL7u3tberiCSGEEBYVHx9PSEiI/nonio4EdYXUvXt3unfvXuDl7z1y9fb2lqBOCCGE3ZIqRkVPgjozS0tLIy0tTf8+Pj7egqUpmC+/hGXLLF0KIYQtcHGBr76CmjULtnxCAjz1FMTFmbZcIm/r14M8KbU9EtSZ2axZs5g6daqli1EomzdDRAQ8+qilSyKEsGaZmbB8ORw/XvCgLiwMfv8devWCwECTFk/kwcXF0iUQBSFBnZlNmDCBMWPG6N/fq2tgSxQFGjRQ776FECI3yclqUKcoBV/HvWUnTIAWLUxTLiHslQR1Zubq6oqrq6uliyGEEEIIOyMdxogCkfquQoj83DtPFCZT9/C6hBC5k0xdISUmJnLhwgX9+8uXL3P8+HH8/f0pX768BUtWdBRFTrBCCPMwRUAoRHEhQV0hHT58mA4dOujf36svN3DgQFasWGGhUhU9CeqEEPmRTJ0Q5iVBXSG1b98epZjdShaz3RVCWJCcb4QwnNSpEwUid81CiPxIpk4I85KgThhN7pyFEOYi5xshDCdBnSgQuWsWQuRHMnVCmJcEdcJocucshDAXOd8IYTi7Ceri4+PZtGkTp0+ftnRRigW5axZC5EcydUKYl80Gdc888wwLFy4EICUlhSZNmvDMM89Qr1491q9fb+HS2Te5cxZCmIucb4QwnM0Gdbt376ZNmzYAbNy4EUVRiI2NZcGCBUyfPt3CpbN/ctcshMiPZOqEMC+bDeri4uLw9/cHYMuWLTz11FN4eHjw2GOPcf78eQuXzr7JnbMQwlzkfCOE4Ww2qAsJCWH//v0kJSWxZcsWunbtCkBMTAxubm4WLp39k7tmIUR+JFMnhHnZ7IgSo0aNon///nh6ehIaGkr79u0B9bFs3bp1LVs4Oyd3zkIIc5HzjRCGs9mg7o033qBZs2aEh4fTpUsXHBzUpGOlSpWkTp0QQlgRCcyEMA+bDeoAmjRpQpMmTbJMe+yxxyxUmuJDUeRRiBAif6Y4T9wLCOWcI0T+bCqoGzNmjMGfnTdvXhGWRAghhKEkUyeEedhUUHfs2LEs748ePUpmZibVq1cH4Ny5czg6OtK4cWNLFK/YkEydEMIQkqkTwrxsKqjbsWOH/u958+bh5eXF119/jZ+fH6C2fB00aJC+/zohhBCWJ5k6IczDZrs0mTt3LrNmzdIHdAB+fn5Mnz6duXPnWrBk9k8ydUIIQ0imTgjzstmgLj4+nlu3bmWbfuvWLRISEixQIiGEEDmRTJ0Q5mGzQd2TTz7JoEGD2LBhA9euXePatWusX7+eV155hT59+li6eHZNMnVCCHORTJ0QhrOpOnUPWrRoEePGjeP5558nIyMDACcnJ1555RU+/PBDC5dOCCHEPZKpE8I8bDKo02q1HD58mBkzZvDhhx9y8eJFACpXrkyJEiUsXDr7J5k6IYShCnuukEydEIazyaDO0dGRrl27cvr0aSpWrEi9evUsXSQhhBC5kEydEOZhs3Xq6tSpw6VLlyxdjGJJMnVCCENJpk4I87HZoG769OmMGzeOX375hYiICOLj47O8hBBCWAfJ1AlhHjb5+BWgR48eAPTq1QvNA7dwiqKg0WjQarWWKprdk0ydEMJQkqkTwnxsNqh7cHQJIYQQ1ksydUKYh80Gde3atbN0EYotydQJIQwlmTohzMdmg7p7kpOTCQsLIz09Pcv0/FrEXr58mT179nD16lWSk5MpXbo0DRs2pGXLlri5uRVlkYUQoliRTJ0Q5mGzQd2tW7cYNGgQv/32W47zc6tTt2rVKj755BMOHz5MYGAgwcHBuLu7Ex0dzcWLF3Fzc6N///68/fbbhIaGFuUu2CzJ1AkhDCWZOiHMx2Zbv44aNYrY2FgOHjyIu7s7W7Zs4euvv6Zq1ar89NNPOS7TsGFDFixYwEsvvcTVq1eJiIjgyJEj7N27l1OnThEfH8+PP/6ITqejSZMmfP/992beKyGEsD+SqRPCPGw2U/fnn3/y448/0qRJExwcHAgNDaVLly54e3sza9YsHnvssWzLzJ49m27duuW6TldXV9q3b0/79u2ZMWMGV65cKcI9sF2SqRNCGEoydUKYj81m6pKSkggICADAz8+PW7duAVC3bl2OHj2a4zJ5BXQPK1myJI0bNy58QYUQopiTTJ0Q5mGzQV316tU5e/YsAPXr1+fLL7/k+vXrLFq0iDJlyhi1rscee4yIiIiiKKZdkkydEMJQkqkTwnxs9vHryJEj9YHY5MmTefTRR1m1ahUuLi6sWLHCqHXt3r2blJSUIiilEEIIydQJYR42G9S98MIL+r8bN27M1atXOXPmDOXLl6dUqVIWLFnxIHfNQghDmOpcIeccIfJns49fL126lOW9h4cHjRo1KlBAFxoairOzs6mKZvfkrlsIYYzCnDPkfCOE4Ww2qKtSpQrly5fnxRdfZOnSpVy4cCHfZcLCwnKcfvLkSUJCQrJNv379eqHLaa/krlkIYQjJ1AlhPjYb1IWHhzNr1izc3d2ZM2cO1apVo1y5cvTv358lS5bkuEzTpk157bXXOHToUK7rjYuLY/HixdSpU4f169cXVfFtmtw5CyGMIZk6IczDZuvUlS1blv79+9O/f38Azp8/z4wZM1i1ahVr165l8ODB2ZY5deoUM2bMoEuXLri5udG4cWOCg4Nxc3MjJiaGU6dO8d9//9GoUSPmzJlDjx49zL1bNkPumoUQhpBMnRDmY7NBXXJyMnv37mXnzp3s3LmTY8eOUaNGDYYNG0b79u1zXKZkyZLMmzePGTNm8Ouvv7J3716uXr1KSkoKpUqVon///nTr1o06deqYd2dsjNw5CyGMIZk6IczDZoM6X19f/Pz86N+/P+PHj6dNmzb4+fkZtKy7uzt9+/alb9++RVxK+yV3zUIIQ0imTgjzsdmgrkePHuzdu5e1a9cSGRlJZGQk7du3p1q1apYumt2TO2chhDEkUyeEedhsQ4lNmzZx+/ZttmzZQsuWLdm2bRtt2rTR17UTRUvumoUQhpBMnRDmY7OZunvq1q1LZmYm6enppKamsnXrVtatW8eqVassXTS7JXfOQghjSKZOCPOw2UzdvHnz6NWrFyVLlqR58+asWbOGatWqsX79em7dumXp4tk9uWsWQhhCMnVCmI/NZurWrFlDu3btGDJkCG3atMHHx8fSRSo25M5ZCGEMydQJYR42G9Tl1YGwEEIIIURxY7OPXwH27NnDCy+8QMuWLfVDen3zzTfs3bvXwiWzb4oij0KEEIbRaEyTqZNzjhD5s9mgbv369XTr1g13d3eOHTtGWloaoA7zNXPmTAuXTgghhBDCvGw2qJs+fTqLFi1i8eLFODs766e3atWKo0ePWrBk9k8ydUIIQ0mmTgjzsdmg7uzZs7Rt2zbbdB8fH2JjY81fICGEEEIIC7LZoC4oKIgLFy5km753714qVapkgRIVH5KpE0IYSjJ1QpiPzQZ1r776KiNHjuTgwYNoNBpu3LjBqlWrGDt2LK+//rqliyeEEEIIYVY226XJ+PHj0el0dOrUieTkZNq2bYurqytvvvkmgwcPtnTx7Jpk6oQQhpJMnRDmY7OZOo1Gw7vvvkt0dDQnT57kwIED3Lp1Cx8fHypWrGjp4gkhhBBCmJXNBXVpaWlMmDCBJk2a0KpVKzZv3kytWrX477//qF69Op988gmjR4+2dDHtmmTqhBCGkkydEOZjc49fJ02axJdffknnzp3Zt28fTz/9NIMGDeLAgQPMnTuXp59+GkdHR0sXUwghhBDCrGwuqPv+++9ZuXIlvXr14uTJk9SrV4/MzEz++ecfNHIrZxaSqRNCGEoydUKYj809fr127RqNGzcGoE6dOri6ujJ69GgJ6IQQQghRrNlcUKfVanFxcdG/d3JywtPT04IlKn4kUyeEMJRk6oQwH5t7/KooCi+99BKurq4ApKam8r///Y8SJUpk+dyGDRssUTwhhBBCCIuwuaBu4MCBWd6/8MILFipJ8SWZOiGEoSRTJ4T52FxQt3z5cksXQQghhBDC6thcnTpr9Nlnn1GhQgXc3Nxo3rw5f//9t6WLVKQkUyeEMJRk6oQwHwnqCmndunWMGTOGyZMnc/ToUerXr0+3bt24efOmpYsmhBBCiGLE5h6/Wpt58+bx6quvMmjQIAAWLVrEr7/+yrJlyxg/frzZyxMRAVpt0W4jI0PumoUQhtFoIC4Orl0r2PK3b99fjxAibxLUFUJ6ejpHjhxhwoQJ+mkODg507tyZ/fv357hMWloaaWlp+vfx8fEmLVObNnDxoklXmaMnnyz6bQghbJ+7O3z0kfoqKI0GHujJSgiRCwnqCuH27dtotVoCAwOzTA8MDOTMmTM5LjNr1iymTp1aZGVatgySk4ts9XotWxb9NoQQtm/rVrh0qXDrKFUKfH1NUhwh7JoEdWY2YcIExowZo38fHx9PSEiIydbftq3JViWEEIVWrZr6EkIUPQnqCqFUqVI4OjoSFRWVZXpUVBRBQUE5LuPq6qrvOBnUzpTB9I9hhRBCCGtw7/qmFKYZtDCIBHWF4OLiQuPGjdm+fTu9e/cGQKfTsX37doYNG2bQOhISEgBMmq0TQgghrE1CQgI+Pj6WLoZdk6CukMaMGcPAgQNp0qQJzZo14+OPPyYpKUnfGjY/wcHBhIeH4+XlhSaf5l33HtWGh4fj7e1tiuJbhL3sB9jPvsh+WBfZD+tjL/tiif1QFIWEhASCg4PNsr3iTIK6QurXrx+3bt1i0qRJREZG0qBBA7Zs2ZKt8URuHBwcKFeunFHb9Pb2tumTyj32sh9gP/si+2FdZD+sj73si7n3QzJ05iFBnQkMGzbM4MetQgghhBBFQUaUEEIIIYSwAxLU2RBXV1cmT56cpfWsLbKX/QD72RfZD+si+2F97GVf7GU/RM40irQxFkIIIYSweZKpE0IIIYSwAxLUCSGEEELYAQnqhBBCCCHsgAR1QgghhBB2QII6IYQQQgg7IEGdEEIIIYQdkKBOCCGEEMIOSFAnhBBCCGEHJKgTQgghhLADEtQJIYQQQtgBCeqEEEIIIeyABHVCCCGEEHZAgjohhBBCCDsgQZ0QQgghhB1wsnQBirvMzEyOHTtGYGAgDg4SYwshhLAvOp2OqKgoGjZsiJOThB1FyaqO7p49e/jyyy+5ePEiP/zwA2XLluWbb76hYsWKtG7d2tLFKxLHjh2jWbNmli6GEEIIUaT+/vtvmjZtauli2DWrCerWr1/Piy++SP/+/Tl27BhpaWkAxMXFMXPmTDZv3mzhEhaNwMBAQP2ylylTxsKlEUIIIUwrIiKCZs2a6a93ouhYTVA3ffp0Fi1axIABA1i7dq1+eqtWrZg+fbpFyrR7924+/PBDjhw5QkREBBs3bqR37976+YqiMHnyZBYvXkxsbCytWrXiiy++oGrVqgZv494j1zJlylCuXDlT74IQophKTtWhKFDCXap1GEUbDZoS4OBq6ZLYHaliVPSs5gifPXuWtm3bZpvu4+NDbGys+QsEJCUlUb9+fT777LMc58+ZM4cFCxawaNEiDh48SIkSJejWrRupqalmLqkQQty3YUcCfd6+zpNvXWPd7/GWLo5tULQQ+QacLwkXQyD5L0uXSAijWU2mLigoiAsXLlChQoUs0/fu3UulSpUsUqbu3bvTvXv3HOcpisLHH3/Me++9xxNPPAHAypUrCQwMZNOmTTz77LM5LpeWlqZ/tAyQkJBg+oILIYqtnUeTWfh9DE+09cTJScOXG2Mp5etIp6YlLF006xY9H2IXQenZkPgLXO8Dlc6Do7elSyaEwawmU/fqq68ycuRIDh48iEaj4caNG6xatYpx48bx+uuvW7p42Vy+fJnIyEg6d+6sn+bj40Pz5s3Zv39/rsvNmjULHx8f/atWrVrmKK4QohiISdAyf3U07Rp5MKKfH2885UvHJh4sWBdDfJLW0sWzXpm34PYk8BsJJd+G4DWgvQ0JP1i6ZEIYxWqCuvHjx/P888/TqVMnEhMTadu2LYMHD+a1115j+PDhli5eNpGRkQDZKn4GBgbq5+VkwoQJxMXF6V+nTp0q0nIKIYqPVb/FoVMURj3rh0ajQaPR8EZfP7Q6hZW/xlm6eNYr5lNAA6XeU987lwO3ppD8h0WLJYSxrOLxq1ar5a+//mLo0KG8+eabXLhwgcTERGrVqoWnp6eli2dSrq6uuLrer4AbHy/1XYQQhRd5J5Of9ybyYncffDwd9dP9vR15tos33/wWR7+u3pT2tYrTvvVQMtXHrr6vgGPJ+9MlqBM2yCoydY6OjnTt2pWYmBhcXFyoVasWzZo1s+qALigoCICoqKgs06OiovTzhBDCXFZujqOEmwNPdfDKNu/J9l64Omv4/g+pw5tN0nbQ3gKfgVmnu9WH9HOgS7FMuYQoAKu5ZatTpw6XLl2iYsWKli6KQSpWrEhQUBDbt2+nQYMGgJp1O3jwoFXWARRC2K/rNzPYdjCJ//Xxxd0t+716CXcH+nTwYt3vCTzfzRtfL8cc1lJMxa8Bl2rg2ijrdNc6gA7Sz4JbA0uUTFi57du3s337dm7evIlOp8syb9myZRYpk1Vk6kDtp27cuHH88ssvREREEB8fn+VlCYmJiRw/fpzjx48DauOI48ePExYWhkajYdSoUUyfPp2ffvqJf//9lwEDBhAcHJylLzshhChq326Jx9fTgZ6tc3+68WR7LzQO8MOfkq3TUzIh8Ufw6gcaTdZ5znd7Xci4bP5yCas3depUunbtyvbt27l9+zYxMTFZXpZiNZm6Hj16ANCrVy80D/y4FEVBo9Gg1Zq/5dbhw4fp0KGD/v2YMWMAGDhwICtWrOCtt94iKSmJIUOGEBsbS+vWrdmyZQtubm5mL6sQoni6fjOD3/9O4vU+vri65H6f7uPpSO92XmzYmUCfDl74e0u2jpS/QRcLnj2yz3MsrXZCnH7J7MUS1m/RokWsWLGCF1980dJFycJqgrodO3ZYugjZtG/fHkVRcp2v0WiYNm0a06ZNM2OphBDivqU/xeHn5cjjeWTp7nm2ixe/7Eng29/iGNHP3wyls3JJW8HBT20U8TCNBlwqSqZO5Cg9PZ1HHnnE0sXIxiRBXVhYGCEhIVkybKBm2cLDwylfvny+62jXrp0piiKEEMXGkTOp7DyazPiBJfPM0t3jXcKRZ7t6s+KXOHq28aRisIsZSmnFkrZCic6gySVr6VwRMi6at0zCJgwePJjVq1czceJESxclC5MEdRUrViQiIoKAgIAs06Ojo6lYsaJBj053796d5/ychhATQojiKiwqg5krbtOwuitdmnkYvNxTHbz44+9k3l96h7mjAvDzckSnUzgfns6R06n8dzmduEQtwaWc6NXWizqV7XQMVG0spB4C31dz/4xTKKTsNFeJhA1JTU3lq6++4o8//qBevXo4OztnmT9v3jyLlMskQd29em8PS0xMNLh+Wfv27bNNe3CdlqhTJ4QQlqYoCjHxOsKiMgiPyiAsKpOL19I5cSGNcqWdeO/lUjmef3Pj6uLApMGlGPtJFP0n3qBMKSdux2pJSNbh4aahZgVXQgKdOXMljRFzo3illw/Pd/M2ahs2IWU/oAOPPJ4SOQdD/A2zFUnYjhMnTuh7vjh58mSWeZb8rRQqqLvXcECj0TBx4kQ8PO7fLWq1Wg4ePKjf6fw83FokIyODY8eOMXHiRGbMmFGYYgoh7EBcopawyAz8fRwpW9o5/wUsJD1D4cSFVNxcHKhV0QUHB+NP8PFJWrYfSmb/vymcC0snPkntLsHBAYJLORES6MyIZ/zo3KwEHjl0YZKfCmWcWTyhDNv+TiIqOhN/b0fqVXGldiVXnBzV8up0Cis3x7H0pziSUhVefcLHvgK7lL/UxhDOlXP/jFMw6KJBlwoORdMALj1D4fe/kzgflk7FYGe6P+KJi7MdHWc7ZY3tAKCQQd2xY8cA9U7y33//xcXlfv0MFxcX6tevz7hx4wxal4+PT7ZpXbp0wcXFhTFjxnDkyJHCFFUIYYPuxGnZvC+RPceSuXAtQz+9VxtPhj3jpw9ArMXlG+lM+vI2129lAlCzggsTXylFUEnDTrWZWoU12+JZ93s86RkK9au60aeDFxXKOBMa5ExwaSecnUyzz/4+6kgTuXFw0PDS476UcHfgi/WxeLhqeKF79vO0zUrZB+6tsndl8iCnYPXfzAi10YSJRdzO5J0vbhEemUFIkDO//pXIpl0JvPdyKSqXK+b1HW3ItWvXAChXrpyFS1LIoO5epDpo0CA++eQTvL1zP0EUVGBgIGfPnjX5eoUQ1is8KoPVW+PZfigJJycNLeu680xnbyqVdeb4uTS+WB9DWobCWy/6W0326PrNDEbPv0kpX0cWjQ8iMVnHR6vuMPyjKD4eHUDZgLyzi1cjMpj19R0uXEvnqQ5e9OvibRXdjjzdyZuUNIVlP8fhVcKBJ9pmH7HC5igZkHIQSk3N+3P6oO6GyYO6iNuZjJoXhYuzhsXvBlEx2IXLN9KZtUL9zowfWJK2DQ2vKynMS6fTMX36dObOnUtiYiIAXl5ejB07lnfffRcHB8t0A2ySOnXLly8v9DpOnDiR5b2iKERERDB79myDH+EKIWxbTIKWlb/G8fPeREp6O/JKL18ea+2Jp/v9E2Slsi54l3Bg5oo7VA91oXc7ywcZqek6Ji++jXcJB+aNCsC7hBqMLXwziNHzoxi74CYfjw7MMWOn0yls3JnA4h/jCPJ35LM3A6keal2NE17s7k18ko4F62Lw8XSkfSMbDzZS/wElGdzz6ZLiwaDOlJtP1zHxy1s4OWn4eEwgJX3U70vFYBcWjAvkw2+imbL4NoN6+vDCo3ZYn9EOvPvuuyxdupTZs2fTqlUrAPbu3cuUKVNITU21WLUxkwR1SUlJzJ49O9fhMi5dyr/zxgYNGqDRaLL1C9eiRQuLDbchhDCP1HQd329PYO22eBwd4NUnfHmyvVeudYs6NyvBf5fT+HJDLI1quFE+0LJ17BZ+H8P1m5l89lagPqAD8Pd25KORAYyad5MxH0cxb1TWwC4qOpM539zh2Nk0+nTw4tUnfAzqmsTcNBoNbzzlS1yilpnLb+PlEUDjGjbcyXrKPtC4glvjvD/n4AsaN5MGdYqiMHdVNDduZbLwzfsB3T1uLg6893JJQss4s/znOG7cymTM8/4me+wuTOPrr79myZIl9OrVSz+tXr16lC1bljfeeMO2g7rBgweza9cuXnzxRcqUKVOgu4rLl7N28Ojg4EDp0qVldAYh7JhWp7B1fxLLf4kjLlFL73Ze9H/UGx/P/B87vvakL0dOpzJrxR0+HRdosfp1e48ns/mvJMb296dS2ez1oEr7OjFvVABj5kcxen4U4weUpFI5FzbtTGDt7/F4ujvw0YgAGll5kOTgoOGtF0uSkKRj0pe3mDcqwOoyigZLPQKu9cEhn/JrNGq2zoRB3U+7E9l+KJn3Xi6Z4/dF3ayGAT18KFvaiTnf3CHyTiZTh5TKcsMgLCs6OpoaNWpkm16jRg2io6MtUCKVRslryAQD+fr68uuvv+pTkAWxcuVK+vXrh6tr1h9Zeno6a9euZcCAAYUtplW6du0aISEhhIeHW0UlSyHM5fTlND5eG8358Aw6NPFgcC9fypQy7j7z9JU0hn8UxXNdvXmll2/RFDQPd+K0vDI9grpVXJk2JO+uRaKiM3l/6W1OXU4HwNkJerfz4sXuPnh6WF92LjcpaTrGfXKTG7czWTA2kBALZ0kL5HIDcGsGZb7K/7NXW6vjwAavLPRmz4enM+zDSB5r5WnwiB7/Xkhl4pfqo/1Zb5TOt26mNbLH61zz5s1p3rw5CxYsyDJ9+PDhHDp0iAMHDlikXCYJ6ipWrMjmzZupWbNmgdfh6OiYYwfGd+7cISAgwG77qbPHL7sQeUlM1vHVplh+/SuRyuWcGdnPn9qVCp7xWbUljmU/xzF3ZAANqpkv25WpVXjr05uERWaw9L0yBmUXtTqFo2dSiUnQ0aSGG/4+tpl5iUvUMmpeFEmpClOHlKJmBRvK2CnpcNYTAueD39D8P3+tDygpEPJboTablKLjf7Mj8XDT8Om4IKO6Lbl+M4MJn98iPknH1CGlqF/VurO6D7PH69yuXbt47LHHKF++PC1btgRg//79hIeHs3nzZtq0aWORcpnk9vD9999n0qRJJCcnF3gduXVgfO3atRy7OxFC2J5/zqUyeEYEOw4nMbSvH1+8FVSogA7g2a7e1K/iyvtLbxMWmZH/AiaQqVX48Js7nLyYxsRXShkU0AE4OmhoWsudrs1L2GxAB+Dj6ciHIwIo7evIyLlRfL89Hp2u0PkB80g7A2Soj18N4VQaMm8WerNfbYwlOl7LpFdKGd0PXdkAZxa+GUilYGfGfHyTT7+LJilFl/+Cosi0a9eOc+fO8eSTTxIbG0tsbCx9+vTh7NmzFgvowESZuoYNG3Lx4kUURaFChQrZhss4evRonstqNBr++ecfateujZPT/ccvWq2Wy5cv8+ijj/Ldd98VtphWyR7vYIR4mKIofP1rHN/8Fk+9Kq68PaCkwX23GSIuUcvo+Te5HZvJs1286dK8BKX9DF+/TqdwNTKDM1fSOReezp1YLSlpCk6O4OykwdVFg4uTBhcXDTotnLyURlhkBuMHlqRT0xIm2w9bk5Gp8NWmWNb/mUDD6q689WJJAv1N9/9aJOJWQcQLUDUWHA1IGNyaCHFfQ5WwAm/y7NU0Xv8gipH9/HiiEK21tVqFDTsTWP5zHJ4eDgx/xo/W9d2tvnWsNV3nMjMzmTlzJi+//LLFy1IUTPLr6927d6GXPX78ON26dcPT01M/z8XFhQoVKvDUU08VsoRCCEvRahU+WHmHPw4lM6inOuSUYwFGWciLj6cjH48JYPGmWL75LZ6lP8fRoJorLeq406SmG+WDnLNsMyZBy5kr6Zy6nMbpy2mcvZpOUqqCRgPlg5wJ9HPE08OBTK1CeoZCYrKO9EyFtAwFDRAa5My4F/xt67FjEXB20jC0rx8t67jzwco7DJ4RwfBn/OnSzMN6A42My+BY0rCADsAxALQ3QVHy7qg4D6u3xlMuwInH23jm/+G8iuKo4elO3rRp4MGCddFM/uo2Leu6M/wZP5PeJNkzJycnPvzwwwLV0z9x4gR16tTBwcEhWzdsD6tXr15Bi1goJsnUmcLXX39Nv379TNLaNSMjg8jISJKTkyldujT+/oZVSLUEa7qDEcLUFEXho2+j2XYwiXcHlaR946LPaiWl6Nh9LJk/Dydz4kIqGZlqo4TSfk44OkBMvJbEFPW05+ftQK0KrtSs6ELNCq5UD3Up0LBbQq0ruWBdNH8cSqZeFVeGP+NnnaMiRLwKacegwmHDPh+/Dm48C1XjwNH4DvbDozJ4aVoEY57357FWhQvqHqQoCnuOp/DpdzEkpegY8JgPfTt6Wd0oK2B917knnniCPn36MHDgQKOWc3BwIDIykoCAABwcHHLshg3U1suWagdgstA+NjaWH374gYsXL/Lmm2/i7+/P0aNHCQwMpGzZsvkub+zBfVhCQgLffvsta9eu5e+//yY9PV1fT69cuXJ07dqVIUOG0LRp00JtRwhhuBW/xPHb/iQmDDRPQAdQwt2B7o940v0RT1LTdZy6lM6ViAxuxWSiU8DH04GypZ2pHupCoL+j9WaUbIynhwPvDCpFl+YpfPZ9DK/NiqR9Yw+e7eJNlRArCu4yroJTqOGfdywNwOXwa9yIr0DLOu5Gjef72/4kPN0d6NLMtN9/jUZD24YeNK7hxvJf4liyKZY//k5i9HOFa3hUHHTv3p3x48fz77//0rhxY0qUyPp/82Dfcw+6fPkypUuX1v9tjUySqTtx4gSdO3fGx8eHK1eucPbsWSpVqsR7771HWFgYK1fm3xRcq9Uyf/58vvvuO8LCwkhPT88yP69+X+bNm8eMGTOoXLkyPXv2pFmzZgQHB+Pu7k50dDQnT55kz549bNq0iebNm/Ppp59StWrVwu62SVjbHYwQpnLoVApvL7zFK7186P+oNHYqTjK1Cr/sTeS7P+KJvKOlQhlnGlZ3pUIZZ8oHORMS6Iyfl4NlAupL1aHEYxA4z7DPp52Ey3UZ8cMGTt5ozFMdvRja18+gRbU6hWffvUHr+u6MfLZonxidC0tn3upozoen878+vjzdyfTDdhaUMde5KVOmMHVq1uHbqlevzpkzZwBITU1l7NixrF27lrS0NLp168bnn39OYGCgweXJawgvQ7Nsu3fv5pFHHsnSDgDUOnv79u2jbdu2BpXF1E8WTZKpGzNmDC+99BJz5szBy+t+JdAePXrw/PPPG7SOqVOnsmTJEsaOHct7773Hu+++y5UrV9i0aROTJk3Kc9lDhw6xe/duateuneP8Zs2a8fLLL7No0SKWL1/Onj17rCaoE8IeRcdpmf31HZrUdOO5rtZzcRHm4eSooXc7L3q29mTfvynsO5HC4dOp/Lg7kXsDDpVw1xAS6Eyj6m50aupBxWAzZPMUnZqpc65g8CK3E/0pBTSvnkCr5r58uTGWLs1KUK18/uU9eiaVO3FaurYo+ix1tfIufPZWIEt/iuOL9bH4ejrSpbltNuKpXbs2f/zxh/79g4HT6NGj+fXXX/n+++/x8fFh2LBh9OnTh7/++svg9T886lVBdOjQIcdu2OLi4ujQoUOegWFRPlk0SVB36NAhvvzyy2zTy5YtS2RkpEHrWLVqFYsXL+axxx5jypQpPPfcc1SuXJl69epx4MABRowYkeuya9asMWgbrq6u/O9//zPos0KIgtHpFGZ9fQc0MH5gSaMeVQn74uiooU0DD9o0UMeKTc9QuHE7k/CoDMKjMrhyI4Of9ySyems8zWu78WIPH2pVLMJHh9qboKSBs+GPXz/+zoFp7TU81S4F51Je/Lw3kdVb45jyaul8l912MInygU7UCDXP42dHBw2vPuFDTLyWeaujqVbehdAy1tNZcUJCAvHx8fr3rq6u2QYcADWICwoKyjY9Li6OpUuXsnr1ajp27AioY8/XrFmTAwcO0KJFi6Ir/ENy64btzp072R7nPujhJ4vvvPNOjk8Wu3btWqAniyYJ6lxdXbP8R91z7tw5/fPn/ERGRlK3bl0APD09iYuLA+Dxxx9n4sSJpiimEMIM1v2RwJEzqcwZXhp/b9vti02YnouzhgplnKnwQKCRqVXYcTiZNdviGfZhFF2aeTDkSb9sY6KaRMZV9V8Dg7p/L6Sy7990Mtv74+ZwGxw0PNXBi89/iCEmQYufV+5lTErRsfd4Ci/28DbrY2aNRsOoZ/3471Iac1dH8/HoAKu5sapVq1aW95MnT2bKlCnZPnf+/HmCg4Nxc3OjZcuWzJo1i/Lly3PkyBEyMjLo3Lmz/rM1atSgfPny7N+/36igLikpiV27duVY3SuvJFKfPn0A9Ti/9NJLWYJSrVbLiRMneOSRR3JdvqifLJokqOvVqxfTpk3T9yWn0WgICwvj7bffNrg7knLlyhEREUH58uWpXLky27Zto1GjRhw6dCjHSP5hf/75J8OGDePAgQN4e2d93BMXF8cjjzzCokWLLNopoBD27tTlNJb9FMtzXb1pUtPd0sURNsDJUUOX5iXo2NSD3/YlseTHWPb/e4Mxz/ubvnFNxhX1XwOCOkVRWPZzHJXLOePsGgjaWwB0bOLBF+tj+PNQEk91zL1qwe5jyaRnKiZvIGEIVxcHxvb3Z/T8m/y0J5HeRvSNp9MpHDmTStNapv/9njp1KkvDyZyu7c2bN2fFihVUr16diIgIpk6dSps2bTh58iSRkZG4uLjg6+ubZZnAwECDnwoCHDt2jB49epCcnExSUhL+/v7cvn0bDw8PAgIC8gzq7g2GoCgKXl5euLvfP04uLi60aNGCV199Ndfl16xZQ2ZmZr5lLPCTRcUEYmNjlc6dOyu+vr6Ko6OjEhISojg7Oytt27ZVEhMTDVrH22+/rcyYMUNRFEVZu3at4uTkpFSpUkVxcXFR3n777XyX79mzpzJv3rxc53/yySdK7969DdshMwoPD1cAJTw83NJFEaJQEpK0ynPvXVOGzolQMjJ1li6OsFGxCZnKlMW3lA6vX1VmLr+lJKVoTbfy2x8oyllvRdHl//08cjpF6fD6VeWvf5IU5Uo7Rbn+nH7exEU3lSEzb+S5/Mi5kcq4T6IKW+JCmbfqjtJjVJgScTvDoM/rdDrl0++ilQ6vX1UuhKeZrByFuc7FxMQo3t7eypIlS5RVq1YpLi4u2T7TtGlT5a233jJ4ne3atVNeffVVRavVKp6ensrFixeVsLAwpW3btsr69esNWseUKVMMjm8eFhwcrMyePVuJiYkp0PJ5MUmHTD4+Pvz+++/8/PPPLFiwgGHDhrF582Z27dqV57PlB82ePZt33nkHgH79+rFnzx5ef/11fvjhB2bPnp3v8v/88w+PPvporvO7du3KkSNHDNshIYRRFEVh7upoEpJ1vDuolFX2lSVsg4+nI5NeKcn4Af7s/SeF4R9FceN2/pkNg2RcVbN0+TwOVRSFZT/HUqOCCy3ruoNTQJahwro0L8H58AyuRuQ8LF3E7UxOXEijq4UbKgx50hdPdwfmr4nOsT+1BymKwpIf49iwI4GRz1pPH4O+vr5Uq1aNCxcuEBQURHp6OrGxsVk+ExUVlWMdvNwcP36csWPH4uDggKOjI2lpaYSEhDBnzhx9HJKfyZMnGxzfPGzUqFF8/vnnhISEMHLkSJN2j2LSXjZbt27NG2+8wVtvvZXlmXd+MjIyePnll7PsWIsWLRgzZgw9e/Y0aB1RUVHZhid7kJOTE7du3TK4TEIIw/15OJldR5MZ278kZUpJz/aicDQaDV1bePLZW0GkZSi88UEkJy+mFX7FBvZRd/BkKqcup/NyTx+1Pty9USXual7bHU93DX8cSspx+T/+TsLdVUPrBpatglDC3YHRz/tz6FQqWw/kXFZQA7rFP8axZls8rz/lyxNtCz6UmaklJiZy8eJFypQpQ+PGjXF2dmb79u36+WfPniUsLIyWLVsavE5nZ2d9tyYBAQGEhalDwPn4+BAeHm7wen744QeeeeYZWrRoQaNGjbK88vLmm29y6dIlvvrqKw4cOEC1atXo27cvBw8eNHjbuSlwULdgwQJSU1P1f+f1yo+zszPr168vaFEAtaXtyZMnc51/4sQJypQpU6htCCGyS0zW8fn6GNo2dKd9Iw9LF0fYkQplnPn8rUAqlHHmnc9vciWXzJjBMq7k252JTqew7JdY6lVxpXGNuyMcPZSpc3HW0LaRB9sPJWXLgCmKwraDSbRt6IG7q+VHJ2lRx50uzTz4/IeYHDOemVqFT9bGsPZuQGfp/u3GjRvHrl27uHLlCvv27ePJJ5/E0dGR5557Dh8fH1555RXGjBnDjh07OHLkCIMGDaJly5ZGNZJo2LAhhw4dAqBdu3ZMmjSJVatWMWrUKOrUqWPQOhYsWMCgQYMIDAzk2LFjNGvWjJIlS3Lp0iW6d++e7/L39ungwYP8+eefaLVaWrVqRatWrdi4cWO+mdVcFfS5bYUKFZTbt2/r/87tVbFiRYPWN2DAgDzrxOVn2LBhSp06dZSUlJRs85KTk5U6deoow4cPL/D6i4rUqRNmp9Mqyp25ipLws0lWt3hTjNJ9VJhyM8awejtCGCshWau8/P4N5cXJ15Xkgtax0+kU5aynotz+MM+P7TicqHR4/apy4vwD15LozxXltKP627nr+Fm1zt2/F1KzLP/PeXX6kTPZr0WWEpeYqbww6boycMp15Wb0/d/p1ch0ZcTcSKXz0KvKz3sSimz7xlzn+vXrp5QpU0ZxcXFRypYtq/Tr10+5cOGCfn5KSoryxhtvKH5+foqHh4fy5JNPKhEREUaV59ChQ8qff/6pKIqiREVFKd26dVO8vLyURo0aKcePHzdoHdWrV1dWr16tKIqir5enKIoyceJEZejQoUaV556LFy8qw4cPV7y9vZUqVaoUaB1WM/br9OnTmTt3Lp06dcpx2I68WqOA+vi1UaNGODo6MmzYMKpXrw7AmTNn+Oyzz9Bqtfphy6yJjCghzO7ObLg1Qf270gVwqVzgVSUm63juves83tqT1/oY1su+EAVx7WYGQ2ZF0q6hB28PKGn8CrTRcL4kBH8H3k/n/BGtwqD3Iyhb2olZQx/oVDZ+PdzoC1Vvg6O6bZ1O4fmJN2hRx51Rz90fBWDK4ltcvpHB8ollrKYrEVCP37hPbpKUqqNxDTfiE3X8cyGNoJJOvPWiP/WrFn7c9Vy3bYfXOQ8PD06fPk1oaCgBAQH8/vvv1K9fn/Pnz9OiRQvu3LmT67KTJ08mLi4ux1dsbCyxsbHodLoCjR9r8sov92JEY/vlWbp0Kb6+vhw5ciRbgwaNRpNvUBcYGMi+fft4/fXXmTBhQpZydOvWjc8++8zqAjohzE4bDXdmge9rEL8W4lZC6an5L5eLTbsSSM9U6GtFQxIJ+1QuwJlR/fyYvTKaxjXc6GxsVyEGdGfy4+5Ert/KZOIrpbLOcLob4GXe1Ad1Dg4aOjbxYPO+JIY+7Yezk4bIO5lq446n/awqoAP1+C0aH8R32xM4ezUNTw8HRj/nT5dmHri6WP4xsa0JCgoiOjqa0NBQypcvz4EDB6hfvz6XL1/O99Hp+++/j5ubGy+99BKNGjXCx8cHb29vvL299X/f6zrFWCYL6pYuXcr8+fM5f/48AFWrVmXUqFEMHjzYoOVN0fojNDSUzZs3ExMTw4ULF1AUhapVq+LnJxkEIQCI+xqUVCg1FXTJkLipwEFdSpqOH/5MoMcjnkXTUawQD+nawpPDp1P5eG00tSq5EmxMoxx9x8MVcpwdeSeTpT/F0quNJ1VDHmr56Xg3qNPeAmrqJ3dpXoK1vyew43ASXVt48vWvcXh7OFi81WtufL0cGdLb19LFsIiGDRsanGw6evRovp/p2LEjP/30Ew0bNmTQoEGMHj2aH374gcOHD+s7KM7N9u3bmTt3LsuWLePZZ59l3LhxBtfly49JgrpJkyYxb948hg8frm+Bsn//fkaPHk1YWBjTpk0zeF3p6elcvnyZypUrZxso11B+fn4FGjNNCLsX9y2UeBycAsGjA8R/C9pYcPQ1elW/7E0kKUVHvy6SpRPmM+JZf/67FMHM5bf5ZEwgjoZ2n5NxFTTu4Jh9lKO4RC3vLbqFj6cDrzzhm33ZBzN1D6gY7EKr+u4s/yUON1cHth1MYvgzfri7SebL2vTu3duk6/vqq6/0Y8gOHTqUkiVLsm/fPnr16sVrr72W57IdOnSgQ4cOnD17lnnz5tG8eXPatGnDm2++SadOnQpXsALVxHtIqVKl9BUGH7R69WqlZMmSBq0jKSlJefnllxVHR0fF0dFRX+lw2LBhyqxZs0xRTKskDSWE2aSeUpTTKEr8BvV92jn1fcJmo1eVlq5T+o6/psz++raJCylE/k5eTFU6Db2qLPspxvCFIkcpysXq2SZfup6mPP/eNaX3m+HKxWu5dLir0ynKaSdFif4s26xrN9OVJ8aFKx1ev6qMnh+pZGql4+2HyXUubzdv3lQmT56sBAYGKg0aNFC+/fZbJTMzs0DrMsntREZGBk2aNMk2vXHjxgYNhwEwYcIE/vnnH3bu3Imb2/0Km507d2bdunWmKKYQxVviL2qmosTd5vbOVdT6QamHjF7VtoNJRMdrea6bZOmE+dWu5MrAx3xYtSWef86lGrZQxpVs9en+OpHMsA+jcHdz4Iu3g6hUNpcOdzUacCqdLVMHULa0Wlft3UElmflGaRytrC6dyN2RI0f49ttv+fbbbzl27Fi+nz9x4oQ+O3fixIk8X8YoXbo0U6ZM4cyZM/Tp04cRI0ZQqVKlAu2TSR6/vvjii3zxxRfMmzcvy/SvvvqK/v37G7SOTZs2sW7dOlq0aJHluXft2rW5ePFinstmZGSwcuVKfVlcXKyjJ2whrErS7+DRDhzu3jRpNOBaD9L+NWo1mVqFNVvjaNPAnfKBuXf4LURRer6bN8fOpDJ58W3mjw6gYnA+5/2Mq+CuVstJz1D4alMsG3Yk0Kq+O+8MLJn/I9OHOiB+UFBJJ4JKSqfbtuLmzZs8++yz7Ny5Uz+ObGxsLB06dGDt2rWULp39ET1AgwYNiIyMJCAggAYNGqDRaHJsFKHRaPJsufrUU0/l2PI1IyNDv76HR80wVIG/hWPGjNH/rdFoWLJkCdu2bdN3AHjw4EHCwsIYMGCAQeu7desWAQEB2aYnJSXlW7lx3LhxdOnSBUVRePPNN/nkk0+M2BMhigFdKqTsgVLTs053rQeJm41a1fZDSUTc0TLttYK1zhLCFBwdNEx9rTRjPo5i9PybTHipJM1r5zGCQ8ZV8OrLP+dSWfBdDNduZjC0ry99OngZVoHeMeBuQwlh64YPH05CQgL//fcfNWuqDV9OnTrFwIEDGTFiBGvWrMlxucuXL+sDvsI07vTw8CA4OBhfX988XwVR4KDu4VRl48aNAfRZtVKlSlGqVCn+++8/g9bXpEkTfv31V4YPHw7c7xJlyZIl+Q7/odPp9H263EuNCiEekPKX2uq1RJes013rQcwC0CWBQ/4t9rQ6hVVb4nmknrvVjA0pii8vDwfmjgxg5oo7TPjsFq3qudO7vRf1q7pmGX84Lj4WH100a3b4snjHTWpVdOHzt4KM+w47BUCG4UNICeu1ZcsW/vjjD31AB1CrVi0+++wzunbtmutyoaGhOf5trG+++abAy+anwEHdjh07TFkOZs6cSffu3Tl16hSZmZl88sknnDp1in379rFr1648l507dy6rV69GURQ++ugjk5ZLCLuQ9IeaaXB9qNm8az1AgbT/wL1ZvqvZ/ncS125m8u6gAnT+KkQR8C7hyMzXS7PjSDJf/xrHmwtu4uwEgf5OODtpiInX4utymmUvwLWYcrw7qCQdGnsY34+cY2lIPZL/54TV0+l0OY4V7+zsnGdi6KeffjJ4G7169cpxelhYGOXLlzd4PdevX6ds2bIGf95qKgG0bt2a48ePM3v2bOrWrcu2bdto1KgR+/fvp27dunku6+LiwksvvWSeggphi5J/hxKdQfNQvSHXGuq/6WfyDepSUnUs/jGOtg3dqR7qWkQFFcJ4Dg4aOjUtQccmHpy9ms6py+lERWei1Sr4eDlSNzAegDcHNQLnAvYh99D4r8J2dezYkZEjR7JmzRqCg4MBNXgaPXp0nl2KPNwtysN16h58jJ9bnbqmTZvSu3dvBg8enGvXa3FxcXz33Xd88sknDBkyJN/BFx5kkqAuNTWVTz/9lB07dnDz5s1ska4hHfkBVK5cmcWLF5uiSEKIe7R3IPUo+A7LPs/BE5zKQvo5FEVBp5Br672lP8USn6TlfzIcmLBSGo2GGhVcqVHhoZuOmCiIcgKnMgVfuWMA6KJByQSN1eRDRAEsXLiQXr16UaFCBUJCQgA1g1a3bl2+/fbbXJd7MLb5448/ePvtt5k5c2aW/nnfe+89Zs6cmes6Tp06xYwZM+jSpQtubm40btyY4OBg3NzciImJ4dSpU/z33380atSIOXPm0KNHD6P2zSTfzFdeeYVt27bRt29fmjVrZvQQYfdotVo2btzI6dOnAfUZ9xNPPFHgToiFEEDSn4CiZupyoHWqzuWL/zJq1jXS0hWqlnehW/MSdGjigXcJdaSI3w8msWFnIkP7+korP2F7Mq6AcwhoCjHyiX5UidvgFGSSYgnLCAkJ4ejRo/zxxx+cOXMGUOMNYzr+HTVqFIsWLaJ169b6ad26dcPDw4MhQ4bo45iHlSxZknnz5jFjxgx+/fVX9u7dy9WrV0lJSaFUqVL079+fbt26FXiECZOcnX/55Rc2b95Mq1atCryO//77j169ehEZGUn16tUB+OCDDyhdujQ///xznjtY1M+ohbBpSb+DSw1wzj6Q9q2YTP79rxwVfA7Rp4MX/t6OHD6dyqffx/DZDzHUr+qGVqdw/Fwaj7YsQZ8OXhbYASEKKeNqrsODGezBUSUkqLNJ+/fv586dOzz++ONoNBq6dOnCjRs3mDx5MsnJyfTu3ZtPP/0UV9f8q5dcvHgxxxaqPj4+XLlyJd/l3d3d6du3L3379i3AnuTOJJ0Ply1bFi+vwp3sBw8eTO3atbl27RpHjx7l6NGjhIeHU69ePYYMGZLnsk2bNuW1117j0KHcO1GNi4tj8eLF1KlTh/Xr1xeqrELYlOQ/srd6BaKiMxk1/ybXYipRodQVXn7cm97tvJj+v9J8N7MsQ570xcVZg4ebA+MH+DOuv3+Bs/BCWFTG1WwdDxvt3vBiufRVJ6zftGnTsvTI8e+///Lqq6/SpUsXxo8fz88//8ysWbMMWlfTpk0ZM2YMUVFR+mlRUVG8+eabNGuWf6OzomKSTN3cuXN5++23WbRoUYGb+R4/fpzDhw/j53e/vo6fnx8zZszIdxzXon5GLaxPSpqOS9czqFDGmRLuMs5irtIvQsZl8Mj66DXyTiZjP1ZPRo93bIRDfApkXgNnNePt7+1I347e9O1o9hILYXoZV6HEo4VbRy7jvwrbcfz4cd5//339+7Vr19KsWTN9Xf6QkBAmT57MlClT8l3XsmXLePLJJylfvry+Xl54eDhVq1Zl06ZNRVF8g5gkqGvSpAmpqalUqlQJDw+PbE2Fo6Oj811HtWrViIqKonbt2lmm37x5kypVquS5bFE/oxbWZfuhJD77PobYRB0+ng5MG1KKulXc8l+wOEr6HXACj/b6SVHRmYz5OAqNRsO8UQH4e9aCeCD9rD6oE8Ju6FJBG1n4x6+aEuowe9IBsc2KiYkhMDBQ/37Xrl10795d/75p06aEhxvWF2GVKlU4ceIEv//+u75eXs2aNencubNFn2iYJKh77rnnuH79OjNnziQwMLBAOzRr1ixGjBjBlClT9KNSHDhwgGnTpvHBBx8QHx+v/6y3d87jTRbVM2phPb7fHs8X62Pp0NiDJ9t7seynWCZ8fosvJwRRtrQMWZVN0jZwbwmO6m8m4nYm4z6JQgPMHxVAgL8TKBUAZ0g/l+NjWiFsWmaY+m9hH79qNHkOFSasX2BgIJcvXyYkJIT09HSOHj3K1KlT9fMTEhJy7L8uNxqNhq5du+bZYbG5mSSo27dvH/v376d+/foFXsfjjz8OwDPPPKMPCu/1/9KzZ0/9+/zGVBP2SVEUVvwSxze/xfN8N29e6eWDRqPh/f+V5rXZkcxcfocF4wJlMO0HKRmQvB383wTgwMkUPlh5hxLuDnw04m5AB2qLQJcqaqZOCHuTfnc4J+eKhV+X9FVn03r06MH48eP54IMP2LRpEx4eHrRp00Y//8SJE1SuXNng9SUlJbFr1y7CwsJIT0/PMs+YvuVMySRBXY0aNUhJSSnUOkw9QoWwHzqdwuc/xLBhZyKv9vblua73M7Ul3B0YP6AkI+ZG8eveRHq1LdrWmWnpOk5dTufEhTQibmcSn6RF98B4zg+HlHklre/dvDg4QJVyLnRvWeJ+oFVI6RkK167upZIunmV/NmP7yRtE3M6kaS033nmpJD6eD3Xt4FJdgjphnzIuA445tv42mmNpydTZsPfff58+ffrQrl07PD09+frrr3FxuT9U3LJlywzOuh07dowePXqQnJxMUlIS/v7+3L59Gw8PDwICAmw7qJs9ezZjx45lxowZ1K1bN1v6MrfHpQ9q166dKYoi7IxWqzB3dTRbDyQx8lk/nsghaKtT2ZXuLUuw9Kc42jb0wNerEH1RPSQxRceZK2mcuJDGP+fTOHMljYxM8C7hQPkgZ3xKOOB8t53GAx2Lq+9zfZP1bVq6wg/b41m1JY5nOnszoIcPLs6GZxx1OoVrtzI5czmN01fSOXMlnYvX03mhySZK1/fh2NXatKzrTpMabjSv45Zz9QiXGhCf8yDWQtg0fR91JrjcOQXIzY8NK1WqFLt37yYuLg5PT08cHbNeK77//ns8PT0NWtfo0aPp2bMnixYtwsfHhwMHDuDs7MwLL7zAyJEji6L4BtEoysOXIuM5OKhXtYcvFsY+Lk1NTeXEiRM5jkqR2zhqtu7atWuEhIQQHh5OuXImuJO0IylpOt5fepu/T6UyfkBJOjfLfXif2AQtA6dG8Eg9d94ekPu4pBmZCr8fTGL74STOh6WTlKrgoAFHRw1OjuDkqMHREZwcNKSk6UhMUX8ePp4O1K3sSoNqbtSv6krFYGfjx47MQ3Kqju+3J7B6axyB/k6MetafRjVybvwRk6Dl9OU0zlxN5/TldM5eTdOXs3ygEzUquFKzgguPBrXA2bMRDmUNGDw6dgVEDoJqSeDgYbL9EsLirvdTGzeU/7Pw67r5NiSsh8oXCr+uYsQer3O+vr4cPHiQ6tWr4+vry/79+6lZsyYHDx5k4MCB+sYT5maSTF1ej07//fdfg9axZcsWBgwYwO3bt7PNk3p0xYuiKBw9m8bC76KJitEy8/XSNKvtnucyvl6OvPakLx+tiqZdIw9a1Mn6ea1ODeZWbo4jKlpLo+puPNvFG08PBxRAq1U/k6mFTK1CZqaCq4uGoJJOVCrrQmiQU5G2aPJwc2DgYz60b+TB/DXRjFtwk9b13enWsgTurg5cvKYGcGeuphF5R/0t+Hk5UKOCK8909qZGBRdqhLri6XE3bZh2Gi6fAm/D+ly6PwbsOXBrYPodFMJSMi6Da97jhxvMMQC0Ufl/Ttg9Z2dnfUIrICCAsLAwatasiY+Pj8EtaIuCSYK6hx+dJiQksGbNGpYsWcKRI0cYNiyHMScfMnz4cJ5++mkmTZqUpcmxKD5S03X88XcyG3cmcPlGBtVDXfj8rdJUKGNYa6Tuj5Rg97Fk5q6K5rM3Awnwd0JRFPb/m8KSH+O4EpFB6/ruzHjdh4rBLvmv0AJCyzgzf3QAv/+dzKotcUxcpN7kuDprqFbehTYNPKhZwYUaFVwJ9HfMPdCMWwYOXlDCwFZZLuooLqSflaBO2JeMy+DZ0zTrci4LukTQxutblIviqWHDhhw6dIiqVavSrl07Jk2axO3bt/nmm28s2n2aSQdx3L17N0uXLmX9+vUEBwfTp08fPvvsM4OWjYqKYsyYMRLQFUOXrqfz61+J/H4wiaRUhUfqujPsaT8aVHM1Kjum0WgY94I/wz+KYsS8KFrUdufo2VSu3cykQVVXPnsrkJoPD/RthTQaDV2bl6BLMw+iorVkahXKlHTC0dHAY5FxA2IWqq1eHQzsv8/RT81CpFvmkYEQRUKXqI7VaoqWrwBOaiezZIaDY+28Pyvs2syZM0lISABgxowZDBgwgNdff52qVauybNkyi5Wr0EFdZGQkK1asYOnSpcTHx/PMM8+QlpbGpk2bqFWrlsHr6du3Lzt37jSqObGwXRmZCn8eTuLnPYmcupyOn5cDPdt48VhrT4JLFfxrWcrXifmjA/n8hxj+OZ9K7UqujHvBn7qVjQsQrYFGoz7+Ndqd6Wonqf5jjVvOpQakSVAn7EjGFfVfUwV191rQZoSDqwR1xZWiKAQEBOgzcgEBAWzZssXCpVIVKqjr2bMnu3fv5rHHHuPjjz/m0Ucf/X97dx4dRZUvcPzb6SUbCQnZAwFZAoQBwhJBlEWG8IDRcTIiRoUhQGR4PiJoZlTAx6aDAqNOZERggOjjwCDOAREXCJiRfY5iIBD25UVZXhYIRBISsnTf90dMD20WsnRSneb3OacP6arqW79f0VR+3Lq3Cr1ez8qVK+vd1nvvvce4cePYt29ftTNotZoeLOzLbFZs3VPAR7sKyPvRTFSEGwum+jOolztGg32KrmA/A69NC7BLWy1O6QXIXw0Bi0Dfun6fde0OxTU/P1mIFqf0pwkNpk72ac8QCugqHqkn7llKKbp06cKJEycIDw/XOhwbjSrqtm/fzowZM6xdjo2xceNGdu7ciZubG7t377bpVdHpdFLUOZDiEgulZarqvc7u4sLlUv68/jrnL5UycqAnsSO96zxeTtTRtQVgCADfu49jrcLUDX5cD8oCOnmernACpafBpTXog+3Tns4IhuCKnjpxz3JxcSE8PJy8vDznKur279/P2rVr6d+/PxEREfzud7/jqaeealBbr776KgsXLmTWrFnWGSXCsRQWWVj5yQ12fXOLsnIIDzMy7XFf+nWrfdxWaZliw44f+XvKTcKCjPz1pZYxtq3FuZ0BNzdA0PKG3ZbE1B1UUUUvhDwDVjiDklMV32t7Dr0whElPnWDx4sW89NJLrFixwqGeK9+oou6BBx7ggQceICkpiU2bNpGcnExiYiIWi4Vdu3YRFhaGl1fd7vBfWlpKbGysFHQOqrDYwh/ezeH/rpUz+VEfAnz1fLq3kD++m8v40d5MerR1tY/oOpVZwtL117mcU8b40d6MH93abpdZxc9c+++KsUM+8Q37vOmn25qUnJKiTjiH0tPgWvex3XViDJOeOsHEiRMpKioiMjISk8mEu7vtbbSuX7+uSVx2mf3q6enJlClTmDJlCmfOnGHt2rUsXryYWbNmMXLkSLZt23bXNuLi4ti0aRNz5syxR0jCjopLLMx5/ypZ18pJejGIzu0qbgfyyygPPtpVwNpP8znxvyXMjvMjwLfiK3Wr2MKabfls21tI1zATq2YH06mtY95GxCnc+hoKt0HIBtA18Dgb7wOXVlByFFqNsmt4QjQ7pSqKOq/H7duuoR2UOMageKGdpKQkrUOoll1vaQLQrVs3li5dyptvvslnn31W56m9ZrOZpUuXkpKSQu/evatMlHjnnXfsHaqog9IyxbxV17hwuZS3ZgRaCzqoGOv49H94072DiTf/J4/4P2Xx6OBWGA06tu0rpKRM8V9jfYh52KvaXjxhJ8oMuYngNhC8n254OzoXcO0Pt2WyhHAC5ZfBchNMEfZt1xQON94HVW6fR4+JFqesrIw9e/Ywd+5cOna008xqO2myb6RerycmJoaYmJg6bZ+RkUHfvn0BOH78uM26lnYrCmdRbla8tvYaGRdKWDI9gIiO1Y+D69vNjTWvBvM/X/zIjn/doqxc8XB/Dyb+qrW15040oR/XQUk6tD/Q+LFDbv2hcItdwhJCU7e/q/jTrb992zVFAGVQ9r9g6mrftkWLYDQa2bx5M3PnztU6lCoc5jdubY8aE82vrLyioDt0spjXpgUQ2bX2yRDennqef7INzz/ZppkiFACU58LVl8D7GfB4sPHtuUfBjXeg/GrFLFohWqribytuQWJsa9927xx7KkXdPSsmJoatW7fy4osvah2KDYcp6iqdP3+eCxcuMHToUNzd3VFKSU+dHRUWW/i/q+UUFlvQ68DNVYebyQVXkw5Xkw5lgR+yy1j9aT4XLpeycGoAA+/y3FWhEaUg+z8BHQQm2adN98EVfxb9E7xj7dOmEFooPgBuA+zfriGk4jYpJcfB6zf2b18AsHz5cv785z+TnZ1NZGQkf/3rXxkwoAn+PhsoPDyc1157jQMHDtC/f388PT1t1mt1GzaHKery8vJ48skn+frrr9HpdJw7d45OnToRHx+Pr68vb7/9ttYhtji3ii2cuVjK2R9KOXuxlHOXSrlytbxOn20fZCApUW494rCUgmtzofATaLvVfr1qxjBwjYTCz6SoEy2X+ToUH6y4vY+96XTg/iAU7wVetX/7gk2bNpGYmMjKlSsZOHAgSUlJjBo1ijNnzhAYGKh1eACsXbsWHx8f0tLSSEtLs1mn5b11Haaoe/HFFzEajVy8eJGIiH8PbI2NjSUxMVGKujo6dmgZ1/PzyL1hJr/AAigMeujjpye6swF/Xxd8vfSYjICCsnIoM1swlyvKLQqAVu4utPF2wUWng2tUbFjFz5c1YBvVkHbt9Jkq+26iWJpqP6Vn4dYOCFhi/94C76fg2ny41g10dXx2bItQ3bFvonXVfrc1iONu62qM04FibEj8t9MAl6brSfMYAVdnw7XXnezfyB18poLex65NFhQUcPPmTet7V1dXXF2rdhy88847TJ06lcmTJwOwcuVKvvjiC5KTk5k1a5ZdY2qozMxMrUOolsMUdTt37iQlJYV27drZLA8PD+eHH37QKKqWx8e8ms5Bl9GHgIuLDr1LxZ82F7AtOii5470OMN6xhQXI//kl7+ougddhmyqXzu3Rbh0+U+0l+yaIpbn28/N96Twh9B/g/UQ1bTWSbwIU7oDrzjjjvLahHLWsq3UISE3rGrivZo2xtnUtPP7AJRVPf2gKPlOgYKOT/hv5idcTdi/qfv48+Pnz57NgwQKbZaWlpaSlpTF79mzrMhcXF6Kjo/nXv/5l13jqKzExkddffx1PT08SExNr3E6n02nWEeUwRd2tW7fw8Kh6F/zr169XW8mL6rUdcExuHyIax6UVdNitdRRCOC69L9z3ndZRtDgnT56kbdt/T1yp7nf7tWvXMJvNBAUF2SwPCgri9OnTTR5jbY4cOUJZWZn155poOQ/AYYq6IUOGsG7dOl5//XWg4qBYLBaWLl3K8OHDNY6u5ZCCTgghhCPy8vLC29tb6zAa7M67dDjqHTscpqhbunQpI0aM4LvvvqO0tJSXX36ZEydOcP36dQ4cOKB1eE3GYrEAkJWVpXEkQgghhP1V/n6r/H1XG39/f/R6PTk5OTbLc3JyCA5uosvpTsRhijpvb29OnTrFihUr8PLyorCwkMcff5zp06dbuzudUeUX15GmagshhBD2lpOTQ/v2tT9X2mQy0b9/f1JTU60PL7BYLKSmppKQkNAMUbZsOqVqnabVbPR6PVlZWVWmK+fl5REYGIjZbNYosqZVXl7OkSNHCAoKwsXFpdZtCwoK6NGjBydPnsTLy6uZIrQ/Z8kDnCcXycOxSB6Ox1ly0SIPi8VCTk4Offv2xWC4e1/Spk2biIuLY9WqVQwYMICkpCQ+/vhjTp8+XWWsnbDlMD11NdWWhYWFuLk56ZRxwGAwcP/999dp28qp4G3btm3R4xKcJQ9wnlwkD8cieTgeZ8lFqzzu1kN3p9jYWK5evcq8efPIzs6mT58+7NixQwq6OtC8qKucFqzT6Zg3b57NDFiz2cw333xDnz59NIpOCCGEEM0tISFBLrc2gOZFXeW0YKUUGRkZmEwm6zqTyURkZCR//OMftQpPCCGEEKJF0Lyoq5wWPHnyZN59990W3a3d1FxdXZk/f36Lv2+fs+QBzpOL5OFYJA/H4yy5OEseonoOM1FCCCGEEEI0XO3TLYUQQgghRIsgRZ0QQgghhBOQok4IIYQQwglIUSeEEEII4QSkqGsCkyZNQqfTodPpMBqNBAUFMXLkSJKTk+v07LvG2LJlCyNHjiQgIABvb28GDRpESkpKjdsvXrwYnU7HCy+8UGWdo+exYMECa3yVr+7du0seDp6HM+Xi6HkAXLlyhQkTJuDn54e7uzu9evXiu+++kzwkjybPQzQ/KeqayOjRo8nKyuL7779n+/btDB8+nJkzZ/Loo49SXl7eZPvdu3cvI0eO5MsvvyQtLY3hw4fz61//2no/wDsdOnSIVatW0bt37xabxy9+8QuysrKsr/3790seLSAPZ8rFkfO4ceMGDz30EEajke3bt3Py5EnefvttfH19JQ/Jo1nyEM1MCbuLi4tTv/nNb6osT01NVYBavXq1ddmNGzdUfHy88vf3V15eXmr48OEqPT3d5nPbtm1TUVFRytXVVfn5+amYmJh6xdOjRw+1cOFCm2UFBQUqPDxc7dq1Sw0bNkzNnDmzxeUxf/58FRkZedfPSR6OlYcz5eLoebzyyitq8ODBkofkoUkeovlJT10z+uUvf0lkZCRbtmyxLhs3bhy5ubls376dtLQ0+vXrx4gRI7h+/ToAX3zxBb/97W/51a9+xZEjR0hNTWXAgAF13qfFYqGgoIA2bdrYLJ8+fTqPPPII0dHRLTqPc+fOERoaSqdOnRg/fjwXL16UPFpoHs6Ui6PksW3bNqKiohg3bhyBgYH07duX1atXSx6Sh6Z5iCakdVXpjGr6X5ZSSsXGxqqIiAillFL79u1T3t7e6vbt2zbbdO7cWa1atUoppdSgQYPU+PHjGxzLkiVLlK+vr8rJybEu27hxo+rZs6cqLi5WSql699Q5Sh5ffvml+vjjj9XRo0fVjh071KBBg1T79u3VzZs3JQ8HzsOZcnH0PFxdXZWrq6uaPXu2Onz4sFq1apVyc3NTH374oeQheTR5HqL5af6YsHuNUgqdTgfA0aNHKSwsxM/Pz2ab4uJiLly4AEB6ejpTp05t0L7+/ve/s3DhQj799FMCAwMBuHTpEjNnzmTXrl24ubm12DwAxowZY/25d+/eDBw4kA4dOvDxxx8THx8vebTAPJwpF0fIw2KxEBUVxRtvvAFA3759OX78OCtXriQuLk7ykDw0y0M0DSnqmtmpU6fo2LEjAIWFhYSEhLB79+4q2/n4+ADg7u7eoP189NFHPPvss/zjH/+wucSalpZGbm4u/fr1sy4zm83s3buX9957j5KSEvR6vcPnUR0fHx+6du3K+fPn69y+5FE/TZ0HOE8ujpBHSEgIPXr0sFkWERHB5s2b69y+5FE/kofQkoypa0b//Oc/ycjIYOzYsQD069eP7OxsDAYDXbp0sXn5+/sDFb0Eqamp9drPxo0bmTx5Mhs3buSRRx6xWTdixAgyMjJIT0+3vqKiohg/fjzp6el1KugcIY/qFBYWcuHCBUJCQurUvuThWHk4Uy6OksdDDz3EmTNnbJadPXuWDh06SB6Sh2Z5iCak2YVfJxYXF6dGjx6tsrKy1OXLl1VaWppatGiRatWqlXr00UdVeXm5Ukopi8WiBg8erCIjI1VKSorKzMxUBw4cUHPmzFGHDh1SSin19ddfKxcXFzVv3jx18uRJdezYMbV48eIa971hwwZlMBjU8uXLVVZWlvWVn59f42dqG1PnyHn84Q9/ULt377buLzo6Wvn7+6vc3FzJw4HzcKZcHD2Pb7/9VhkMBrVo0SJ17tw5tWHDBuXh4aHWr18veUgeTZ6HaH5S1DWBuLg4BShAGQwGFRAQoKKjo1VycrIym8022968eVM9//zzKjQ0VBmNRhUWFqbGjx+vLl68aN1m8+bNqk+fPspkMil/f3/1+OOP17jvYcOGWfd95ysuLq7Wz9RU1DlyHrGxsSokJESZTCbVtm1bFRsbq86fPy95OHgezpSLo+ehlFKfffaZ6tmzp3J1dVXdu3dXf/vb3yQPyaNZ8hDNT6eUUnbo8BNCCCGEEBqSMXVCCCGEEE5AijohhBBCCCcgRZ0QQgghhBOQok4IIYQQwglIUSeEEEII4QSkqBNCCCGEcAJS1AkhhBBCOAEp6oQQQgghnIBB6wCEcEZms5mysjKtwxBCCKdhNBrr9Hzye5kUdULYkVKK7Oxs8vPztQ5FCCGcjo+PD8HBweh0Oq1DcUhS1AlhR5UFXWBgIB4eHnLiEUIIO1BKUVRURG5uLgAhISEaR+SYpKgTwk7MZrO1oPPz89M6HCGEcCru7u4A5ObmEhgYKJdiqyETJYSwk8oxdB4eHhpHIoQQzqny/CpjlqsnRZ0QdiaXXIUQomnI+bV2UtQJIYQQQjgBKeqEEEIIUW8XL16kVatWZGRkaB2K+IlMlBBCCCFEvYWGhpKenk779u21DkX8RIo6IYQQQtSbwWCgS5cuWoch7iCXX4UQQgghnIAUdUIIIYQQTkCKOiGEEELUS3p6Ok899RTBwcGYTCY6d+7Ma6+9Rnl5udah3dOkqBNCCAdw3333kZSUpHUYTiklJQWdTlfra+fOnVqH2WIkJyczYMAAgoKC+Pzzzzl16hRz584lKSmJ+Ph4rcO7p0lRJ4Rg0qRJ1l9uRqORjh078vLLL3P79m3rNjqdjq1bt2oXZAty5/E0mUx06dLlrr0Yhw4d4ve//30zRnnvGDp0KFlZWdaXn58fc+fOtVk2YsQIrcNsEXbv3s3UqVNZs2YN7777LlFRUXTu3JlJkyaxZMkS1q1bx/nz57UO854ls1+FEACMHj2aDz74gLKyMtLS0oiLi0On07FkyRKtQ2uRKo9nSUkJX375JdOnT8doNDJ79myb7UpLSzGZTAQEBGgUqfNzd3e3Pjf0ypUr5OXlMWTIEIKDgzWOrOWZOXMmY8aMYeLEiVXWDRs2DICjR4/KrFiNSE+dEAIAV1dXgoODCQsLIyYmhujoaHbt2qV1WC1W5fHs0KEDzz33HNHR0Wzbto1JkyYRExPDokWLCA0NpVu3bkDVy6/5+flMmzaNoKAg3Nzc6NmzJ59//rl1/f79+xkyZAju7u6EhYUxY8YMbt261dxptjhHjhwBoF+/fhpH0vIcOXKEY8eOMX369GrXFxcXAxW3OhHakCMvRFMqK4Lrp5t/v226g9GjwR8/fvw4Bw8epEOHDnYMyj5+/PFHCgsLadu2LQBZWVm4urrSpk0bysrKyMrKIjAwEDc3N27evMnNmzdp164dADk5ORgMBvz8/DCbzVy5coWAgADc3d0pKCggPz+fsLAwAHJzc3FxccHf398ucbu7u5OXlwdAamoq3t7eNRbNFouFMWPGUFBQwPr16+ncuTMnT55Er9cDcOHCBUaPHs2f/vQnkpOTuXr1KgkJCSQkJPDBBx/YJd66ul1q4WJ28w+Obx9swM1U/36Jw4cPExYWhp+fXxNE1QiWIijV4Fxh6g4udTtXpKenA9CnT59q1x8+fBiA3r17Nyqk6OhoVqxYQXh4eKPauRdJUSdEU7p+Gtb3b/79TkiDoPr1RHz++ee0atWK8vJySkpKcHFx4b333muiABtu3759HDx4kDfeeAOAtWvX0rVrV5588klu3LjB4sWLeeGFF+jatSvffPMNKSkpvPXWWwCsW7eOkJAQJkyYQGFhIYsXL+a5556jV69eHD58mM2bN7Ns2TIANm7ciJeXF1OmTGlUvEopUlNTSUlJ4fnnn+fq1at4enqyZs0aTCZTtZ/56quv+Pbbbzl16hRdu3YFoFOnTtb1b775JuPHj+eFF14AIDw8nGXLljFs2DBWrFiBm5tbo2Kuj4vZ5fzn4uxm21+llbOC6dq++uNXm8OHDztmL13pafheg3PFfWngVrfjUVpaClDj9+v9999n6NChdOzYsco6s9ls/U/J3Zw7d47OnTvXaVthS4o6IZpSm+4VBZYW+62n4cOHs2LFCm7dusVf/vIXDAYDY8eObYLgGmfIkCH07dvX+j4+Ph5XV1cAfH19mTVrFoGBgQAMHDiQiIgI67YTJ060Xhpq1aoVs2bNso5l69evn03h9PTTT+Pi0vARKpVFcllZGRaLhWeeeYYFCxYwffp0evXqVWNBBxU9Iu3atbMWdD939OhRjh07xoYNG6zLlFJYLBYyMzNtcm5q7YMNrJzV/GPT2gc37NfX4cOHefbZZ6ssX79+PcuWLaO4uJj27duzZcsW6/eqWZi6VxRYzc1U93NFZGQkAHv27CEmJsZm3VtvvcWpU6fYv3+/ddljjz1Gu3btOHToENOmTcPNza3aY3zixAmmTJlCcXExsbGxBAcHN+rf3r1MijohmpLRo949Zlrx9PS0Dm5OTk4mMjKStWvXOtwtClq3bk3r1q2t70NCQqw/G41Gm+dQent74+3tbX0fFBRk/Vmv19ts6+XlhZeXl/V9ZWHYUJVFsslkIjQ01GackaenZ62frRzUX5PCwkKmTZvGjBkzqqxr7udwuplcGtRjpoVr165x6dKlanvqxowZw4QJEwCYOnUqu3fvZtSoUc0XnItHnXvMtPLAAw8watQopk+fTllZGVFRUeTk5LBmzRo++ugjPvnkE5tLsxkZGYwYMYL3338fgLy8vCrH+OGHHyY2NpaNGzfSq1cvYmJiGn359l4mRZ0QogoXFxfmzJlDYmIizzzzzF2LDFHVnUVyffXu3ZvLly9z9uzZanvr+vXrx8mTJ2WGYT1Vjvn6eVGnlGL16tVs3ryZ0tJSLl26ZC0+hK0tW7awcOFCXnrpJS5fvozZbGb06NGcPXuW0NBQ63YFBQWYzWZmzpwJ1HyMt27dyrBhw+jVqxcAERERNu2I+pH+TSFEtcaNG4der2f58uXWZZmZmaSnp9u8ZMal/Q0bNoyhQ4cyduxYdu3aRWZmJtu3b2fHjh0AvPLKKxw8eJCEhATS09M5d+4cn376KQkJCRpH7tiOHDlCUFBQlaLhww8/5PTp0+zdu5ejR4/i6+tLjx49NIrSsXl4eLBkyRK+//57ysvLiY+P5+zZs3h42E62OHHiBA8++KD1fU3HOCMjw6Z3Ly0tTXrqGkGKOiFEtQwGAwkJCSxdutRauCUmJtK3b1+bV+UtIoR9bd68mfvvv5+nn36aHj168PLLL2M2m4GKnrw9e/Zw9uxZ6xjDefPmSQ/HXbzyyitkZ1ed1HHixAkeeugh3N3dWb58OUVFRXLfwDpavnw5U6ZMqXIeyMjIsPa+Qc3HuE2bNhw/fhyAnTt38tVXX0lR1wg6pZTSOgghnMHt27fJzMykY8eOzTr7UAjROEePHuWJJ57Az8+PwYMHc+zYMXlsWCPNmDGD6OhoHnvsMaDmY5ybm8uYMWOwWCz07NmT/fv3k5mZWWO7cp6tnRR1QtiJnGyEEKJpyXm2dnL5VQghhBDCCUhRJ4QQQgjhBKSoE0IIIYRwAlLUCSGEEEI4ASnqhLAzmXskhBBNQ86vtZOiTgg7MRqNABQVFWkciRBCOKfK82vl+VbYkseECWEner0eHx8fcnNzgYo7r+t0Oo2jEkKIlk8pRVFREbm5ufj4+KDX67UOySHJfeqEsCOlFNnZ2eTn52sdihBCOB0fHx+Cg4PlP8w1kKJOiCZgNpspKyvTOgwhhHAaRqNReujuQoo6IYQQQggnIBMlhBBCCCGcgBR1QgghhBBOQIo6IYQQQggnIEWdEEIIIYQTkKJOCCGEEMIJSFEnhBBCCOEEpKgTQgghhHAC/w8p11dJm/iUiwAAAABJRU5ErkJggg==",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "q(s,a) = \n",
+ "[[2.56 4.458]\n",
+ " [6.448 6.346]\n",
+ " [3.256 2.165]]\n"
+ ]
+ }
+ ],
+ "source": [
+ "model = Q_Learning_Agent(env, eps_min=0.01, eps_decay=0.001, alpha=0.1, gamma=0.9)\n",
+ "model.learn(total_episodes=10)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Qblijq9WVobQ"
+ },
+ "source": [
+ "Since our environment has been defined with one-dimensional state and action spaces, we can plot the q-function after training as follows:\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 21,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 408
+ },
+ "id": "z_u32nxmzSDm",
+ "outputId": "a8faf1dc-3bd4-4cf3-bc52-e838ee7fe16d"
+ },
+ "outputs": [
+ {
+ "data": {
+ "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAGHCAYAAACamdTSAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAADE9ElEQVR4nOy9d3wchZ33/5mZ7bvqzbJkyyq23KtkWzYmEPqBsSEHSY4jJCS5u4SQPEcuR8jlSPKkQXIF8qRduBzwS+BIwBgIGEyzsY1tMFhdlizL6pLVdlW275TfH6sZza52V1tmdmfleb9eJKAyO9oyn/m2z5fgOI6DioqKiopKnJCpPgEVFRUVlfRGFRIVFRUVlYRQhURFRUVFJSFUIVFRUVFRSQhVSFRUVFRUEkIVEhUVFRWVhFCFREVFRUUlIVQhUVFRUVFJCFVIVFRUVFQSQhUSFRUVFZWEUIVERUVFRSUhVCFRUVFRUUkIVUhUVFRUVBJCFRIVFRUVlYRQhURFRUVFJSFUIVFRUVFRSQhVSFRUVFRUEkIVEhUVFRWVhFCFREVFRUUlIVQhUVFRUVFJCFVIVFRUVFQSQhUSFRUVFZWEUIVERUVFRSUhVCFRUVFRUUkIVUhUVFRUVBJCFRIVFRUVlYRQhURFRUVFJSFUIVFRUVFRSQhVSFRUVFRUEkIVEhUVFRWVhFCFREVFRUUlIVQhUVFRUVFJCFVIVFRUVFQSQhUSFRUVFZWEUIVERUVFRSUhVCFRUVFRUUkIVUhUkg7HceA4LtWnoaKiIhGaVJ+AyuUFx3Hw+XxwuVwgSRJarRYURYGiKJCkel+jopKOqEKikjQYhoHP5wPLssI/Pp8PBEGAIAhoNBpoNBpVWFRU0gyCU3MMKjLDcRxomgZN0+A4DgRBwOv1CkLBp7pYlhVSXiRJqsKiopImqEKiIissy4KmaTAMAwAgCEJIbxEEEfJ3xKLCC48asaioKBdVSFRkgRcDn88XIAaAX1y8Xm/A16I5lvitGiwsGo0mqmOpqKhIjyokKpLDcRy8Xi8YhhHEQnyRj1VIQh0/OGIhSRIURQUU71VhUVFJDqqQqEgKH4WcOnUKZWVlWLp0acifSURIggknLMGpMFVYVFTkQe3aUpEEjuPAMAxomgbLskm9aBMEAYqihPMA5sTK4/GowqKiIjOqkKgkDF885wvqJEmCJMmUDB3y4qAKi4pK8lCFRCUh+As0y7IgSVK4GPPdWakmkrB4PB54vV4AoduNVWFRUYkOVUhU4oJPZfFdWWIRASILSSov0GJhoShKqKtwHDdPWPjCvUajmff3qaiozKEKiUrMBM+GhLrILhSRKOWiLC74BwuL2+0WfoYXFj5iUYVFRWUOVUhUoibSbEgwSkltxYoqLCoqsaMKiUpUiG1OAEQUEf774YRkcnISFy9ehMViQW5uLjIyMhQ7pb6QsIyOjoIgCBQXFwfUWFRhUbmcUIVEZUH4KIRlWQCI6qIfSkg4jkN3dze6urpQXFwMu92OgYEBsCyL7Oxs5OTkICcnBxaLRbEX4WBhcTgcAPzPER+x8F1rqrCoXC6oQqISluDZkFgvhmIh8Xg8aG5uhsPhQE1NDUwmk/A9u92OyclJ2Gw2dHd3gyAIQVSys7NhNpsVfRHm51jEEQvDMGAYJmy7sSosKosJVUhUQhJqNiSWC594jmRiYgJNTU3IycnBrl27oNFohO4ogiCQkZGBjIwMLFu2DCzLwm63w2azYXx8HBcuXIBGowmIWIxGo2IvwnzEEuxsHCwsfDeY2CdMqX+TispCqBYpKvPgU1kMw8R959zU1ASj0QgA6OnpQXV1NZYtWyakvMRCstC5TE9Pw2azwWazYWpqCjqdThCVnJwcGAyG2P9IiTh//jxIkkRVVVVUPx9smc8LCC8ovLiowqKSTqhCoiKw0GxILDQ2NsJms4GiKGzevBkZGRkBjxOtkATDMAympqYEYZmZmYHBYAgQFp1OF9c5x0OsQhJMJGFRt0eqpAtqaksFQOKpLDFjY2MYGRmByWTCzp07odFI9zajKAq5ubnIzc0FANA0LdRXent70draCrPZHFBj0Wq1kj2+1IRLhdE0rW6PVEkbVCFRCViBm4iAsCyL8+fPo7+/Hzk5OcjMzJRUREKh0WiQn5+P/Px8AIDP54PNZhNajB0OBzIyMgRRyc7Olv2cEiGcsPh8voh2LqqwqKQS5X6iVGQneDYkERFxOp1obGwEy7Koq6tDX1/fgpPtcmRVtVotCgsLUVhYCMDfLcYLS2dnJ9xutyAsOTk5yMrKEny44kHuzPBCwqJGLCpKQBWSy5Tg2ZBEiruXLl1CS0sLli5diurqaqFYrITym16vx5IlS7BkyRIAgMvlElJh586dg9frRVZWliAsmZmZir4IhxIW/rX0+XzCz6jbI1WSiSoklxniC0+iqSyGYdDe3o7h4WGsX79euFgD/osZL1JKwmg0wmg0ori4GBzHweVyCYX7gYEBMAwT0GqckZGx4POTahNKcUQlfn35iIXfHhncFaaiIhWqkFxGSFlQt9vtaGxsBEmS2LVrV8CAIZAeXlsEQcBkMsFkMqGkpAQcx8HhcAjC0tvbCwABwpIuw5E8kYRFXUusIhWqkFwmSDEbwjM4OIi2tjYsX74cK1euDJkKSgchCYYgCFgsFlgsFixbtgwcx2FmZgY2mw1WqxUXL14ESZIBHWFK/xujFRZ1F4tKIqhCsshJ1OZEDE3TaGtrw9jYGDZv3oyCgoKwP5uOQhIMQRDIzMxEZmYmysrKwLIsZmZmYLVaMTo6is7OTgBz6TJ+6l7JqGuJVeRAHUhcxASnshIpqM/MzKChoQF6vR4bN25ccJr8woULcDqd2LhxY9jz4gfw0hWGYdDS0iJ0vU1PT0Ov1wcMR+r1+hSfZfSIhYXvDlOFRSUa1IhkkSLVbAjHcejv70dHRwfKy8tRWVkZ1bEWQ0SyEBRFQa/Xw2KxoLKyEjRNC1P3/f39aGtrg8lkChAWpQ9HAoFricNtj1SFRUWMKiSLDPFsSKI2Jz6fDy0tLZicnMTWrVuRl5cX9e9eDkLCwz+/Go0GeXl5wvPk8/kCXI1bWlpgsVgCaizpMBwJzN/F4vF4AlJh6pKvyxvlvotVYiaaFbjRMjk5icbGRpjNZuzevTtm/6rLSUjCodVqUVBQINSSvF6v0BF24cIFuFwuSYcj5UbdHqkSDlVIFgGxrMCN5lg9PT3o7OzEypUrsWLFiriOFc3vpHuNBIhtsl2n06GoqAhFRUUAALfbLQgLPxyZmZkZICzpMBwJRBaW4BkWVVgWH6qQpDlSFtS9Xi+am5sxMzOD7du3Izs7O+7zUupAopIwGAwoLi4WhiPFwjI0NASapgOm7pW8khgILyzq9sjFjyokaYyUsyFWqxWNjY3Izs7G7t27Ey4KqxeG2CAIQmgjXrp0KTiOg9PpFISF9y5Ll5XEQGRh8Xg8cLvdqrAsElQhSUOknA3hOA5dXV3o7u4OWD6VKGpEkhgEQcBsNsNsNqO0tBQcxwmbI0OtJM7JyYHJZFL0BTg4Wo60lpifule3R6YHqpCkGVLanLjdbjQ1NcHtdmPHjh3IzMyU7DwXOiex43C6k4yLnHgl8fLly4XhSJvNhrGxMWElsbgjTMkriYHIa4lpmla3R6YRqpCkEfwEcqJRCOBfPtXc3Iz8/Hxs3bpV8jbUcBEJP5fS3t4ecEedm5ur+DtqJUGSJLKyspCVlYUVK1aAZVlhhmV4eBgdHR2KWkkcDdEu+VK3RyoPVUjSAClX4LIsi87OTvT19WHt2rUoKSmR+Gz9hDo/mqbR2tqKiYkJbNy4ERRFYXJyEuPj4+jq6gq4o87NzVX8hQ+Qfx9JtIg9wIDAlcSDg4Nob29P6UrieFCFJX1QhUThSJnK4pdPMQyDuro6WCwWKU81gOCIRGyxsnv3buH7WVlZgo8Vf+EbGhpCR0cHDAYDcnNz02IqXGkstpXEwMLCMj09DZ/Ph+LiYnXJV5JRhUShSDkbAgAjIyNobm5GcXExVq9eLfvgm/hcebfgFStWoKqqCgRBCEuYeMR31BUVFcKFz2q1BkyF88KSnZ2t6OE9pRG8ktjr9QrC0tXVBafTOW84UslT98B8YZmZmYHdbkd+fr66PTLJKPudcpkSvAI3ERFhGAYdHR0YGhqat3xKTviIo7m5GWNjY9iyZYtwEYuG4AsfvzLXZrOho6MDHo9HmLHIzc1V/IyF0tDpdCFXEoufX/FwZGZmpuKFm+M4oTDP/7e6PTI5qEKiMIJX4CZycXQ4HGhoaAi7fEpOPB4PHA4HNBoNdu3alXC9Q7wyN9RmQ5ZlhRmL3NxcxS+gUhqhVhIHD0fywqJU4eabUHii2cUiFhZ1e2T8qEKiEKRcgQsAQ0NDaG1txbJly7Bq1aqkfuiHhobQ3t4OiqJQW1sr+WOH2mzIz1jwC6goihIuenLvCVmMF57g4chQwh08dZ/q5yFYSIJRt0fKhyokCkDKgjpN0zh37hxGR0exadMmIXWRDPgd7pcuXUJFRQWGh4fDfrCl/HCGmrGYnp6G1WoVWmH1en1A4V7pHUtKIpRwK3ElMcuyMdV11O2R0qEKSYqR0uaE74zS6XTYvXt3Uttng9NoTqcTQ0NDSXt8MSRJIjs7W/AKC9WxJJWdu1Laf5MJQYRfSTwxMYGuri4hIuT/ScZw5EIRyUKo2yPjRxWSFMHPhvT392NkZASbN29OyOZkYGAA7e3tWLFiBSorK5Oayrp06RJaWlpQUlKC6upqkCQJl8sV8SKbzAtwqI4l/m76/Pnz8wrLSnfdVRoEMX8l8fT0NGw2G0ZGRnD+/PmA4Uh+6l5qEhUSMaGWfPGPoQrLfFQhSQHiVBb/xkxk+VRraytsNlvMy6cShWVZtLe3Y2hoCBs2bBDs0QFl7yMJtnPn8/9WqxWDg4NgGAbZ2dlCKkzp5ohKQxwRlpeXhxyOlGMlsZRCEoy6PTIyqpAkmeAVuBqNJm5zw6mpKTQ0NMBkMmHXrl1J3Q/ODzdyHBeyI0zJQhJMcGGZz//zMyziGZdkpWkWE6GGI+VYSSynkAQT6/ZIvtV4sTobq0KSJIJnQ/g3FEmSMQsJx3Ho7e3F+fPnUVVVhfLy8qS+OUdHRwOGG0N9eNNJSMQE5/95c0Sr1RqQpuGjFYZhFuWFQU7kWkmcTCEJ5nLfHqkKSRIIng0Rv+liFRLx8qna2lrBWykZsCyL8+fPY2BgAOvWrUNxcXHYn01XIQlGbI7Ip2n4i15/fz9mZmYwNTUFj8eD3Nxcxe9hVyKRVhJ3dnbC7XZHtZI4lUISzOW2PVJ9x8tINLMhJEkKbb8LYbVa0dTUhMzMTOzatSupLawulyvAp8tsNkf8+cUiJMFQFBVwN93U1CRcKIIverm5uWrhPg6iWUksnmHJzMwUbsiU+lxHsz0ynYVFFRKZiHY2hCTJBS+4HMfh4sWLuHjxIlatWoXly5djyD6EEp08zr3BjI2NoampCUVFRVizZk1UVhkLCUk6fDiigSRJWCwWlJWVAZi76FmtVrS2tgrrcvlUmBIG99KN4JXEwcORDMMgKysLbrcbbrdb0YLCE+v2yJ6eHhQVFSW1mSYWVCGRgVhmQxZKbXk8HjQ1NcHlcmH79u3IysoCANz10l3IM+Zh+9LtqF1ai5riGmTqpVtMxf8dFy5cQG9vb8yW84s1IgmF+PUNvug5nU5YrVbYbDb09PSk3VZDpRFuOHJychJTU1PCDVc6rSQGFt4e+cUvfhH/8A//gHvuuSeFZxkeVUgkJJ4VuJFSW+Pj42hqakJeXh62bNkSkHunCArD9mG8fP5lvHz+ZZAEiVW5q7CjZAdql9ZiQ8EGaKn4bcHdbjcaGxvh8/nispy/nIQkHAQxty5XXLgXbzXUarUBO1iS2Xm3GBA3R/T29mLdunWgKCpgJTHfjpxO4h3sbOxwOGRd+5AoqpBIRLw2J6EiEnEksGbNGpSUlMw7lo7SwUW75n6HY9E+0Y72iXY83fQ0DBoDNhdtFiKWypzKqP+WiYkJNDY2Ij8/H9u2bYureHy5pLZiIXiroXi+YmBgAOfOnRPaYPnCvdJ3hCgJlmVBUVRMK4nl9mGTAj6yVYVkkRM8GxLLRZK/4+B/ly9q0zQdMRJYKNpw026cHjyN04OnAQB5xjzUFNcIEUuecX6uleM4dHV1obu7O6yARQv/e/wuFZX5iOcrKisrA9pgxTtC+PpKuG4lFT+haiOhxJufuhf7sEk9HCk1drtdFZLFing2JN4VuGIhGRsbQ0tLS1RFbS0Z253qhGsChy8exuGLhwEAlTmVqC2uRe3SWmxZsgUEQwi1mJ07dyIjIyOm4wdzuQiJlOm74DZYj8cj1FfOnTsHn883z3FX6UXlZBJNkV3sAQZgXjt38HBkdna2Igw+1dTWIoVlWdA0nbBjL//G511zF5rP4IlVSILpsnWhy9aF59qeA8ESKEIRthZuxb6afTBbIrf2RoNYSELBFxNVwqPX6+cV7vncf19fHwAE7GBJh9y/XLAsK9zMxUJwOzcfFU5OTqKnpwd2uz3lK4n5hgJVSBYRUq/A5XvIJycno5rP4EmkkC6Gt3PwGDy4NH4Jh944hAxdBmqX1goRy9KMpTEfN5KQ8Fb3LpcLeXl5KbMdl4pknLe4cF9aWhrguDs+Po6uri4h98+nwpLp/pxq+PdZohFaqOHISCuJk7HymTdAVYVkkRBcUE9URIaGhtDW1gYA2Lx5c9QiAiQekfB3uCzLwmw2B3wYZrwzeLfnXbzb8y4AoDSj1C8ss23GFt3Cb+hwQmK321FfXy+4wfK24xqNRrgAqt1LCxPKcTfYGNFgMATsYFnMhXspNoqGQgkriR0OBwCoQrIYkHJvCMMwOHfuHEZGRrBx40Y0NjbGfAwNGf9LR9M0XC4XKIqKqsd+YGYAAx0DONhxECRBYk3+Gn83WHEt1heuD3kuoYRkeHgYLS0tWL58uWA3smLFCuEiaLVahe4lPp2g2o5Eh9hYsqKiQtjBwhtPtrS0JP1OOpmI7YfkJJqVxFLXsex2OyiKUnSEqX46FyCe2ZBIzMzMoLGxUdhlbjQa4zJu1FHxFQD5VJZer4/rrp/lWLSOtaJ1rBVPNj4Jo8aIrUu2ChFLeXY5gEAhYVkWHR0dGBwcFLY2BteX+A8e373ET4fztiOZmZlCh5NaZF6Y4B0s4jvp9vb2AJuRxfCcsiwbMHeRLIKdo8V1rP7+frAsm/BwpNPphNlsVvTrowpJBKRcgStePlVWVoaqqirhjRGPkMSa2uKtJRiGgclsgoaS5qV30S68P/A+3h94HwBQYCoQ6isz9Azcbjfq6+vBMExIu/lQaLXagHSCy+USupf6+/sBQNgXkpubm1Jb93RpGBDfSYfbwS7ewZJuNSsl2KKEqmPZ7XbYbDZMTk6iu7sbBEHEvJKYL/grGVVIwsAvnJIiCqFpGi0tLbBardiyZYtwl8gTl5DEUGxnGAZOp1PwhZLzAjHmHMOhC4dw6MIhTE9N47+H/hs1S2qwd8teaPTxvd2MRiNKSkoESwze1n1sbAydnZ2CrTt/EVRCu6aSCWUzwl/wglfl8s+p0of2lCAkwRAEMW840m63w2q1xrSS2OFwqEKSbvCpLL4rK1ERmZqaQmNjI4xGI3bv3h0ynRSLAzBPtBGJx+uBxx1/KitePB4POHAY58ZxxHoER945Ah2lw8bCjdi+dDu2LdmGMnNZzMcVF5n5AbNQ+9h5YVGH+BYm1AVvenoaVqs1YGhPXLhXmlgrUUiCIUlSeO8CWHAlMd95xwuJnDeAg4ODePDBB/H666/D6XSiqqoKTz75JGpqaqL6fVVIREidyurt7UVnZycqKipQUVER9ljROAAHs1BEEpDKMpmSVqwWPy4AaLRzj+tlvPho+CN8NPwROI5Dli4L24q3oba4FjVLalBoLoz58YLnALxeb8ghPjndd9MpBRQN4lW5AITCfbBYx7p4Sk7SQUiCWWgl8ZkzZ/Dggw9i1apV4DgOIyMjAeuspcJms2H37t24+uqr8frrr6OgoACdnZ0x7TpShQTSz4Z4vV60tLRgenoaNTU1C74gUtdIkpnKivS4MzMzQAR9nPRM4t3ed/Fur7/NeFnGMtQU16C2uBZbirbApF24nhKMTqcLqAXwxU+r1Yqenp6Awj5fX1GJTHDhXrx46vz58wEtsLm5ucJ+kGSSjkISTPBKYrvdjh/84Ad44YUX0Nvbi6VLl2L16tX45Cc/ie985ztRDS5Hw6OPPoply5bhySefFL5WXl4e0zEueyEJXoGbqIjYbDY0NjbGtHwqntRWuPZfr9cLt9ud9FSWFI/bP9OP/pl+HDzvbzNem7dWEJY1eWtAkbGlqIKLn+K1uZcuXcL58+cvq1kLqQhePMUX7q1WKwYHB8GybEAUmIybmcUgJMFYLBbceeedGBoagtFoxBNPPIH33nsP7777rqStwK+88gpuuOEG3HHHHXjvvfdQUlKCr371q/jyl78c9TEuayEJXoGbyBtRvHxq5cqVKCsri/rDI0X7bypTWW63Gz6fT9LHZTkWLeMtaBlvwVPNT8GkMWHLki2oWeIXlmWZy2I+ZvDa3FCzFuI762i2G6ZL15acBLfAOhwOQVh4G/eFCsqJshiFhId3/s3Ly8Ptt9+O22+/XdLjX7x4Eb/5zW/wwAMP4Dvf+Q7OnDmDr3/969DpdFHvP7kshSSaFbixEG75VLQkmtoSp5SS2W/OsiycTicA/92TnI/rpJ0BbcZF5iLULKnBtiXbUFNcgyx9bM85EHrWwmq1Bmw3FKfB0q0lNhWI94OId7BYrVahoCx2283NzZWkcL+YhUTuri2WZVFTU4Of/OQnAIAtW7agpaUFv/3tb1UhCYeUBXVgbvlUbm7uvOVT0ZJI+y+fUtLpdEmdfPX5fHC5XNBqtTAYDCGfQ4IgwEUqkiTAiGMEr3W9hte6XgMArMxZ6S/aF9dgQ8GGuAY2g00SHQ6HULi/ePFigJeVauMSHcFRYCi3XSlcDBazkNjtdllX7BYXF2Pt2rUBX1uzZg0OHDgQ9TEuKyGR0uZEvHxq9erVKC0tjft48QgJBQoul0vylFI0uN1ueL1eGI3GhWsKScr8dNo60WnrxLNtz0JH6bCpcBNqltSgprgGldmVMb824jtrviWWt3EZHBwUllCxLAuDwQCaplPeuZQOhHLb5Qv3vItB8A6WaARiMQuJw+FAWVnsrfLRsnv3bnR0dAR87fz58zE95mXxzpd6NkS8hlaK3R0URcUkJHa7HX3dfWBZVvaUkhiWZeFyuUIaPSoJL+PFmeEzODN8BqgHsg3ZgqjULqlFvil/4YMEIc7zA3MXwM7OToyOjmJoaEi1cYmDYBcDt9stRIFi76qF2rcXs5A4nc6oHCHi5R//8R+xa9cu/OQnP8Gdd96JDz/8EL/73e/wu9/9LupjLHohkTqVNTo6iubm5qiWT0VLLBHJ0NAQWltbkZOVA9N08vZPxGr0qCQm3ZN4u+dtvN3zNgDg02s+ja9u/WpCx+QvgMPDw8jPz0dubq5QYOYtR8ST4ZfzrpBYMBgMWLp0aYB3FS8sPT09IAhiXvs2QRCLWkjk3kVSW1uLgwcP4qGHHsL//b//F+Xl5Xjsscdw1113RX2MRS0kiazADYY3HhwYGMC6deuwdGnsOzrCwX8QIsEwjLD8atOmTZi2ToMYSM6FKVGjR6Ux7ZmW7Fh811Zw5xJvhcHvCddqtQE2+UqbDFci4vZtceFevH9dq9UiJycHPp9v0bZuJ2Op1S233IJbbrkl7t9flEISPBuSqIg4nU40NDQAAHbt2iV5BwVFUfD5fGG/73A40NDQAJIkBcdg7aT8HxopWorlKrYngtznJLYcKSsrEyaWrVYr+vr60NbWJti4LEZLd7kItX+dnwS3Wq3wer2Ynp4OeF4Xg7gofTsisAiFJHg2JNEBw+HhYbS2tgpTpXKEz5FSW5cuXUJLSwtKSkpQXV0tPH68NvLRIkVLMUEQSSu2x0KyZz+CJ5b5yXCr1Rpg6S6ur6hpsIURP680TQtpL6vVKmwzFC+dSkffNT69pwpJkpB6NkS8fGrDhg2yeNzwhBISlmXR3t6OoaGhkI8f65R3LPCtvcluKU5XYn2fiSfD+aiPrwPwu9jFbcaqjcvC8N1z4jW5/FxQsO+aeAdLOgi2GpEkCakL6na7HQ0NDQHLp+QkWEicTicaGxvBcVzYHR6JrtoNhXhKParW3jRFSek2saV7sI2LeIBPbJO/WF+XRAhVbA+eCxIvneIFm98Nkpubq9iGCFVIkoCUsyEcxwkzAsHLp+RELCR8V1hxcTGqq6vDhuJSp7aSOaWeaqRMbUmdJgtn42Kz2QJW5opt8hfzaxUtC3VthVo6xRfux8fH0dXVFTBwylu4pxr+c6kKiUxIvQKXpmm0trZiYmIi5PIpOSFJEjRNo729XegKW8jZU8qIhKZpOJ3OiFPq8RLu7t/r84L20aAoChqNJuHXb7ESzsbFZrMJNi7izYbp1potFbG2/4r32pSVlc0bOG1vb1eEoafT6QTHcaqQyIHUqazp6Wk0NDTAYDBg165dSb8ToWka09PToGkadXV1UXWFSSUkfGuvwWCQvCU11GvC1wRomoZOpwPDMPB6vQD8F01KQ0FDadLuLvvBIw9iS9EW1BbXojKnUrbHCWXjEmyQKG4zVsJddTJIdI4keOA0lKFnRkZGwA6WZBTuHQ4HAKhCIjVSzoZwHIe+vj6cP39+weVTcsGviyVJEjt37oz6zRnLqt1Q8Dlj2afURQFJcPoMmPXjmo0uGYaBz+uDm3GDJEloNBq/uFBUUqKkRPhg6AOcHjqN39T/BjmGnLlJ+uJa5Bnl8UkKZZDIbzYcGhpCR0cHjEZjwF31YrVxkXogMVQkyNdXxJ124sK9HDc/DocDWq1W8fNbafOuEs+GSGFz4vP50NLSgsnJSWzbtk1ozUwWYq+u0tJSTExMxHQxT6RGkorFV8HpM2CuvkAQhCAaer0eLMeCof1pS7fbDY7jQFGUZGkwqYUk+Fxsbhve6nkLb/W8BQAozyoXRGVT4SYYNPJECeKNexUVFfD5fMJddVdXF1wuV8oXUMmF3JPter0+YGEav4PFZrMJTgbiFKNUTtEOh0OxTQBi0kJIWJYFTdOSpbL45VMZGRnYvXt30qeMxV5ddXV1cLvdGBsbi+kY8bb/JnvxFQcuZPosXJGaGB4GaTaDNJmgNRr9bd0iYVFqGiySOHVPdaN7qhvPtz8PDanB+vz1glPxqtxVIAl5zl+r1Qa0w/I+VuIFVGK7kXS4YIUjmRYp4k67kpISwcnAZrNhYmIioHAv3sESD3a7XVYLealQtJBIvQKX4zh0d3ejq6sr5uVTUjExMYHGxkbk5+dj27Zt0Gg08Hq9sS+2ImMTP7kWUC2Ez+t/7aJNnxHT0yCmpwAAnN4AwmwGYTKBMhmh0+kCmiz4NJg4WokqDSZhQBJr1xbN0mgYbUDDaAOeaHwCmbpMbF2yVVjYtcSyRLqTCyLYx4q3ceG7lni7Eb4jLJ1sXFLptSV2MuCdovkU4/DwMDo6OmAwGAKEJdrnNh1afwEFC0lwQT1REfF4PGhubobD4Yhr+VSicByHrq4udHd3Y82aNSgpKRH+nljdf4HYaiSpaO3lax58Hj+e147wuAGPG4R1AhxBAiYTYDJBYzZDM5seY1lWEJaANJgoWgl+bCXNkUx7p3G07yiO9h0FAJRmlAr1lS1FW2DRyXMRiWTjIt4TwouK0m1clGTaKE4xAgho4e7t7UVrayssFkvAJs5wN3ZOpzMtFqopUkhYlsXExARGR0dRWRn7LolgJiYm0NTUhJycHOzatSvpbXziDYqhbOcT3ZAYiWgWUEkN/5gEQUCn08X2mARCRgwExwIOOwiHHRgDOI0WMJn8aTCzeS4NJhIWr2cuDcZHK0q52IRjYGYAAzMDeKnzJZAEiTV5a4Q02Nq8tbI5GoSzcbHZbOjo6IDH41G0jYuShCSY4MJ9qOc2XO1K7tTW97//ffzgBz8I+Fp1dTXa29tjOo6ihESctnA6nbh06RKqqqriPh7Lsujq6kJPTw+qq6uxbNmypL/5rVYrGhsbkZOTE3aDohxCwnH+2kTUC6gkQrz0KpIRZTg4ECCiiBgI2gdMT82mwQhwBgMIkwmE2QzKOD8N5vV6wTAMKIqC1eafw5BqmE8u7y6WY9E63orW8VZhb/3mos2oLa6Ne299tISyceHbjINtXOTawx4LShaSYMTPLYCA55avXen1erz55pugKErWXSQAsG7dOrz99tvCf8eT9laMkASnsrRarfDv8cAXtL1eryTLp2KF4zhcvHgRFy9eXFDEeCHh60DRECm1lYoFVKHaieMRkjgfHYTbBbhd/jQYSQJGE2A2h0yDMTQTsJOdv8uOt9icrAuok3bi5OBJnBw8CQAoNBWiprhGSIXFs7c+GkIVl5Vm45JOQhJM8AoCh8OB1tZWvP/++6ivrwfHcfj0pz+Na665Btdcc43kYwoajQZLlsRfm2NZVhlCwrKsUHDmc9oajSZuIeFtRgoLC4WCdjLxer1oamqKuh7DfwBiEZJw7b+pWEDFtxNL8pgSOAYTrCgNBoDTagGTWegGy8vPw+7du4Wd7HynDb8zhL8YRlsQTVXNZdQ5ikNdh3Co6xAIEFiZu1IQlXj31keDeCqct3MPNbwnXpcr580MfxOWrkIihq8p7tixA4cPH8YPfvADtLa2Yt26dXjmmWfw9a9/HYODg5LucO/s7MTSpUthMBhQV1eHn/70p1i+fPmCv8dfr0iSTK2QRFqBS1FUzELCsizOnz+P/v5+yZdPRQvfWpyVlRV1PYb/AMRyV0USJEiCBMvNpcSEBVQGPfS65AwwRWonJghiwYtsLOIZL4TPB0xNgpiaBECA7KShu7ACVG0tLGvXYvny5QHFZr4gGo2nVbIt6cPBgcN563mct57Hs23PQk/p/XvrZyMWOaftg/ew88N7VqsVbW1tAetyc3NzJb/B4V+DxSAkwbjdbpSXl+Phhx/Gww8/DJfLJamJ7I4dO/DUU0+huroaw8PD+MEPfoA9e/YINwORIAgCJ06cwDPPPJM6IVnI5oQXkmgvNLxjLsuysiyfWgiO49DT04MLFy7E3FosFpJY0JAaeBmvJAuoYiXqduJYr7OyB1AciPFxaI89Be1TT4HLyAC7ZQuY2lrkbd+O3NmaHO9pZbVa0draCoZhkJ2djby8vIAVr4ByxESMh/Hgw+EP8eHwhwAgTNvzhXu5pu2B+cN7/Lpcq9WKnp4ewY6EF5ZEbVz4z81iFBKHwyHMAQGQ3In8pptuEv5948aN2LFjB8rKyvDnP/8ZX/ziF8P+3p/+9Cf87Gc/Q39/PyYnJ1MjJBzHCamscG29/IWJYZgFL4z88qelS5dGdMyVC5/Ph+bmZkxPT6O2tlZo+4uWeIVES2rh8roSXkAVK/K2E8ufihNf9omZGVDHjoE6dgwAwJaWgq2tBVVbC/2WLYKnVfDqXJ1Oh9zc3ICbISWTqmn7UOtyg2csjEZjQOE+1huhxS4kFRUVSXu87OxsrFq1ChcuXAj7M9u3b0dTU5MwHAykqNjO59UiRRu8GEQSEn6P+fDwMNavX59QwShepqam0NDQAIvFgl27dsU1xMWLaawXJI72F+aSuYBKTqdgJUAODIAcGIDm4EGAosCsXw+mthZZtbXIWLVKmLngawJutxsd5zvgcDiEFmOpvcHkIHjafkPBBqG+Iue0fagZC74VlrdxEacUo7FxiXRDmu44nU7Zu7bE2O12dHV14e677w77M7zXGACYzWbceOONqUttLfTC86mucBdXu92OxsZGYY95Mp9sINDwsaqqCitWrEjojRxLCzC/PdHtdCe1tVdOp2CBJFwLok5EMQyoxkZQjY3Af/83uKwsMNu2gamtRX5tLfJWrvTvCC/IhbZfC4aeczIWJu01lKzbLKWAZmnUj9SjfqQ+YNq+trgWNUtqZJ2212g0IW1cbDYbmpubo/KwSueOrYWw2+2yTrb/0z/9E/bu3YuysjIMDQ3he9/7HiiKwmc/+9mwv+NwOHDDDTfgvvvuw8aNG2GxWJTRtRUOjUYDmqbnfX1wcBBtbW1Yvnw5Vq5cmfQ3kc/nQ2trK2w2G2pqagTr6USIVkjcbrfQEliYV4hLzksJP/ZCxFuD4Z19Y0O5d5XE1BQ0774LzbvvAgDYsjKUFheD/MQV0FEaQKsThiJpmobP54PbPedkzE/cy3W3LxWRpu23LtkKs1a++mMoG5dgDyuxTb5er1/UQuJwOGQdXRgYGMBnP/tZTExMoKCgAFdccQVOnz4dUJcJpqWlBRUVFQHNNSmNSBYiuHOLpmm0tbVhfHwcmzdvjvjHygW/u8RoNEpq+BiNkIyPj6OxsRFFRUVYs2YNDCPyp7P4egif65bqAxufyEgDJ5FWkb29KGxrA068B/LaHnBGv4ULZTaDmu1i4zgONEODoRl4PB6wLnZetKL0lEyqpu2DPawYhhHqKwMDAzh37pxQf+HdwRebTT5vkSIXzz33XMy/s2bNmnlfU/SzLhaS6elpNDY2Qq/Xp2T5FMdxGBgYQHt7uyy7SyIJiXi4cc2aNSgtLQUAaCh5Xz4f7YPLmVx7FSThMWSZ++A4EE4HCKcDGB8DR2n83mBmM7RmM7SioUjeydrrDEyD8Rb5SibUtP2WJVuEjjA5p+0pihJMDysrK+Hz+WCz2TA8PAyGYXD8+PGk7AhJFvxwomramCD8dHRfXx86OjpQXl4uifdWrPBreK1WK7Zu3SrpMBBPOONGn8+HpqYm2O127NixA5mZmcL3pFy3G4zb44bX44XBaIBOmz4usKkgVIRDMDQwMw1iZtr/Mzo9MOsLpjMaAd3CaTCNRqP4aMVJO/H+wPt4f+B9AHPT9rXFtdi2ZJts0/aA3/2isLAQFEXB7XZj48aNQptxf38/AATY5KfaxiUeVCGRAJIk0d3dDY/Hk5LlUwAwMzODhoYGIRKSa4dHqIiE7wjLyMgIOdwoh5BIujkxjAFj5N9Jbvtvso5HeD2A1wPCZgUIEpzRCJjNEdNgvEOBOFpR+oUw0rT9xoKNCW/2DAVfIzEajSgpKZln4zI6OorOzk7o9fqANuN0sMlXhWQBFvpATE5OYnp6GiaTKSXLpwAIedgVK1agqqpK1g8xQRCCkIjTaJWVlSgvLw/52FJbYEi9OZEAETB5rzILx86lwcbgT4OZzf40mMk0Lw02b6GXaFOkkknWtH2oYnskGxexlbvYuUBpNvkMw8Dlcil+sdWpU6eUF5GIJ8TNZjOKi4uTLiIMw6CtrQ2jo6PYsmWLYP8sJ3xqi3/ssbGxBdNoGlK6l4+3fk/WTArfSizezU6SZFJ6tqQqtgvHS/D3CYYWORn7F3ph1hdMJ17oxfoNJ+elwWZ3ryg9Wgmets815GLbkm0JT9tH07UVbOPi9XqFNNi5c+fg8/lktXGJB4fDAQBJN5yNhZGRETz++OPKEhKx2WFtbS0GBgaSPjVst9vR0NAArVaL3bt3J62oT5Ik3G43Tp8+DYqiomookCIiSbbdvLiV2GA0gGM5IRICADPLgmI5gCCSkeWSBKmFKeRCL7MJGpMZGnEabLZo73F74GLTLw1mdVtDTttfuexKbCzcGPVx4mn/1el082xceH+wYBuXRFblJgIvJEpMbfHPeWdnJ959913lpLb4vR3Z2dlCPYDvxkgWQ0NDaG1tTcl8is/nQ1dXF5YtW4bq6uqoHjvRiIT/AHEcB7PFLPvgXHArMQCAQsAdNwgCLMcBs63B4sFVhV8XZSFgoRfmFnoJ3WCzwp/uaTBgbtq+d6oXP//kz6P+vUTnSMQ2LqWlpWBZVqiviFflirdFJmMI2Ol0Qq/XJ92SPxaGhoYwPj6e+ohEvII2eG9HsnZa8FYrly5dwqZNm1BYWCj7Y/KwLIvOzk5MT09jyZIlIXu0w5FIsV1s/R7vHo5YCGWtIm4uIAjCn56hKBAU6dcRzt+oy/8c4f/BWXGJ/1yUZ7EYPcELvdiVKwGSBEmS0Ol0C6bBot5rn0JiLchLPZBIkiSysrKQlZWF8vJyYVWu1WqdZ+PC2+TLIdR2uz0pn8144GfArFYrgBR3bbndbjQ1NcHtds9rbQX8QuJyuWQ9B4fDgYaGBsFqJZkhrMfjEZZv5efnxxzCxtsBE8n6XUr4oUOP1wOPOzZrFYLw/w8BAJx/ZyI3G6mw4mhl9odj+axJnYpKHVzIMI0XZQ2l8U9+cywYem6vPcuy86IVJV2sdGRsKVu5J9uDV+W63W4hDTY0NCQ4Q/OpMKl2rMu9ZlcKhOg3lSdw8uRJ5OfnY+vWrSEnUuPZSRILvGtwSUlJ1OkkqbDZbGhoaEBubi62bt2K9vb2mN1/Y62RRG39LhH8nITX403s8Qh/NOL/cBKB0cqsuAREK7O/kyxStdgqWkiCBKklodVq/a+JSFgC0mCzRftUp8FiTdkm2yLFYDCguLhYcIbmF6RZrVZcvHgRGo0mwCY/3ps1vvVXSSIfDH/NSpmQ6PV6bN26NaLleiJbEiPBmx4ODQ1hw4YNwu7kZMBxHHp7e9HZ2YlVq1Zh+fLlghtyrH9rLKktea3fQz+e1+vflRLu8ULapETxoQmIVmaFRapoJR5SGuHE+McRBAGKoEDpqIC99gzDwOf1wc2kPg2W6tRWLPAbDS0WC5YvXw6WZYUFaYODg4KNCy8s2dnZUd9QJdv5Nxb490RmZqbf+j+VJ5OTkxPRb4miqJCmjYnAL8DiOC7prsE0TaOlpQWTk5Pz9pbE4v7LE+2dW7Kt3/nHI2dz93J/yAkiKFrBbDQUpmiv9AgiehJ/Hfm11hpN6DQYx3GC2SQfrcj9/okntaUUjy2+24s3cuVtXGw2Gzo7O+F2u5GZmSlEK5FsXOR2/k0E/j2wZMkSaLXa1BfbIyF1aovf5V5cXJz0BVj8hLzBYAi5t4QkyZhFM5rUVlKs30Xw9Re+dTnmZolEhyAJ/v9DpMFmhdo3m9KRqptpscgSECINNjvbRNM0vJ65NBgfrchxk5BOEclC8DYufAOPy+USbPL7+/vBcVxAGkxs4+JwOJJWI3nkkUfw0EMP4Rvf+AYee+yxBX+eP8fdu3fjj3/8o7KFRKrUFr/LfWBgAOvWrUNxcbEEZxc9fFtxpAl5qSOSAOt3s0l2g8dQ9Revz5vyNbShivYEQYD2+eCZ7Wai0qSbaR5y77onCGFRlzgNxtdWGIYRvi/l8xdrN6KShSSYcDYu/OZNrVYLnU6HxsZGTExMJCUiOXPmDP7rv/4LGzdGP7vDk5WVhWuvvTa1QrKQlbgUEYnL5UJjYyMYhkFdXV1SuyD4Wszw8PCCbcXxCEm4DxzDMnA5XZJbv4cjXP2FUNpukdmiPamhYDKbwbEsaIYBE5TG0YhnL6K4MKZKKpP9uOI0GICAaEXKNFisTSTpJCRiwtm4fPzxx/jVr36F7u5u5OTk4Fvf+hauu+46XHHFFZKn4u12O+666y488cQT+NGPfhT3cRT97CcqJGNjYzh58iQsFgt27tyZVBFxuVz44IMPMDk5ibq6ugVnU+ISkhApAJ/PB4fdv/bVZDLJ/gGjGRoOh0O6nfFJNG0kSH8Kx2A0wmI2wzwbSdEMA4fTCbvDAbfLBZ/PBy7G1yYppDh6ImefP6PRCLPZLESiDO2fUbLb7XDNPn+xvLcXc0QSCd7G5frrr0d9fT3uvfdebNu2DVarFV/60pewZMkSeDweSR/zvvvuw80334xrr702oeOkRWor0m73ULAsiwsXLqC3txdr165FSUmJjGc5n+AFVNHUYuIRkuCipNvtTprVCZC8eZSkQBAgKQq6UGkcjwdulp29255Ng4milcUzlxI/UqbBlN7+myy8Xi+2bNmCRx55BBzHYXBwUNLP2XPPPYezZ8/izJkzCR8r5amtSPAXYIZhou7KcLvdaGxshM/nQ11dXVK7HsItoIqGRCISSa3fo0DWeRSF1ChCpnFmLUicXi8IAJRGM9cdlpqTTM3jRkGkNBg/ZCwWFXEa7HJJbS2EeDsiQRAxXU8Wor+/H9/4xjfw1ltvSeInqOiIJFYhmZiYQGNjI/Lz87Ft27aktgR6vV40NzfD4XCEnNJfiHCLrSKhJbWSW78vBMuycLlcEedDBAikvNgeinjOiCRJkDodtPxCKoYBPXth9HjcYBj2svcFiwTfBi7uBqMZevb58wgRjUajARFjiLdYhUTOXSQff/wxRkdHsXXrVuFrDMPg2LFj+OUvfwmPxxPTDamihYS/S6FpOmJIJ/brWrNmDUpKSpLafSNeQFVXVxdXWimeiGTKNgWHwwGdXgeDXn6X4lj9uRRXbJ8l0TkSgiBAaTSgNBrQPh90BgNIgpDFF2wxIk6DQYd5abDOc504M3MmYFdIJKFYrEIi5xzJNddcg+bm5oCvfeELX8Dq1avx4IMPxpzVUHRqC1i4Bdjj8aCpqQkulws7d+5Mqnd/tAuooiEWIWFZFh0dHRjsH4TRZIRWk4R6iM8LtysJ9ZA0vOoSBAGClMcXbIEHluhAqSU4DbZx/UaUWkphtVrR0tIClmXDzloAi1dIHA6HbNezjIwMrF+/PuBrZrMZeXl5874eDYqOSIDInVu89XxOTg62bNmS1FQWwzBobW3F+Pi4JHvcoxUSj8eDhoYG+Hw+bN64Gc8ffz6hx12IgH0lSRItuZF8sZX4eAr3BUsHzAZzgJeV3W4PmLXQ6XSCqOTk5CxaIVGyRUowaSEkwRPf4qJ2sPV8MuAdg6NdQBUN0QjJ5OQk6uvrkZubi23btqFhrCHhx41EsveVJItkVm1k9QVbJBFJMOJiO0EQyMjIQEZGBsrKygJW5nZ3d6O1tRUAMDw8DI7jkJmZuShEhTeDTGaG5ejRo3H/btqltsRbFLdv346srCw5T3EeIyMjaG5uRmlpKVatWiXZmzaSkHAch/7+fnR0dGDlypUoKysDQRAxexLFgiT7SuK5zqXhxTEWYYrVFywNn46EoYjwNyzBK3M9Hg9OnjwJj8eD5ubmeWmwdLmjD0UyLVISJS0iEl5IbDYbGhsbkZWVJWxRTBb8Aqq+vj5s2LABS5YskfT44dx/xTvct23bhtzcXOF78e4jWQh+f3ui9RACkZ0LFgvx/oXR+IIJabJQ0coiVZlY2n/59+eqVatgMBhgt9sxMTGB0dFRdHZ2Qq/XIy8vT0iDKcXcMRqcTqei97WLUfyzyqe2uru7ceHChYA78mQhXkAl12xKqIjE5XKhvr4eBEGETKElsiExHMkaauQ4zr8Lg/BHnSQxG9klcbJdsuNJdMrhlnnx//h/hkjqez8VxCIkLMuC4zihw5NPg/GWI/wCKn6zYbTOu6mGt51Jh4iEZVnlCwlBEOjr6wPDMPOs15NB8AIque5oSJIULhgEQQjT8fz63VBv+ER3totJ5lCj+LEAwO1yCzMEOo4DxUHWAnRaTKJHKtrPPm8cw4D2eucN9KU7sbyveYEN9fmgKGreZkN+AVV/fz8AzOsGUwp2ux0A0iIiIUlS2TWSqakpTExMQK/Xh7RelxPxAqpkFPT5DwLDMOjr60NXV9eC0/GxTgCHI5lDjeLaC3+3xXEc6NmpcZ/PB7CsrIup0nEfSahohZ1dPRA80KfRaNJaVGKNSIDQQhKMwWDA0qVLsXTp0gDn3ZGREZw/fx4GgyGgGyyVaTCHwwEAit1HAvgzJq+88gqOHj2qzIiE4zj09fXh/PnzyMjIQFZWVlJFhKZpNDc3Y2pqKmlREP9BaGxsxMzMTFSNBFJEJHw9RKfTSdJ9FkDQtSz4sfgIjCRJ6HQ6/2us14OY7RQLVYBWYrts0oWJdzEmSZhMpoCBPo/HA5fLFSAq6RatxLr5E4hOSMQEO+/SNC10gwWnwfLy8pCRkZHU59DpdMJgMCR1Z1K08FmTvr4+/P73v8fbb7+tPCHx+XxobW2FzWZDTU0NxsbGhL3SyWBmZgb19fUwGo1JjYLcbjcAv4hF+7iJRiRy10PExXZ+wdZCjzVbV45cgE4wWpG8RiLx8aJG6Oya72vFR3hK3Mm+ELFGJFLUjTQaTUAajF9AJU6D8dFKbm6u9DddQdjtdpjNZkXeAPBC0tnZiYaGBgAKa/+dnp5GQ0MDjEYjdu/eDZ1OB6vVKsve9lBEs4BKDviWYgDYsGFD1OIVb7FdvPQqGSaPTqczhscKfM7FKR1utrihpGhFiUkycYQnjlb4nexyLKOSklgibbmGEYMXUE1PT8NqtWJ4eBgdHR0wGo2CqMSyhz1aeCFRIvzNYX9/P8bHx/03Mik+JxAEAZZlBauRiooKVFRUCG9uqbYkRkK8gGrz5s0oKCiQ9fF4OI5DZ2cnent7sWHDBjQ1NcXULhtP+y/DMnA6klMP4f8WvoCf6Ac+qnbZKKIVWSfbk0kUr10sLrxSrR5OlFhTW3KfM0EQyMrKQlZWFsrLy0HTtNANxu9hz8rKCugGS/Rz5XQ6k2LCGg/85/rSpUsAgKKiotQLCU3TaGlpgdVqDWk1IvXe9mBcLhcaGhrAcRzq6uqSNsDED1Y6nU6hpZj3FYqWWCMSWeshQdAMDafDvzUxphA9hs+NUqIVJUYl4Qjpwjvb5OCeXT0sToOl4kIWa2or2eKn0WhQUFAg3HA6nU5BWHp7e0GSZEA3WDyfNYfDofhhysnJSQD+z1xKhYRlWZw+fRparRa7du0KOfwWyiJFKsQttqtXr05aYWt6ehr19fXz3IJjdQAmCAIUQYHhFhbaaGsUUiA2eORrP9ET34UrlmglZftDpCbBi7zYhVev1wvdcwzDwOP2wMW65tVWkiEsSkhtxYLJZILJZEJJSQlYlsXMzAwmJiYwNDSEjo4OmEymgDRYNNcZOZ1/E2XOdcH//yMjI6kVEpIksW7dOmRlZYV9g8qR2hLbzid7g+Lg4CDa2trmpfCA+JdbMXT45yfZ9RC+gG8ymUBRVBxCIg3hhvtYjoOP9sHhcMztZldgnSAVEAQBrVY7F61wLBia3xI5V7TnxUUYIpWQdF+zS5KkkAarqKiAz+cTopWOjg54PB5kZ2cLwhIufaVkexT+fEtLS2GxWGC321Of2srNzY148ZQ6tSVOKSXTdj6aOkxcQkJq4UboizXLsnA6nSAIIin1EGGgcdbgMS57FDnOMWi4j9JqodPphDoBh9kL5Owq3VgvTKnu2pLn0P5ol9KFWZ3rYuYK9hoKFCmNGC+27YharRaFhYUoLCwUbur4brCenh6QJBnQDcZnZcTbEeXgN7/5DX7zm9+gp6cHALBu3To8/PDDuOmmmxb8Xf75vuOOO/DrX/9aGUKyEFIKydTUFOrr65GZmRn3Aqp4cLvdaGhoAMMwEeswiazbDYamaTidTmi1WhgMBnmHDJNYwJcCgoBw5w2OAzNbJ/D6fGDdbpAUJYhKNF1NaTEpHw0+HxDmMxGyxZihwdAMvE5RizEf5cV5cY+1gUTpQiKGIAghDVZaWgqWZYVusMHBQZw7dw4OhwOHDh0CRVGyTtqXlpbikUcewcqVK8FxHJ5++mns27cP9fX1WLduXVTHKCsrw+OPP44f/ehH6SEkidZIxO65VVVVWLFiRdIudrzFSl5eHtatWxcxtRTOuDESoVIBfD3EYDDIPgfDC1akAj7fdx4NnGC5niSC6wQsK6zQ9fHRinDnHfoCmY6T8qEgensAkgLMZnBmM2AyAWEu0iRJ+t2ntbMbDlkGDM3MK9rH2mKsIWK7JKWTkARDkiSys7ORnZ0tpMEaGxtht9tx4sQJTE5OYmhoCNdffz2uv/56bNiwQbLr1t69ewP++8c//jF+85vf4PTp01ELCX+cDRs2pF5IFnpiNBqNYMwWz5MoXkAV7J4rJ+Lp/FWrVmH58uULnn88EYk4FZDseshCgqXUyCTSZZ8gSWhnu5rE0cr8riZN6qeOpfeOAeHzApNeEJM2gCDAGU3gTCbAbAbC3CgQBAENpYGG0kCv1wfUVtxuNziOC0iDkUT4ov1iS23FglarRU1NDf7whz/gK1/5CsxmM9avX48333wTTzzxBNrb22V5XIZh8Pzzz8PhcKCuri7m31+xYkXqhWQh+A8rwzAxD/3wC6g0Go1kC6iigReviYkJ1NTUICcnJ6rfi0dI+A4XcT1EipmNSHAcB7fbDZqmYTKboKEU/zaKD1G0gjDRCuB/vTluEbi6B9e0OA6E0wHC6QDGx8BpNIBpNloxm4EwQkoSJEhtUIsxQ0flC3Y5C4kYl8uFNWvW4P7778f9998f9410JJqbm1FXVwe32w2LxYKDBw9i7dq1MR8n5e2/0cALCU3TMQmJXAuoFsLpdKK+vh4URaGuri4m8Yq3RpLMeggvWADkESwF28iHilacTidomg5sMU7WUiqJH4BYoDmCoGlgegrE9BQAApzBMJcGMxhCno+4xRg6LOgLFqt/3GIVkuDtiHJ8pqurq9HQ0ICpqSm88MILuOeee/Dee++FFZMvfvGL+Od//mdUV1cDmFtxoJjJ9kjwvevR1g74BVT9/f1Yv3695AuoIsHPpRQXF2P16tUxv8FjFRKO4+BxegSDN7nrIWLnXqPRGNWbmyAIxdUQJDmb2QskQRDQGfT+1zrEUirwA5HpHq3MgwPhdgFuF4iJcXDi2orZDIS56VvIF8w2bkNra6uwjGqh9/RiFhK52391Oh2qqqoAANu2bcOZM2fw+OOP47/+679C/vxTTz2FZ599Fp/61Kfwuc99Dtdff/2cA4msZyoR0c6SeDweNDQ0wOfzoa6uLml92OId8onMpcQiJHz6zGV3JSW9lNDWxJh20cZ26HiQo8sq3NzKvN3sUkYrUqqTf4Iz/lNhGWBmGsTMtP9wegNgNoEzWwCjMey5BvuCFWQXwGAwoL+/H21tbcjIyBBaY7OysuaJxmIVklRsR2RZFh6PJ+T3vF6vsIzuz3/+Mw4dOoTS0lLs2bMHO3bsSA8hiaYF2Gq1orGxEXl5edi2bVvSdgnQNI2mpibMzMxgx44dyMzMjPtY0QoJvzmRJEmUFpdidHQ07seMhmRtTfSTfrfu8y6/USylUl60Im3USHjcgMcNwmoFR5CAyeSPVixmQBs6yiAIAhajBZWVlaisrITX64XVasXExIRgH5STkyNEK0ajcdEKid1ul9Ui5aGHHsJNN92E5cuXY2ZmBs8++yyOHj2Kw4cPh/x5iqLw/e9/Hz/96U/h8XgwOTmJyclJnD9/Hi+88ELqhSSa9EikFmCO49DT04MLFy4kZQGVGLvdLljO19XVJZxaikZIbDYb6uvrUVhYiLVr1+KFd19I6DEjkewusHRloUuwbNGKpBGJdIcKhuBYwGEH4bADowCn1YVtMRbPkeh0OixZsgRLliwBx3HCPnZ+ERWfXjWbzWAYZtG8PzmOm1cjkZrR0VF87nOfw/DwMLKysrBx40YcPnwY1113XcifpygKDz/8MO6++258//vfxx/+8AcA/khlbGws9UISDeFSW6lYQMVz6dIlNDc3o6ysDCtXrpREvCiKCisk4lkYsWBKtSUxmGROxQeQlGK7tFfNmFJlSo1Wkug/FrLFeLa2Es4iJXgfO+/A29XVBavViuPHjyMrK0uIVpS6yyNa5K6R/P73v4/r98rLy/HEE0/goYcewh//+Ee8+uqraGpqSg8hCZXaStUCKnExf+PGjSgqKpLs2OEGElmWRVtbG0ZHR+fNwlCk9HdhUnaBRSq28/NByd7gp6TSf0LRiqTPWYqeFXGL8Rhgan4FutMZYLZvB1NTA4QxLuQdeEdHR2E2m1FYWIiJiQlYrVZcvHgRWq1W2G6Yk5OTNBcLqeBt5JWITqfD6tWr8Z3vfAdf/vKXMTAwkHohiTa1Jb7ApmoBldfrRWNjI9xuN3bu3Cn5C02SpH9nuQjeXoVlWdTV1c2zTdCR0gqo1+uF2+2WtgssxDWKL94Dcy2iGo0G6fVxh7R38jFGK5KiEHXV2V3QvPYaNK+9BpAkmLVrwe7YAWb7drDV1fPEk2VZUBQlWI8sW7YMDMNgamoKVqsV3d3daG1tRUZGhhCtZGZmKjpa8Xq98Hq9SS+2x4rJZEJZWRnKyspSLyTRwNdIWJbFuXPncOnSpaQuoALmfLqysrJQV1cnSzE/uEYyOTmJ+vr6iPYq8Sy3CgU/ZOjz+WAymWRtVhBPxFMaCizDCnMFtNsNI8PKmtqRfLGVTFfhhaIVn9cL1u2WZneIQqz1teIXh2VBtbSAammB9ve/B5edDWbbNjA7doCprQVmDV+Di+0URQmdXlVVVXC73WHX5ubl5cXehSgzDocDABQbkQBzHV68r1paCIlGo4Hb7cYHH3wAjuOwa9cuWQ3NghkYGMC5c+dQWVmJ8vJy2e5mxELCP+bKlStRVlYW9jHjXbcrhnfu5TgOFotFti6Y4Il4iqSEO0qKomAwGMDRtP8KKmPbrOQ725NxdxsiWtFotfAC8Lg9cHPuqG1IQqMMIdGx4c+ZmJyE5p13oHnnHQAAW1WFopIS6K68EliyJOykvcFgwNKlS7F06VJhbS6/L6S9vR1ms1mIVrKzs1PeBcYLiRJt5HnhPnToED7zmc9g1apV0Gq16SEkXq8X4+PjKCkpwZo1a5L2QosjoC1btiA/P1/Wx+NrJG1tbRgeHg65MTKYRIVEPGRoMplkE0mxWPGFUI7jAGJudScHDiQfiZAKKkQrEIIAKJL0i++sDYl4Ypwf/Is6WlGGjgRGJAtAXriAJY2N0L79NjRZWWC3bvVHK9u3gwsziCxem8sbJfLRSltbG2iaFrYb5uXlRT14KyX8dsRUC1ooxGt2nU4nGhoaAChgIDHSi8QvoBobG0N2dnZMrpSJ4na7UV9fn9QIiGVZ2Gw2uN3uqNf+JtK1Jaze1etg0MvjQ0YQBFiWFawwxH+Tf2MhIYgJN5skIiDKtMymdkgiSUN+cZDqfSRiGxJ+sI+3eBdHKxE3HSoltRUhIgkF/34hnE5QJ06AOnECAMAuWyaICrt5MxAmfaXValFUVISioiKh7dZqtWJ8fBwXLlyAXq8XopWcnJykzKfZ7XbFd51NTU0BgPC+S7mQAJi7OxUhXkC1fPlyeL3epJ2P1WpFQ0MDCgoKsHbt2qT0p09NTeHixYsgCAI7d+6M+jHjnWhP1updPp2l0+ki5qIFUQlzF8aJfy5ctMLXExaIVqRPbUl8wAQhCAJajRZaTehNh0K0MpsKI2ZTiUpAx8Z4Fx7GLZPs7wfZ3w/tCy8AWi2YTZvmopUVK0Ieim91t1gsWL58ORiGEbYbdnV1weVyISsrS4hW5GqLV/J2RD5KWrduHfbu3YvOzk60t7crQ0iCCS5sDw0NCR0+csJxHHp7e9HZ2ZnU4UZ+/W5hYSEcDkdMwhXr/oZkDhl6PB6wLAu9Xi+kYBZitr4cQAhT2tkfJkASmLdKF0HRCsEfmP99hV344yZKr7Nwmw5500SNRgMtw0DHYd5zlWy0MepZVD/u84H66CNQH30E/OpX4AoL/e3FfItxmIs2RVHIz88XUtr8dsOJiQn09vYGFPWj8QWLFj61pcSIhD+nG2+8EceOHcNf/vIXXHPNNcoSknALqKRetxsKmqbR2toKq9WatOFGlmXR0dGBoaEhbNmyBQzD4MKFCzEdI5bUVrKGDMVFdZIkQVExrN0NcU7iL4UVFfjf5GQU0crlTLBpIr+QipkVff5noons5CDmiASx6x4xOgrNq69C8+qr/hbj9ev9KbDt28GuWhX2PWI0GlFSUoKSkhKwLCu0GPf19QX4guXl5SEzMzPuGofD4VBsxxZfbD98+DCef/55AMBXv/pVZQgJQRDChXxiYmLe0J1Go0l4S2IkeOt3fm9JMtoBvV4vGhoa4PV6hXrI2NiYZKt2g0mW1XxwUd3pdAr2FVI8ZsAhuPl3pMHRCjcbfoijFf5uXKPRgCLJhMUlVe7GUnSLUaQ/WgHDgJi98Imfq4DmBkD2aCVS11ZIEl0Ew7KgmppANTUB//3f/hbj2lr/P9u3A2F2CZEkiZycHOTk5MzzBWtubgbLsgHRSiw1Vr5GokT4G8LXX38dXV1dyM/PR2lpqTKExOFw4OOPP4ZWqw25w0POiGRsbAyNjY0oKSlBdXV1UjolpqencfbsWWRlZWHr1q3C3WFc+0ii6NqSZcgwBMEdYIC/mMkPWIn3ToR7njmKAkFSABvF600EXteijlZIAizDwOn1ggDmVsJq4pvFUEZ1IUE4Trgei1uMQzY3+P9FlmhFE2exXSqIyUlo3noLmrfeAgCwq1YJaTB2/fqwLcbBvmAzMzOwWq24dOmS4AsmbjGOlFJW8lQ7LyT8PM74+DgOHz6sDCFpaWlBfn5+2AVUcggJ3xHW3d2NdevWYenSpZIePxzDw8NoaWlBRUUFKioqAi5cUgtJMocMxbvbxREdbxEu3jvBr6zVarVC0VfAaARbVQXC5QLsdsDhAOENbW0dTKQUmPhrJEnCaDLNqxewLhcojUbY0U5GGa0sippLiCeMIIKm7CF/tJL0iGQByPPnQZ4/D+0f/wjObAa7bZtfWHbsAFdYGPJ3CIJAZmYmMjMzA3zBJiYm0NHRAa/Xi+zsbCENFlwPUXKxnT9PsdA9/vjjyhCSmpqaiHeCkdx/48Hn86GpqQl2ux07d+5MihUBx3E4f/48+vv7sWnTJhSGeBMmurM9+PGSMWQIAB6vBx63BwajATqtLmQ9JHjvBC8q/LZFYUPebFTAmUx+Z1gAnM8HghcVpxPgFn6OFizYs/7peb59kTAYAsTO453rbuLFTnHFTynPZ6HtiEmKVmKZI4m67iYRhMMB6tgxUMeOAQDYsrK5aGXzZiBMtM/7ghUUFAifS352RewLlpGRAb1eL2uN5Kc//SlefPFFtLe3Cz6Fjz76qLD1cCH4m779+/fj5ZdfFoY7FTHxstCdspQRyczMDE6dOiXMhyRDRHw+Hz7++GOMjo5i586dIUUEiOz+G45Qq0kZhoHdbofc+9v5DjCP2wOTySS0my4EQRDQarUwGo3IyMgQhq+8Xi9mZmbgcDjg8XjmXnOtFlxODrjSUrArq8CVlILLzgY00bct+++uZ4fm+S/wHV4cB45lhfMymUywWCwwzEZWbrcbdrsdTqcTXq834DVaFGmtOCAIgCQJkBQJkiSF9xg7OxzJMixYlot5X1ascyRA6prMyN5eaJ9/HoZvfQtkfX1Uv8N/JpctW4ZNmzZhz549WLNmDTQaDQ4dOoSVK1filVdeQWtrKz744APJMzHvvfce7rvvPpw+fRpvvfUWfD4frr/+emGaPho4jsNtt92Gz3/+88LXFBGRLIRGoxGcYhO5K+TTSsk0e+Rdii0Wy4IeXfHubBcjDBnqdDHti4+V4IgnkeeSjwr0en3YaEWr1c5GBSQ4iwWwWMAVAfB4/Hsu7A5/OiyKqxbHp2yIuUK8ICr8fxMEqNm6iUGvn4tWfD54ZlNzFP9apmoGI4kRSTSnIUW0ooslT8ifsxIixTgbdMQtxCtXrsS1116Lb3/72+jt7cWNN94IkiSxf//+uG3fg3njjTcC/vupp55CYWEhPv74Y1x55ZVRHYP/rP/4xz/G6tWr8ctf/jI9hIQPpxiGiSvPz7Iszp8/j4GBgbBpJTngd5ZEK1wkSfonvGMQTHGNJFlDhgzLwOkILKpLBR8V8OcvrquwLBuQAiNJEtDrwen1QG4eOJYBYXf4U2AOOxDmbi74kjlPVPz/EvDzBElCq9NBJxI7hqYBjoPT4wHLsMJxlHBdixkJxTCR2kosEYnQoCfJWSeIRDdtFRUVMBgMuPvuu/Gtb30LZ86ciXkkIBb4CXVxl2y0mM1mfPnLX8ZVV12lDCFZ6KLJCwlN0zELicfjQWNjo9Bmm4wiFsdxuHDhAnp6emLaWSKkB2aNDKNBS2mTOmQYrqguF+KZh1AFe7GwgKTAZWYCmZn+i4zLBcLhAOx2EB4PoolWiLnba2EWJVS0wkdJPpr2Pw+z7gyc+A5c7pbZNFCsUNEKZp+nUNFKzMV28YOkEE7CzwK/HVGj0aCurg51dXWSHVsMy7L4P//n/2D37t1Yv359XMegKAqrV69WhpAsBJ+DjTVfODk5iYaGBmRnZwe02coJX8h3OBwxF/LjERLWxwa4hcpZVA9uI052sTO4YM93XLndbnAcN2dSqNGAJEjAaARnNAL5+eBoWhAVLmjnS1iChhiDo5V5opGgdUtKSdJrScxaFxBhohXG4YbHQ0Y34yNzx1ZMSJhGTlbX1n333YeWlhacmPUnS4S0EBIg9oJ7f38/2tvbAybk5SZ4h3us6SWxkETD5OQkWptbBQt2uSfV+TbimCbVZSLUhDbto+Hz+uB2uQNmViiKAjQacFlZQFYW2FVF8N35EMhTp0CeOgWiry/qx5z9F3AcB4/H35bMp/UTtW6J40lI4JdTT6hoxURqA2Z8KD7ipChhYJJH6hmSRJA6IpF7juRrX/saXn31VRw7dgylpaUJH08RQhLPlsRwiNfSRmPDLhWjo6NoamrC8uXL497hzv9ONH8nvyWyfFk5jKPyORMHTKpbzP47fQVCkRQofeSCPf8PRxBga2rA1tQA998PYnAQ5MmTIE+f9nffRGEQ6na7hVQiQc2fc4lo3ZLEAb+oSfGNAeAXFpNGB6PJH3GysxGn1+uFm2FAim8OSFJZEYlE7uD8502uiITjONx///04ePAgjh49ivLyckmOqwghiYZobFJcLpfgjx9qLa0ciAcb169fj+Li4riP5b/gRO7cEs+jbN68GS6dC/g47oeMCMv602YkSSp2QCoUwQV7RnRBcrlcMDJG9PT0ID8/3/93lZSAueMOMHfcAbhcID/+GOSpU6BOngTGxgKOzXEcXE4nOCDA6jvaYUiIRQUJDvgppGtLSvhiOyHqmtPD/15kaBr0rHsCAYCkqLnW7VTv7pAwtWW322UbS7jvvvvw7LPP4uWXX0ZGRgYuXboEAMjKykroepk2QrJQRDIxMYHGxkYUFhZi7dq1SbE6oWkazc3NmJ6elmywMZKQ8PUXp9MpNA4M24cTfsxQJLuoLifi9mKWZaHT6mCz2dDV1QWdTic4vObm5oIyGsFecQXYK64ADYC4cAHkyZOgTp0CmpvhnBXWSN1qCw5DholWkm1HEvEkU0S4gUSSJEHqdNACQrTi9fnAwX/hFaKV2dc6qZGKVivp48kZkfzmN78BAFx11VUBX3/yyScD5kJiRRFCkkhqi+M49PT04MKFC1i9ejWWLVsmxynOw+FwoL6+Hnq9HnV1dZJ5WIUTEofDgbNnz8JoNGLnzp3C3bYUq3aDSXVRXU7I2a2CvNuyzWbD+Pg42tvb4fV6kZubKwiL0WgEV1UFpqoKtv370fz++1gxPIyy/n7ggw+A6WkAC5s2RhutBBTsESFa4Y+nlLSOhETTtcVHK9rZ4Uej0TgXrYhrK/y+FZlvKjkJoxF+uZZcNRK5PsuKEJJoCGWTQtM0WlpaMDk5mTTrd2DO6LG0tDSsP1i8hBKS8fFxNDQ0oLS0FNXV1QHCm8iGxFC43W54vV7FFNXlgP+bxPsmqqur4XA4MD4+jpGREXR0dMBkMqGgoABarRZdXV2oXLMGS2+8ETRBgGZZEK2toE6dAnv2EIBoC/bB5xL+vwmCiLgZkmEYkAkO6YY9kRQR6xwJgcBoBXw332zE4na75Y9WJBQSj8cDmqaT4rghJWkjJBqNJiAi4SMCnU6Hurq6pKRfOI5Dd3c3urq6ZDN6FAuJeNHW2rVrUVJSMu/nQ1mkxANf5GNZVvY2YiUi3o63YsUK+Hw+TExMoK+vD1NTUyBJEtPT0xgeHkZ+fr4/UtuwAfSGDfCN7AD72t/Pzaw4HUCUnXcLRiui8wtuL/Z6vaDt9qhclRc8D4WYvcTitRWy2C6urej14FgW9GyNzOX1ggPmRCWB5yvgNCRu/QWgWPffcChCSGJNbfEdUnJEBOEQRz/bt29HVlaWLI/DCwnLsmhtbcX4+HjEaEuKiCR44ZWK/2LjcDjgcDiEGaSxsTFhiVFmZqYQzXAcF9BezHGc31ySFxZfdGuiAz4G3PzxyeD2YqPJBMZoBE3T8M3efQs7tGM1mlSAjmjY2XXLURLNKRMkCe2s03RwtMLy0Qrv9hyvMafErb8EQUjuGCE3ihCSaKAoCj6fD52dnejp6Um4QyoWxIuv5I5+KIqCx+PBhx9+CJZlQ+5nEaMhNf7hrjivBMELry4HFnquWJbFuXPnhG2ZvLhmZWWhqqoKHo8H4+PjGB8fR09PD3o8PXC73QEXb85s9q9wLSwE5/WK/MCc0aWRiMCmrVApMA5zQ5p6vT5gNzu/mloY0qQWuPtWQGpLdsPGSNGKyxV3tCJ1RCLuBkwXFCMkxOyQV6Tvj4yMgCTJpFm/A/5usIaGBhQXF2P16tWyRz+8L1h+fj7Wr18f1YS7htTAx0Y5rS1iMRfV44WmaTQ2NsLn86G2tjakuOr1+oCVq8c6jwH9CO8HptOB0+UCObn+VlWn0x+pOOxAlOsRwqXA+NeM92fjrVu42UI0Tc8OaTLueSkwYqG8WpKJKa0FJDxHMi9amX2+gqMVaqHoTuLWX1VIZGJmZgaDg4MgSTKuifF4ENcn1qxZI8n050IMDw9jZmYGRUVF2LhxY9RvpniERFxU12g0SRER58BKTDfvgaWqHqYVLSC1sYufFIT7W91ut9CJV1NTE5WlDkmSyMrOEgRnYT8wkXsx4Hcvttv9wuJ2Izo/MP7/CRCzRp/iv8vvBxYUrbAsaIYGQzPzhjQpDQX53NmiJ1afLUkn2wkioE1cHK34+GiFTxkGRysSComcrb9yongh4Se4c3JyhEEzuWEYRtgfn4xuMN7ksbe3F5mZmcjLy4vpjkRLaeGiXVE/ltjgkRRdiOSG9Rgx2fBJTDZ8EqTGC1NZGyxVZ2Guqoc205aUcwBCp7Z4e5vc3FysWbMmtshTdLho/MDEC7wE9+K8PHAMM1ewdzgWXjcsDEPO/b8gJuAC/p0gCGg1WmHxWOBmSBZGjwdalpPfaDICMae2ZJxsDxet8LUo/gaB0mgktUex2+3zNiamA4oRkuDUFsuy6OjowNDQEDZv3gyPx4OhoSHZz8PlcqG+vh4kSWLXrl2yd4PRNB2wrbGzs1OyLYnBpLqoTpBzfxdL62Dv2gx712bgMGAo7IW5qgGWqnoYirtAkPKJW7BwWq1WNDY2Yvny5fPWH0d1vDBRxDw/sKAJ+3l+YFQI92J+M6THHdW5CKIyW2ThOC5QVGZTYPzdt2F2MyTs9tQNQ86i42JLGyctGSd6vhAiWhm+dAnW5mZhdW4itcZk+GzJgWKERIzH40FDQwNomkZdXR1MJhMuXbok+bawYKxWKxoaGpI2He90OnH27Fno9Xrs3LkTOp1O8r3tPIooqhPh/y73aBnco2WYOLkPGvMUzBVNsFTVw1zRBFIX3UU0HoaHh9HW1obVq1eHbK+WkmgWeAVEK7x7cUGB372YT4Hx64ajuMITs067vKgAs8LHDz3ORiskQYCgSGmsW+IkHgv5VNy3B0crxeXl8M3ajZw/fx4mkwl5eXnIzc1FdnZ2TNcRvkaSbihOSGw2GxoaGpCbmxtQbJZy3W4wHMehr68P58+fR3V1NZYvXy7L44gJV8SPxy5/ISHx+rxwu9zQG/TQ6/QpK6qLI5JI0I4sTDXvwVTzHhAUDdOydpgrG2BZeRa6nFFJzoV3ROju7samTZuQn5+f0LFiZSE/MIqioNVq5wr2Go1/vXB29lx7cYxp3kjRCsdxQSm6OUfekNYtITzGEiWu1Faq550IAvqsLKxYsUKYP7LZbJiYmEBbWxsYhkFOTo4gLAv5Wak1Egno6+tDR0cHVq5cibKysoAUQ6jJding5zXGxsZQU1ODnJwcyR8jGP7vDGXpIsW6XTGpKKqHJUJEEg6O0cDRsx6OnvUYfedvocsdhmXlWViq6mEsPR+1OAUck+PQ3t6O0dFR1NTUIDMzM+ZjBBxPgiRLsB8YH614PJ6AFJlGo/Hb2EtwsRFHK8GX8OAJ+7C7Vvw/EGjdEifxFNsVMTYrivC1Wi0KCwtRWFgIjuNgt9sxMTGBkZERnD9/HkajEXl5ecjLywsZraiprQRpb2/HwMAAtm3bFnLtoxwRCd+lAwC7du2SPeXDzyeMjIyE/TulSm2lsqgejngu+sF4rcWwfnAzrB/cDMrggLm8Geaqs7BUNoIyOqI6xrR9GjabDdu3b0+KQ3SsiAv2QOC6YY7jJJlkD4UgBFEMQ4p3rUgVrSS7/VcqwhXbCYJARkYGMjIysGLFCtA0DavViomJCZw7dw40TSMnJ0fY2W4ymWC322UVkmPHjuHnP/85Pv74YwwPD+PgwYPYv39/wsdVjJAUFxdj+fLlYS/mwRYpicKn0PLz87F27VpZ19MC/pmN+vp6oe4T7gJGkmTMkVewTYq4qK6onnRS2hsBxm3G9LmdmD63EwTBwljSCXNVPSxV9dAXDM77ed4GJkObgdra2qR0AEpBpHXD8wr28RIQgiw8DCn86ALRSix77GNNbaV+8mWWKG9GNBpNQLTicDgwMTGB0dFRvPzyy/jVr36FiooKFBQUCPNdUuNwOLBp0ybce++9uP322yU7rmKEJDs7O6JQUBQFlmWFrpNE4Lcnrlq1CsuXL5f9QjszM4OzZ88iMzMT27ZtizifEE9EIu7aohkaLqcLGo1GcZPqBCFnJxYJ50A1nAPVGDv6GWizxmCZ7QIzLm8DSC+cTicoioIlwyKpiCQz0gtuL16wYB8tEYeBI/9YpGglls2Qsk+2y0Q87b9ib7eysjJUVFQgKysLTz75JN566y3k5ubi6quvxr59+/B3f/d3kp3rTTfdhJtuukmy4/EoRkgWetPzd1sMw8S9e51PLV26dClp2xNHRkbQ1NSE8vJyVFZWLvh3JpLa8vl8cLlc0Ov1ktnaS0ocNZJ48U0VwPbxdbB9fB1IrQvaknpkrGxA9upWcJD2PKSokcRDNAV7SaKVgMcM/O9od63Mi1ZCtBena2pLioHE7Oxs/O3f/i3eeOMN3H777di7dy9ef/11dHV1SXCC8qMYIVkI/oNA03RcQuLxeFBfXw+WZbFr1y7Zc+Mcx+HixYu4ePEiNmzYgCVLlkT1e/EIiYbUCEV1o8kIrUab8npIKAiJU1tRwXGg3Tow3XXw9OzCxNscLEsH8D8OCrt3s6iuTvx5UspzHam9WFywD2X3wS1Z4p8lEbcXR0Gs0Qo3Kxah2ou1DBeTOEg62Z4AUnttZWRkYN26dVi3bp1kx5WbtBESkiTjao0FgKmpKZw9e3ZeS7FciJ2Cd+zYEVNXUKxCwjAMbBM2+Hw+xRTVwyJBsT0WOM6/hpUkSeHixHEEnENl+J//0eB//gfIzwd27WKwaxeLmhpWSreLlBIcrYjrKiH9wEK1F9vtgMMOwhedlU1C0QrHAW4adrtdmBhfMD2nlIhEwqFlp9Opdm0lQiJbEiMxODiItrY2VFVVYcWKFbLXQ/jJeIqisGvXrphTTHwtKBrcbjfOnj0LglVYUT0MctZIguFm62liEQnF+DjwyisUXnmFgk4HbN3KYtcuFrt2MYgyiExZaisWIhXs5/mBid2LUeR3L+Yn7J1ORFvmjilaIQlYtAYYZm3xvR4P3EHpueDXUjHPugzuv+mGYoQkGmKZJRFbrGzZsiWhgbNosdlsqK+vT2gyPtqIhI+y8vPzUVZahpYLLfGcclIhSOnngELBsSw4cCApErEkP7xe4PRpEqdPk/iP/9CgspITRGXdOi7ls29SEbMfmE4HLjcXyJ11Lxb7gTGxuxeHay/WcBCEAwgUPI/XO5eem3XkBRSS2pIoIpF7za6cpJWQRNsC7PV60dDQAK/XC693D0ZG9MjL42SNggcGBnDu3DlUV1dj2bJlcUcH0QjJpUuX0NzcjKqqKpSVleG1U6/F9VhJJwkRCf/ckWTi6cuuLgJdXRT+8AcKWVnAzp1+Udm5k4X4plGxqcQoiNkPjCTBZWQAGRl+MXC756KVKN2Lw7UX6xhiLs2FufTcPMHzeMC53f76F02DmE17pwyJbeTlXJFht9tx4cIF4b+7u7sFJ5FEHD0UIyRSpbamp6dx9uxZZGdnY+vWrfj+9/X44Q9J5OUBV1zBYs8eFjU1HKRqahJHPlJ0gkUSEnEBf9OmTSgoKPB3sRGKeRkjI2vXFjf7vBGyXFSmpoDDh0kcPkxCowE2bvSnwHbvZpVxWywRMfuBGQz+YnN+PjiGBmF3+EUlGvfiWfiPvg7+HSkcMLu5S2yLT8xFK7O7VhwOh3/bod0OkiSFukpMmyElQMpiu9wWKR999BGuvvpq4b8feOABAMA999yDp556Ku7jpskVyM9CqS3ecr6yshLl5eUgCEK4WZiYAF5+mcTLL5MwGoEdO1hceaU/dRGvQ4bX60VjYyM8Ho9gLpko4YSEYRi0tLTAZrNh586dMJvNYBgGLMtGtEhRElJMtodk9sLC7+dY8McTfDiaBs6eJXH2LIlf/hIw5W/AaOZdMFfVw7S8Xb6/M8ks1F48r2BPidYNA4DTOZcG83oWfDwdOzu86H9wAKJoL2jnCv86G41GEAQBhq/5LLQ7RA4kEpJkpLauuuoqWSLotBOSUBEJv1VwYGAAmzdvRkFBgfC9UF2+Lhdw9CiJo0f9x9y4kcOePSyuvJLF0qXRnYvdbsfZs2dhsViwc+fOuGdbggklJB6Px19UJwjBJZh/HkiSlGRve1KQo/2XF5HgjX9JZHTIAGvHTbCeuQmU3glzRZN/wr6yEZTRnpJzkoOY/MAAwGQCZzL53Yt9vjlRCdNeHGogUXhNxWsmOE74jAi2+BoNNOF2h0ixlz0SEgmJy+UCy7JJ2/4qJYoRkmhe3FA1Ej4qcLvdqKurmxcWLvQaMwxQX0+gvp7CL35BobLSLyp79nBYvTp0XWV0dBRNTU1Yvnw5Vq5cKekbM1hI+FRdbm4u1q5dO3vO/ueAnxSOxkZeCUjdtcW396ZSRGbPRPg3xmMKsG0xLO2ChbdtKexP4TlKS8x+YFptyPZiwm4HaH978UKmjfxrzM76yGm0WpAEMS8Fxp8bscBedo1GE1UEGxGtVrIWZD59qBbbZSY4IuGtRzIyMlBXVxcyKtDrY7t48QXWp54CCgo4XHGFX1i2beOg0XDo7u5GV1cX1q9fj+Li4kT/pHmIhYSfiq+oqEB5eblQcAzet62h0uRllDAi4TgWHLtwe28q4TgSrsGVcA2uxNh7d0KbNQ5LZb0/BVbWBlKTnC62ZBCqvZiPCOYV7EXtxVxREeD1gLA7oPVWAqNT/ru7MPA+csEWQMEpMOFrxNwee/AFe4aBVxyt8OcWx3uJk3Cw2T5b61GatVE0KOoKFLwlMRhxjYTvXFrIeiSR13lsjMDBgwQOHiRhMnGoqBjBypUO/M3f7EBxcWLW4+Hghy4vXryIrq4ubNy4EYWFhUI9JFhEAEBHpkdqS6raQbztvaIjSHIeseKbyoft7HWwnb0OpNYD04pWIVrRWCZTck5yELMfmE4PLlcP7tavwvPwbpAffgjq1CmQH3wA2OZWMPMiotVooA+62IpTYMCcK/G8aIWioBOl5xiaBs3499gTgFCw11BUdNGKhMOI/AyJ0ufBQqEoIVkIjcZvBXL+/Hn09fVh48aNKCoqivg7UtwwsCyL0VEXRkcz0NS0Ba+8QmDLFk7oApM6MGFZFn19fdixYwcsFktEEQHmu/8qFgm6tqRs700lrE8Pe+dW2Du3giA46It6hf31hiXdSg2yYiZqPzCtxn9DlJEB9pprwF5zjd8+5dw5UKdOgT12DO7GRui02qjWX/Ouw0D4aIUgCGi0WmhF7cUMPwzJssK5URGiFSk7tngLeVVIksD4+DgoisLOnTujyiUmKiQ0zcDl4kNpIwjCH3l/9BGBjz6i8NhjFKqq+GI9l5B3k8fjQVNTEwBg+/btMBgMAUX1cG+wdCm2JxaRcGBZ/3Ob0pkBGeA4Au5LK+C+tALjJ26HxjIJS2UDzFX1MK9oAalbuOMpXQgo2HMsGNovLE6HE+dazmGJcwny8/ORm5vrX8S2di1sy5bh48pKrMjIQMXwMNhTp0B+9BHgiG7/TEC0MmseGam9WI+59BwjHoYUCYtwTIntUaTo/EwFihKSSKktu92O/n5/sbKuri5qG/BYayRivF4f3G4X9HoD9PrwF+sLFwhcuEDhySeBwkIOe/b4hWXrVg7RNnPNzMzg448/RmZmJqanpwPqQeJlQaFIl2J7/BHJbGcWomvvTXdoezYmG6/CZONVIDU+GJefE1Jg2qzxVJ+eZJAECVJLCp/llZUrQZEUOjs74XK5kJOTA4vFgsHBQZSXl2NFeTlYAOwttwA0DbKpCeTJkyBPnQLR2xvdg0bbXhxpGNLlElJgsa47joSa2pKZkZERNDc3IycnBzRNx7RLIt6IRHDTNZqg1Ub/NI2OEjhwgMCBAyTMZv809JVXsqir4xAugBodHUVjYyPKy8uxfPlyvPvuu7BarcjLy4vKYDJdUlv+bAMLjotBDIT2XgIEoUwRkbPiwtJaOC5uhOPiRoy8eQ/0+QOwVPkL9saSThBk+k7VB1OQU4DqJdWorq6Gw+FAf38/+vv7wXEchoaG4PV6kZ+fj5ycHJAaDditW8Fu3Qp87WsghoZAnjrl/+fsWb/fTRSEay8OOwwJgJ0t2NM0jfHJSXSeOiWsz83JyYk7YpZ7O6KcKPoKxHEcurq60N3djQ0bNgDwj/THQqwpTP8WPX8/t9lsAUXFf/FyOIB33iHxzjv+aegtW/jWYhZFRf7H6unpwYULF7BhwwYUFRWBpmmUlJSgra0NAJCfn4/CwsKIomLUKm9dbFiIEEZLYZCrvTeNHU3gGS+FZ7wUE6f3gjLakb/7ReTUvpnq05IE8Q2Rx+PB0NAQqqurUVxcDKvVivHxcbS2toKmaeTl5SE/Px/5+fn+KfylS8F86lNgPvUpwO0G+dFHIE+dAnXqFDA6GtXjhyzY+/8l4L8JkoR2tqGguLwcXFUVJiYm0N7eDp/Ph9zcXEFYYunASlefLUBhQiK+WNA0jaamJszMzGDnzp3IyMjA2NhYzO6/sUQkDOPvCiFJEhaLtCEmTQNnzhA4c4bCf/wHhVWrWFRUDKC8fBS33rodmZmZYBgGHMdhzZo1WLNmDSYnJzE2NobOzk40NzcjNzcXBQUFKCgoCCg47i7djef2P4cTAydwov8EmkabwEa5TyLZECQLjl04ykqH9t5Uw7gs8IzF74+kNPha38TEBBobG1FdXY2SkhIACFhROzMzg/HxccHZOyMjA/n5+SgoKEBmZiYIgwHsFVeAveIK0ACIri6QJ0+COnUKREsLEKW7djTRCmcwIC8vDwUFBQHrc0dGRnD+/HmYTCZBVLKysiJGKw6HQ62RSInD4cDZs2dhMBhQV1cnDD3F4v7LE22NhG9P1Ol0svdxcxyH+noXzp7NhclUioMHCezaReOKKwhs2TJ3952Tk4OcnBysWrUKDocDo6OjGBoaQnt7OzIzMwVRMZvNKM0sxWfWfgafWfsZTHmmcGrgFI73H8eZ4TNw+pyy/j0xEUWdRLCAj7u99zIiiVsn5UZLajE2Nobm5masWbMm5JwWQRDIzMxEZmYmKioq4PV6MT4+jvHxccH9gY9U8vLyoNVqwVVWgqmsBHP33cD09Fx78enTfhO1KAjXXszp9cI1iSRJGI1GLFu2DGVlZfD5fLDZbBgfH0dLSws4jguIVoJXTKgRiYSMjY2hsbERpaWlWLVqVYCCR+v+KyaaiMTj8cLjccNgMEKnk7dw7Y96HKAoSrj7GB4GDhzQ4MABDTIygLo6Fldc4XeZ5W9QzGYzysvLUV5eDo/Hg/HxcYyOjuLixYvQ6/UoKChAYWEhsrKykKXPwo2VN+LGyhvhZbw4e+ksjvcfx/sD72Pcmdpi7UJbEv3tvXKLiNS5rRTmyhaJrxcATNmm0N3djfXr1y/Y1s+j0+mwdOlSLF26FCzLYmpqCuPj4+ju7kZLSwuys7MFYTGbzSAyM8Feey3Ya68FWBZEW5tfVE6eBNHZGfW58u3FlNkMjUYDdvbmR7xKmCRJITXNR1ITExMYHBzEuXPnkJGRIdRVMjIykiIkv/rVr/Dzn/8cly5dwqZNm/D//t//w/bt2xM+rqKEpK+vD21tbVi3bh2WhjC9imexVSQh4TjA7XaBpmmYTGZoNPLOJvh8NFwuJ3Q6/Wxqav4FaGYGePNNEm++SUKr9S9auuIKFrt3Mygs9P+MXq9HSUkJSkpKwDAMJiYmBAEGAusqOkqHnSU7sbNkJ/6J+ye0T7QLotJlS8E+6Ag2KYtlRiSZEIskIvH5fLjYeRF7tuxBIf9GjxGSJIUofuXKlXC5XEK00tXVBZ1Oh4KCAqFgT1EUuPXrQa9fD3z5y8DYGKgPPvB3gn30EeBcOJInjEah+YcXE37uS3yt8qfLLcjMzER5eTm8Xi8mJiYwMTGBQ4cO4ZFHHkF1dTWKiopgtVqRm5sb13MQiT/96U944IEH8Nvf/hY7duzAY489hhtuuAEdHR1xP+c8BKegZQpWqxUMw4RdTet2u3H06FFcf/31UXdGeDzA1VfPjzJYlpudtOVgMplkn00IjnpifdoJAqiu5rBnD4MrrmBRWTn/9zmOw9TUFEZHRzE2Nga32x22rgIAw/ZhHO8/jhP9J9A42ggmStvvRLjw+K9BO4Nf3+S29xIAqvOqJTvejHcGgzNDkh0vFnJrX0fhtc+k5LGlgrdSObD/AKpKqmR5DIZhhIL9+Pg4vF4vcnNzhWjFGHzH6fOBbGycay/uD+2TRn/5y2DuuSfk93gx4QVG3FrMz4WRJAmv14ujR4/iscceQ2dnJ8bHx7Fz507ccccd+MY3viHZc7Bjxw7U1tbil7/8pXB+y5Ytw/33349vf/vbCR1bURFJVlZWxBqIePlOtBd+vR4gycD6GjNriUBRGhiNBln7tkNFPfFoN8cB7e0E2ts1eOIJYOlSfrKewaZN/u19BEEgOzsb2dnZQl1lbGwsbF2l2FKMO9fciTvX3IlpzzRODZ7C+wPv4/TgafnqKsGpmDRo71U0SVxfLAdenxcetwdGoxFFBdGls+KBoijhvc8XxcfHx3Hp0iV0dHTAbDYLopKVlQVSqwVbUwO2pgb4+tdBDAz4ReX0aZANDXPtxRHqqaRo4RbLsgH/BEcr1157LZ566il885vfxGc/+1m88cYbmJiYkOzv93q9+Pjjj/HQQw/Ne9xTp04lfHxFCclCF3S+/ZVhmJhmSfR6v3U84L/7cblc0Ov1UVktJIK/ldgJjuNmB41IyXYBDA0R+POfKfz5zxQyMoBdu/yisn17YF3FbDZjxYoVQl1lbGwsZF0lU5+JGypuwA0VN8DH+FA/Ui9EK2POMUnOGQhMxSjHvTcxUhrTp3Fqy+v1wuPxwGQygaKopHnGEQQBi8UCi8WCFStWwOfzYWJiAuPj42hsbATHcUInFl8U50pLwdx5J5g77wRcLqG9mIuylhMsKnw9hf93mqbR3t6OjRs3orS0FF/60pck/ZvHx8fBMMy82lNRURHa29sTPr6ihGQh+FAw1jqJweAXErfbA6/XE/OQYTzwrcQURYpa+uS54szMzG3v0+n8dZU9e/wFe35hYzR1Ff6Do9VosX3pdmxfuh3f3PFNdEx04ET/CRwfOI4L1gsRziQKyLk9Etysf5ja3hs/6bpEK1hEAKRsQZtWq8WSJUuwZMkSIT08Pj6O3t5etLa2IjMzU/h8WCwWEEYj2D17wO7ZE9fj8YLC/90sy+Kxxx7D4OAgVq9eLdnflUzSSkiAeFuAWTidbjBM4kOG0SBuJdbrDUhmV4/XC5w+TeL0aRL/9m8arFnD4YorGOzZw6K83H8eFEUF9OXzdZULFy6gpaVlXl2lOq8a1XnV+OLmL+KS/RLeH3gfx/uPo2GkATQb22tBEOxce2+KRCS9k0FBpGFE4vF64PV4A0SEJEiQCkhtitPDVVVVcLvdQrTS09MDjUYjpMB4P7BE4DgOv/71r/Hv//7vOHr0qCQdVKHIz88HRVEYGRkJ+PrIyAiWLFmS8PHTUkhiiUicTiccDi84ziT5kGEoEi2qSwnHAW1tBNraNPjd74DSUn9d5YorGGzcGF9dZYllCT61+lP41OpPwe614/TgaRzvP44Phj6A3bvwNkCG8yVoAa8iJt0iEo/HA6/XC5PZBErUnadU41GDwSBE8izLCnMhYj8wvhMs1mFCjuPwxBNP4Mc//jFef/112UQE8LdJb9u2De+88w72798PwB8JvfPOO/ja176W8PEVJSTxbkkMx8TEBBoaGpCZuQPT02ZZb379RXU3fD5fQkV1ORkYIPDccxSee45CVhawezeD3btZ7NjBCjXDWOoqFp0F15Zfi2vLrwXN0mgYaRBaiy/ZLwU8Nl8vAsEswvbeFL7OaRSR8CJiNpvnNcukKq0VCyRJCsOEvB8Y//k4f/48jEajEK0s5LnFcRyefvppPPzww3j11Vexa9cu2c//gQcewD333IOamhps374djz32GBwOB77whS8kfGxFCUk0RBuR9PX1oaOjA6tXr0ZRUTYGB+U7J3FR3R/1SFdUl4upKeDQIQqHDlHQ6YCaGn9dZfduBnwLeyx1FY1Gg5riGtQU1+Aft/8jOq2dONF/AicGTuDc2LnZehEFjZaE/E3Glw/pYtrodrtB03RIEQHSyMFaBH/TVVZWBpqmYbVaMTY2hpaWFjAMM88PjIfjODz77LN48MEH8fLLL+PKK69Myvl++tOfxtjYGB5++GFcunQJmzdvxhtvvBH18GckFDVHwnEcvAu4dn744YdYunQpSktLQ36fZVmcO3cOIyMj2Lx5M3Jzc/Gtb1F4/3158q9ify6TKY3ME8NAksDatf66ypVXsli+PPy8ytjYGEZHRyPOq0xNTeHomaPoJ/txkb2I5390G1wjy5L5J4VktYRzJNOeaQzZhyU7XiwUXPUc8upeTcljRwsvIpHmtZZYluCF219I8pnJg9gPbHx8HNPT07BYLDhz5gzWrVuH0dFRfP3rX8cLL7yAG2+8MdWnKwmKikgSTW15vV7U19eDpmnU1dUJQ0ZyWWf5i+ou6HTapBfV5YJlgZYWAi0tGrz0EvDCC/OXKonrKitXrhTqKsPDw2hvb0dGRgYKCwtBUf7dEptXbsaty28FQRCwPkOi3umG3WuH3WsHo1BzyXRB6RGJy+UCwzALDv2mY0QSjnB+YI8//jh++MMfwufz4YorrsD09DQmJyeRnZ2d6lNOGEUJSTSES23NzMzg7NmzyMzMxLZt2wK6KeQQEq/XC7fbDYPBICy/WWxE2xwnrqt4vV6MjY2hr68PdrsdOp0ObrcbNpsN2dnZ0Gp0yNBlIEOXAQBw+pywe+2Y8dnhY3wy/jWLFEK5iUJBRMymBTuylFpslwLeD+yuu+7CO++8g3/913/FzMwMfvjDH+Kuu+7CyZMnUVtbm+rTTAjFCUmkLYlA6PbfkZERNDU1oby8HJWVlfMiGymFJLCobvKvA12EIgL4VwrHilarhcvlgtvtxpYtW8CyLMbGxoQVwtPTtfD5MqHRaEAQgElrgklrQiEK4WG8sHtnYPfa4abdiyC+kx+lem25XPxOn+g6JdNlOVu8HD58GPfeey9+//vf49Of/jQA4NFHH0Vvb29Il+N0I+1ePXFEwnEcLl68iIsXL2LDhg1h+6GNRmkuSfzSK45j06aonggxjuuAZVm0tbXBZrOhtrZWcDIVz6totSQ8Hg/cbhcoyr+uVKv178DWUzrojXnIM+aBZml/+stnh9PnBCvx88xBuubjlL4DFJja4htPTCZT1O32iym1FcyRI0dw991347e//S3uvPPOgO+VlZWl6KykJe2ERKPRwOPxgKZptLS0YHJyEjt27Ahr9AhIE5GwLAuHw19UN5vN8F+GlPchlpJYIhKfz4empib4fD5s3759nv0MX1fJydHCYiHBsixomhYM+yiKhEajhVarAUmS0JAaZBuykW3IBsuxcPics9GKAwwnRTpHSilJHYTCUlvxiAiweFNbx48fx2c+8xn84he/wF133ZXWVkCRUJyQRJPa8ng8+PDDD0FRFHbt2jVvQUww8e5t5+GL6lqtdnbpFYfFLiJA9BGJ2+1GfX099Ho9ampqIk778tuCydlVpXx9yS8qNLxeDwiCgEajhUajgUZDgSRIZOgsyND5IxynzwW7zw67dwZeRdRVUrmPRDnvw3hFBFicQnLq1Cnccccd+NnPfoYvfOELi1ZEAAUKyULwQ3IlJSVYu3ZtVC7AiUQkl0NRPRzRCMnMzAzq6+uRn5+P1atXL/h6hPo2QRDQarWCEScvKq5Zp02/oGhEdRUjTFojCk0Fs3UVv6hcjnUVJUQk/BwVQRBxiQiw+GokZ86cwac+9Sn86Ec/wj/8wz8sahEB0kxIBgYG0NfXB6PRiHXr1kX94sQbkbjdbni9i7+oHg6W9TcXhHuaJyYm0NTUhLKyMpSXl0f1ekTj/s+LBuB3eqZpeoG6Si7yjLmgWQYOnx0zXjucPofkdRVFkuKIJFhE4mUxRST19fXYv38/vvvd7+L+++9f9CICKFBIQj3pLMuio6MDQ0NDqKiowPj4eEwvTrR723n4ojrfdUKSi7uoHgmaBkI59g8NDeHcuXNYs2ZNyG2W4aBidEehKAoURUGv10dRV6GQpfevGmY5Dk6fQyjY08FLuxZHiSSlFilSiQiweIrtzc3NuPXWW/Gtb30L3/zmNy8LEQEUKCTBeL1eNDY2wuPxoK6uDg6HY56D5ULEktpiWVb4cFgsl0dRPRLBQsJxHLq7u9Hb24vNmzcjj/epj5JEFiDGVlchYNFZYJmtq7hol39exWuHl/EuHh0hU5Pa4kWEJMn52wXjYDEISVtbG/bu3Yv7778fDz300GUjIoDChcRut+Ps2bOwWCzYuXNnQMdWLET7Pqdp/+ZEf1Gd7zq6fEUECKyTsCyL9vZ2jI+Po6amBhkZGTEfL9aIJByx1lWMGiOMGiMKTAXwMl6Qi+VDnoKIhN8wSFGUJCICpIdpYyTOnz+PvXv34otf/CK+973vXVYiAihQSPgXYHR0FE1NTVi+fDlWrlwpfD1WG3kgOiHxen1wu12XZVE9EryQ0DSN5uZmuN1ubN++fbZ7LXakEpJgwtVVXC4XNBpKiFZIkpA8H5/KtwqR5FW7LMfC6XBKKiJAekckXV1duOWWW/A3f/M3+PGPf3zZiQigQCHhhwy7urqwfv36eVOf8QiJwRD5w+Yvqnsv26J6JGja3ylXX1/vd/itqYlpzXEwiaS2oiWWukrak8TUFp/21Wg0cd9IhCNdi+09PT245ZZbsH//fvz85z9fHO+pOFCckPT29qK3txfbt29HVlbWvO9TFCXsOo72RQv3nuc4btYPyL858XIuqodjZsaJ9vaPkZ2djXXr1iX8QUn25yyWukr8pO49k6yIRE4RAdJTSAYHB3HzzTfjxhtvxC9+8YvLVkQABQrJsmXL5lmRixGnL6J94UJF4OKi+pwfkCoiYhiGwdmzTdi0aQmqqqokCdk1mlRedGOrq6QFSYhI/K4ODtFArvSk2xzJpUuX8Fd/9Ve4+uqr8etf//qyFhFAgUKi0WjCiggAYcczwzBRp1iC3/tzRXXx3ZUqImJ8PhputwulpSuwcmXiO515lPR5i6WuolTkjkgYloHT4YROp4v4uUyUdIpIRkdHcfPNN2P79u144oknhGvS5YzihGQhCIIASZIxdW7pdP4LGMvOFdX1esOstYoqIMF4vV54PB4YjUYUFBRDyudISUIiJpG6iklrQrYhG3avHTQbo9Nlosi4s51h/DdccosIkD7F9vHxcezduxcbNmzA008/rYrILIoTkmjSJ/EU3PV6wGbzF9WNRhO0WrWoHgq32yNY5Pst+yNvrIyVdPjcxVpX0VE6LDEXAeYiuGk3ZmaXdnmY+UvBpEYuG3leRPR6/YJedlKQDkJis9mwb98+VFZW4o9//GNET7nLjbR8JiJtSQwFTdOgaRd8PgJmswUUpRbVQzHXeDC3zS6enSSRUGpEEo5QdRWa5usqnEhU/HUVg8YAg8aAAlM+fIwPMz47HF6/Fb4s7zgZIpJkiwig/NTW1NQU9u3bh6VLl+JPf/pT0p6XdCHNPtZ+YolI3G43PvzwQ2i1tNqZFQb/gJkTLMsFiAgQ+06ShUg3IQmG71rKyLDMro8l4PF4MDMzA6fTCa/XB5b1v7+0lBa5hhwsy1yGqtyVWGopRqYuY8FtgbEgdURCM7RfRAzJExFA2cX2mZkZ3HbbbcjNzcWBAwfiTvMRBBHxn+9///sxH/OJJ57Anj17kJOTg5ycHFx77bX48MMPw/48byD52GOPxfU3hEO5r14EohWSyclJ1NfXo6CgAEuWZKOrC1BrIoH4u9dcoKjQVhdSRyTpkNqKlmjqKhqNBhRFgiJIZOozkanPBAdOWDFs99rhS6SuIqGQ8JGWwWBIaFYoHpQakTgcDvz1X/81jEYjDh48mFDX2vDwsPDvf/rTn/Dwww+jo6ND+Bq/CK6vrw/Lly+P6phHjx7FZz/7WezatQsGgwGPPvoorr/+erS2tqKkpCTgZw8ePIjTp0/H5I0XLYoTkqjWcmo0Cxbbh4eH0dLSgpUrV2L58uWzQ4nK7b5JBf4UhivIEiYQmpa2LXoxCYmYWOoqBAiYtWaYtWYUzdZVeHNJNx1jXUWi1FYqRQRQppC4XC5ho+Ff/vKX2YV28SPe4JqVlQWCIEJudS0vL8fVV1+Ne+65B5/61KciGmI+88wzAf/93//93zhw4ADeeecdfO5znxO+Pjg4iPvvvx+HDx/GzTffnNDfEQpFJhoWEpNIEQnHcejs7ERrays2bdqEsrIysCwr6d72xYB/WZcTer0urIj4f07ax0331FY08HUVk8mIjIwMYRmay+XCzMwMXC43fD5asFYxaAzIN+VjRdYKVOZUoMhcCLPWBCKKGx8pUls+2geXywWj0ZgSEQGUV2x3u9347Gc/C7fbjVdffVWIFpJBW1sbtm/fju9+97soKirCvffei/feey+qlLzT6YTP50Nubq7wNZZlcffdd+Nb3/oW1q1bJ8s5p+XHOpyQMAyDxsZGDA0NYceOHSgoKADDMOA4ThUSEV6vb/bu07hgHlxqIVmsEUkkYqqrkFrkCHWVKiy1LEWmPgNUuLpKghGJz+eD2+WG0WhMaReSkmokHo8Hd999N6xWKw4dOhTSYUNOqqur8ZOf/AQ9PT145ZVXwHEc9u7di8rKSnz/+99Hd3d32N998MEHsXTpUlx77bXC1x599FFoNBp8/etfl+2clfPqxYC/LTXwCseveyVJEjt37oRWqwXDMIKViiokfjwej7CsK5oeeOm7ti7vGlXsdZUMZOozwAFw+ZyY8c4E1FUSiUj4x021iADKSW35fD58/vOfx+DgIN555x3k5OSk7FwIgsDVV1+Nq6++Go8//ji+8pWv4Ac/+AEaGhrw0ksvzfv5Rx55BM899xyOHj0q1HI+/vhjPP744zh79qysZpKKFJJo9raLI5KpqSmcPXsWeXl5QujGf58kSRAEAaPx8r6AAXx7LzOvMysSampLPmKrq/gHH01aE4rMRfDQHsz47KjIWYFBDMb82F6fFx63B0aTERoq9ZcBJdjI0zSNL33pS+jq6sKRI0di3rUjB2fPnsX/9//9f/jf//1fEASBBx54AF/60pfm/dy//du/4ZFHHsHbb7+NjRs3Cl8/fvw4RkdHA4r3DMPgm9/8Jh577DH09PRIcp6pfwfFAb+XBPB73jQ3N6OyshIrVqwAy7KCCIkvlpdzRMKbU3IcJ/IViw41tZUcYp1X0Wv00Gv0+LfrfgZ97ghODJzAif4TqL9UDx/ri/hYvHNBtFFpMtCRqY1IGIbBV77yFbS0tODIkSMoKChI2bkMDAzgj3/8I/7whz+gq6sLe/fuxe9//3vceOONISPHn/3sZ/jxj3+Mw4cPo6amJuB7d999d0CaCwBuuOEG3H333fjCF74g2TmnpZBQFAWfz4euri5cvHgRGzduRGFhYUAqK/hiKeHqhLSCZec22ZlMppjNCBlG2nBYjUiigxcNgyGyD5hGAxSaC3F79e24vfp2OLwOnB46jfcH3sepwVOY8cwEHFeJIgKkNiJhGAb3338/PvzwQxw9ejRkJ1UyKSsrQ01NDe677z589rOfjZhee/TRR/Hwww/j2WefxYoVK3Dp0iUA/lZii8WCvLy8eZGVVqvFkiVLUF1dLdk5p6WQEASByclJTE1NYceOHcjIyIgoIsDCO0kWIwzDwuVKzPpbjUhST6S6SkNDK1yuLBQUFMBiscCsM+OaFdfgmhXXgGEZNIw0CNFKz0SPsHdHSSICpK5ri2VZPPDAAzh27BiOHDkyb/YiFbS2tmL16tVR/exvfvMbeL1e/PVf/3XA17/3ve/FNeAYL4oUkkipF4/Hg76+PjAMgz179swrqof73csttUXTDFwuJ3Q6PfT6+NMGqpAoi+C6SmnpUszMXEJ3dzf0ej0KCgpQUFCA7OxsUCSFbcXbsHXJVtycczPOdJ3BdO40zoyeQYe1Q1EOD6kotrMsiwcffBCHDx/G0aNHUVZWlrTH/vznP4/Pf/7zIb8XrYgAiKvGIVVdRIwihSQc09PTOHv2rNBlwosIgIgiAlxeQsLfrUoxXKYW25ULQRBYurQIeXlFYBgGVqsVY2NjaG5uBsuyyM/PR0FBASYnJzEyMoJbr7gVZrMZf4+/x5hzDCcHTuJ4/3GcvXQWXkZac85YSXb7L8uy+Nd//Ve89NJLOHr0KCoqKpL6+IuNtBGSkZERNDU1oaKiAhaLBR0dHaBpWrCVXwi9Xjl3X3Li8Xjh9XpgNJoS3PrnR41IlA3/1qcoSohGOI7D1NQUxsbGcO7cOdA0jezsbFitVlAUBYPBgAJTAfat2od9q/bB6XPiw6EPcbz/OE4NnsK0Zzrpf0cyIxKO4/CjH/0I//u//4sjR45g5cqVSXvsxYoihUQcWXAch+7ubnR1dWHDhg0oKiqCw+GA1+vFyZMnUVRUhMLCQsFyIByXQ7Hd7XaDpmmYTGZQlDS3/qqQKJtQzydBEMjKysLQ0BC0Wi02bdqE6elpXLp0CR0dHcjIyBBEx2KxwKQ14aqyq3BV2VVgORaNo4040e+vqwzOxN5aHCsEQSQtIuE4Do8++ih+//vf491338WaNWuS8riLHUUKCQ/LsmhpacHExAS2b9+OzMxMMAwDnU6HPXv2wGazYXR0FPX19aAoCoWFhSgsLER2dva8KGXVKg7XXsvi5EkSTmeK/iCZ4DjA5XKC4ziYTGZJN/pd7jbySifU88lxHFpbWzE1NYWamhoYDAbk5uZixYoV8Hq9GB8fx9jYGLq7u6HT6VBYWCjUVUiSxJaiLdhStAX319yP7sluHO8/jvcH3kfbeJssdZVkFdo5jsN//ud/4le/+hXeeecdbNiwISmPezlAcEqquM3CMAwcDgfq6+vBcRy2bNkCnU4HlmVDFtVZloXNZsPIyAjGxsbAcZwgKrm5uQGi4vUCZ86QeO89EsePU5iZCXUG6QPHccLueaPRKPn06uc+x+Dv/k66sOSNN0j86EepHz5bLLz5pgdiTz/+5stut2Pbtm0RLc/FdZWxsTGhrlJYWIi8vLx5MwsTrgm83/8+TgycwEfDH0lWV7HoLHjjM29IcqxwcByHX/7yl3jkkUfw5ptvora2VtbHu9xQZEQyMzODDz74ANnZ2Vi/fj1IkoxYVCdJUuiX5jhOKC62tbWBYRgUFBQIHw6djsLu3Sx272bBMDTq60kcOULi2DEKNlsq/tr48VvAO0FRGhiN8nQTqDbyykb8fLIsi+bmZjidTtTU1CzooxaurtLV1YXm5mbk5uYK3zcYDMgz5uHWVbfi1lW3wk278eHQhzjRfwInB09i0j0Z998gd0TCcRyeeOIJ/PSnP8Xrr7+uiogMKDIisdlsGB4eRnl5ecCkOr8AJlo4jsP09DRGRkYwOjoKr9cr3HHl5+cH3HGxLNDURODYMQrvvUdhZETyP0tSkrVP+9OfZnD//dJFJO+8Q+J731MjEqk4csQDrdb/fmhqaoLH48G2bdsS7tZzOp0YGxvD6OgopqamYLFYhBSYxWIJzAhwLJpHm4V5lf7p/pgeq9BciBc/9WJC5xsOjuPw9NNP49vf/jZeffVVXHnllbI8zuWOIoWEZVl4vV6wLAuGYRZs7Y0GjuNgt9sFUXG5XMjLyxM+HMEfvHPnCBw9SuG990gMDChrj4nPR8PtdkGvN0Cnk/ei/KlPMfjHf5ROSI4eJfHd76pCIhXHjnnAcQwaGhrAMAy2bNkiuRW8uK4yMTEBrVYrRPmh6pG9U7040X8Cx/uPo228DSwX2ViyJKMEf7rtT5KeM+D/zD/zzDP45je/iVdeeQVXX3215I+h4keRQnLhwgX/lK7ZLImIhMJut2N0dBSjo6Ow2+3Izc0V6irBKYGuLgLvvUfi6FEKFy+mVlR4i4tkObbu28fgW9+STkiOHSPxne+oQiIVR4860NDQINQS5X5PxFpXsbqsODV4Csf7j+Oj4Y/gpt3zjlmeXY4/3PoHSc+T4zg8//zz+NrXvoYDBw7ghhtukPT4KoEoUki+9KUv4ZlnnsE111yDffv24eabb16wvTcRnE6nICrT09PIzs5GUVGRkBsW099P4OhREu+9R6G9Pbmi4nb7FyKZTMakWVzccguDb39bOiF5/30SDz6oCokUEASHf//3E6AoCps3b0667QmfOh4dHcXY2BicTue8uooYD+3BmeEzONF/Au8PvA+b21+UXJm7Ek/e8qSk53bw4EH83d/9Hf70pz/hlltukfTYKvNRpJBwHIeWlha88MILOHjwIDo6OnDVVVdh//79uOWWW5CbmyubqLjdbkFUJicnkZmZKcyqBO80Hx6GUFNpaSHASrc+ex5+C3hWWIyULG66icW//EtkN9lYOH2axD/9kyokicJxHDweB373u2Zs2rRJEd5ZsdZVWsdacWLgBKwuK/5l979Idh6vvvoqvvCFL+CPf/wjbrvtNsmOqxIeRQqJGI7j0NHRgQMHDuDFF19EU1MT9uzZg/3792Pv3r0oLCyUTVQ8Ho/wwbBarbBYLIKoBO9vnpjgRYVEfT0pWbeTv73XBQAwmaRv712I665j8b3vSSckH35I4oEHVCFJBL7lW6cDTp0io94tk0xiratIxRtvvIG7774bTz75pLBvXUV+FC8kYjiOw8WLFwVR+eijj7Br1y7s27cPt956K5YuXSrbhdbn82FsbAwjIyOwWq0wGo2CqATfbU1NASdOUDh6lMRHH5HwxXkd9rf3ukBR5LxoKFlcfTWLH/5QOiH56CMS/+f/qEISLxzHweFwgqJI5OQY8fbbnlSf0oKEq6sUFBTM655MhHfffRef+cxn8Nvf/hZ33XVX0m+6LmfSSkjEcByHvr4+vPjii3jxxRdx6tQp1NbW4tZbb8X+/fuxfPly2d5INE1jfHwcIyMjGB8fh16vF0QlMzMz4HEdDuDkSX+h/oMPSLjn1xpD4m/vdUGr1cJgkK+9dyH27GHx059KJyT19QTuv18Za1XTDX63DEVRMBoNMJuBw4eVLyRiYq2rRMuxY8dwxx134Be/+AU+//nPqyKSZNJWSMRwHIehoSEcPHgQL774Io4fP46NGzdi//792LdvHyorK2V7YzEMg4mJCUFUNBpNgFWL+HHdbn+N4NgxCu+/T8LhCH1MfjOeXq9fcKhMburqWPz859IJSWMjgfvuU4UkVvwi4ggYPs3IAF5/Pb2EJJhY6irhOHnyJG6//Xb87Gc/w9///d+rIpICFoWQiOE4DqOjo3jppZfw4osv4siRI1i9ejX279+P/fv3o7q6WrY3GsuymJiYEO62CIIQRCUnJycgL+zz+dM8R4/6rVqmZw1XvV4fPB43DAYjtNrUGw9s387iP/5DOiFpbibwla+oQhILvINB8IKy7Gzg1VfTW0jExFNXOXPmDPbt24cf/vCH+NrXvqaKSIpYdEIihuM42Gw2vPzyyzhw4ADefvttVFRUYN++fbjtttuwdu1a2Yp+vP8X3wHGcVyAVYv4cVkWOHuWwIsv2nD8uBY0naWILhwA2LaNxeOPSyck584R+PKXVSGJljkRmZ/izM0FXnll8QiJmEh1FYPBgOzsbNTX1+OWW27Bd7/7XTzwwAOqiKSQRS0kwUxOTuIvf/kLXnzxRRw+fBglJSVC+mvz5s2yiQrv/8WLCk3TyM/PR1FREfLy8kAQhODWunnzFly8aMF77/nbimdXMKeMTZs4/OpX0i096ugg8MUvqkISDSzLwuFwQqfThrTBKSgADh5cnEIihq+rjI2NoaWlBffccw82btyI/v5+3HvvvfjpT3+qikiKuayERMzMzAwOHTqEAwcO4PXXX0d+fr5QqK+trZVVVPhi4+joKNxuNyiKgkajwZYtW+a1Fbe3E7OiQqKvL/kflnXrOPzXf0knJJ2dBL7wBVVIFoJh2NkW39AiAgBFRcCBA4tfSII5dOgQfvKTn8Bms6G/vx+bNm3Cvn378I1vfANZWVmpPr3LkstWSMQ4nU688cYbOHDgAF577TVkZGRg79692L9/P+rq6mRLMzmdTpw9exaA38HY5XIhNzdXmKoP9kzq7vZP1R89SqGrKzmiUl3N4fe/l05ILl4k8LnPqUISiTlDTj30+vDPVXExh+efT+2K3GTT0dGBm266Cffeey9+/OMfw2q14rXXXsOhQ4fw5JNPpqxN/nJHFZIg3G433nrrLbz44ot4+eWXodPpsHfvXtx2223YvXu3ZIZ4MzMzOHv2LAoLC1FdXQ2SJOFwODA6OoqRkRHY7Xbk5OQIohJ8VzowMOf/1d5OQK5XsbKSw9NPS3ex6u0lcNddqpCEgxeRaDr2Sks5PPfc5SMkXV1duPHGG/GZz3wGP//5zxU5iHm5ogpJBLxeL44cOYIDBw7gpZdeAsuyuOWWW3DbbbfhE5/4RNytuRMTE2hqasKKFSuwYsWKkPldl8sliMr09DSysrKEWZXgXvuRkTmrlqYmaa1ayss5/OEP0l2s+vsJfPazqpCEgqYZuFzOqF2dly/n8Oyzl4eQ9PT04KabbsKtt96Kxx9/XBURhaEKSZTQNI1jx47hhRdewEsvvQSXy4Wbb74Z+/fvxyc/+cmoB6mGhoZw7tw5rF27FsXFxVH9jtvtFqbqef8vvq3YJF6PB8BqBY4f99dUPv44cauWZcs4/O//SnexGhoicOedqpAEw4uIwWCIOuqVWuSVysDAAG644QbccMMN+PWvf62KiAJRhSQOGIbB+++/jwMHDuDgwYOYmprCjTfeiH379uH666+fd3EH5uxd+vr6sGnTJuTm5sb12F6vVyjU8/5fvKhYLJaAn52e9rvtHj1K4cwZEt44rjlS5+EvXQL++q9TN6mvRPgB1FhEBJA+7ahEhoeHceONN2LPnj144oknFNMWrxKIKiQJwrIsPvzwQ8GpeGRkBNdddx3279+PG2+8ERkZGfB4PHj22WdRWVmJrVu3zrvgxwvv/zU6OoqJiQkYjUYUFhaiqKho3lSw0+m3annvPQqnT5NwuaJ7DKlbTEdHgdtvV4WEZ05EYh9Arari8NRTi1dIRkZGcNNNN6GmpgZPP/20KiIKRhUSCWFZFvX19XjhhRfw4osvoq+vD1dddRWGh4dht9tx/Phx2doTef+v0dFRjI+PQ6fTCaIS7P/l8QAffOAXlZMnSczMhD9uTg7wl79IJyQTE8C+faqQAHObLuNdUiZ1R52SGB8fx1/91V9h3bp1eOaZZ5KyxE0lflQhkQmO43DkyBHcddddYFkWbrcbdXV12L9/P26++WZZd6rw/l+8VQtFUQFWLeLHpWm/Vct77/mtWiYnA48ltZ+TzQbs3asKic/ng9vtTmjT5Zo1HJ54YvEJidVqxc0334yKigr8+c9/lnx1sIr0qEIiE+fOncNNN92ET3ziE/jd736H7u5uHDhwAAcOHEBzczOuvPJKYadKQUGBrP5fVqtVqKsQBIGCggIUFRXN8/9iWb+p4pEjFI4fpzA2BphMwJtvSick09PAX/3V5S0kcyJigkYTf7pm/XoOv/3t4hKSqakp7N27F0uWLMGBAwfCDmOqKAtVSGSCX8b10EMPBYgEx3Ho6uoSdqqcPXtWiFRuvfVWFBcXyyoqYqsWhmEEUcnNzQ3IQXMc0NpK4ORJCn/3d9Kt2nU4gBtuuHwvDrwpp8lkSjjnv3Ejh1//evEIyczMDPbt24esrCy8/PLLcVvKqyQfVUhSCL9ThReV06dPY/v27YJVy7Jly2QTFY7jMDU1Jcyq+Hw+wf8rPz9ftsKmywVcd93lKSRerxcej0cSEQGAzZs5/PKXi0NIHA4Hbr/9dmi1Wrz66qshOx9VlIsqJAqB36nCL+o6ceIENm3aJJhKVlRUyCoqMzMzgqi43W7k5+ejsLAQ+fn5kuaovV7gk5+8/ITE4/HC65VORADpnZlThdPpxB133AGGYXDo0CHJuhpVkocqJAqE4ziMjIwIO1WOHj2KNWvWCKIi504V/ypXB0ZGRjA6OgqHw4G8vDxh2VCii7YYBvjEJy4vIfF4PPB6fbMiIt0wXW0ti//8z/QWErfbjU9/+tOw2+04fPgwMjMzU31KKnGgConC4TgOVqs1YKdKVVUV9u3bh/3798u6UwXw3y3yojIzM4OcnByhAyyeQijHAXv2XD5CwouI2WyS/HXauZPFv/1b+gqJx+PBXXfdhbGxMbz55pvIyclJ9SmpxIkqJGkEX9d45ZVX8OKLL+LNN99EaWmpICqbNm2SVVR4/y9+LWpWVpYgKrG4ru7Zo5fNZFJJuN0e0LQ/EpHjddm1i8XPfpaeQuLz+fC5z30Ovb29eOedd5CXl5fqU1JJAFVI0piZmRm89tprwk6VgoIC3HrrrbjttttQU1Mjq6h4PB5BVGw2GzIyMgRRCd6pEsxVV+lBS9cIpkjcbjdompZNRADgiitYPPJI+gkJTdO499570d7ejiNHjqCgoCDVp6SSIKqQLBIcDkfATpWsrCxhp8rOnTtltZfwer0BVi1ms1mYqjebzfPqOUeO+P2/Tp0i4XTKdlopw+Vyg2FomExmkKR8e2M+8QkWP/5xegkJwzD4+7//ezQ0NODdd9/FkiVLUn1KKhKgCskixOVyCTtVXnnlFej1+oCdKnLaTfh8vgCrFoPBIIhKRkZGgKh4vcCHH/qtWk6ciGzVki64XC4wDDsbici7fOzqq1n88IfpIyQMw+D+++/HyZMnceTIEZSUlKT6lFQkQvVjXoQYjUbceuuteOqpp3Dp0iU8+eST4DgOn/vc51BZWYn77rsPb731Frzx2AEvgFarRXFxMTZt2oSrrroKVVVVcLvd+Oijj3DixAl0dHRgcnISHMdBp/OnZ/7lX3z4y188+I//8GHfPgbpWnN1uVxgWXa2sC7/Bst0clNnWRb/+I//iGPHjuHtt98WROSnP/0pamtrhdTo/v370dHREfC7v/vd73DVVVcJnnGTwT4+MXDVVVeBIIiw/1x11VUxH/Po0aPYt28fiouLYTabsXnzZjzzzDNhf/65554DQRDYv39/3H+H0lAjkssIfqfK888/j5deeglutxu33HIL9u3bF9NOlXhgGCbAqkXs/5WdnR3SquW99ygcO0ZhdFS205IMp9MFjvNHInK1Zgdz/fUsHn5Y+REJy7L453/+Z7z22ms4cuQIKioqhO/xGw9ra2tB0zS+853voKWlBW1tbUKt7bHHHoPb7QYAPPTQQ7DZbMjOzo7pHPr6+rB8+XJYrVbhBqq/vx/bt2/H22+/jXXr1gEAdDpdVCse+OMBwE9+8hO4XC7cdNNNKCoqwquvvooHHngAL7/8Mm655ZaA3+vp6cEVV1yBiooK5Obm4qWXXorp71AqqpBcpvA7VXj7++npadx0003Yt28frrvuOlkni1mWhc1mw8jICMbGxsBxnCAqubm5AaLCccC5cwSOHvUv6xocTM5FOhacTic4jkuqiADAjTey+O53lS0kLMviu9/9Ll544QUcOXIEK1eujPjzY2NjKCwsxHvvvYcrr7wy4HtHjx7F1VdfHbWQjI2N4dlnn8XTTz8NjuNQX18f8P2enh6Ul5ejvr4emzdvXvB409PTeP755/H000+jpaUFVqs17M/efPPNKCoqwv/8z/8IX2MYBldeeSXuvfdeHD9+HJOTk4tGSNIoOFaREoqicOWVV+IXv/gFent78cYbb6CkpAT/8i//ghUrVuBv//Zv8cILL8But0v+2CRJIi8vD2vXrsWVV14ptC23tbXhvffeQ0tLi+AFRhDA2rUcvvpVGn/6kxdPPeXFF77AoLxcGfc/fhEBTKb5TQVyQ1HKeA7CwXEcfvjDH+K5557D22+/vaCIAH7TRuD/b+/c43I+/z/+vO8KTUeSMsekrGI05+M2piIq5hBLNG2GOc2+2MNsfDHMYWz4WnOIHzYqzHGOlTMhpxQ55dBB9W0d6HDf9/X7o/X5amEO3QrX8/Ho8XB/jteVu8/rc13v6/1681yF3zZu3Ii3tzdvvvkmy5cvZ8CAAWzfvv2ZrqfT6di9ezcfffQRNjY2zJw5k06dOhEVFfXY8/78888SfZg6dSrW1tZ8/PHHz9SW8ow0+ZegVqtp3bo1rVu35vvvv+fUqVOEhIQwbdo0hg4dSqdOnfD29qZr164laps8LyqVCktLSywtLXF0dCQzM5Pk5GQuXbpEfn5+MasWQ0ND7O0F9vYaPv4YEhJU7N+vJjLSgLi4F/sQF6JQRFQq/hqJvNDbA+U7RiKEYObMmSxfvpz9+/fTsGHDfzxHp9MxevRo2rZti4uLy1Pd79SpU6xcuZK1a9dSoUIF+vfvT1RUFI0bN36m9l+6dImVK1eyevVqcnJy6NOnD3v27KFNmzb/eO769es5ceIES5cuVbYdPHiQZcuWER0d/UztKe9IIZEUQ61W06xZM5o1a8aMGTM4f/48ISEhzJ8/n2HDhvH+++/j5eWFp6dnidomz4tKpcLc3Bxzc3MaNGhAdnY2ycnJXL16lQsXLhSzajEyMqJ2bYG/vxZ/fy2JiSoiIgqXFV+4oNJrwqMQgnv37isiUlaUVyERQjB//nwWL17M3r17n1gUhg8fzvnz5zl48OBT37Nnz57cunWLr7/+mkmTJj33cvdPPvmEiIgIhg4dyrx585444Xb//v0MHjyYoKAgJe6SlZWFn58fQUFBWFlZPVe7yisyRiJ5IoQQxMbGKtUfz58/T8eOHfHy8tJ7TRWA7OxsJVCfnZ1NlSpVlLjK3/2/UlMhMrIwpnL6tBqdrvTaUSgi91Cp1LzxxpNn8+sDHx8tX3xRvjI7hRD89NNPzJo1iz/++IPmzZs/0XkjRoxg8+bNREZGUq9evYce87gYyYEDB1i+fDkhISHY2tri5+fHRx999MhrweNjJNHR0Sxbtox169ZhbGzMgAED8PPzU8ThYURERNCtWzfmzZvHJ598UuxaTZs2LSZuur++lGq1mri4OOrXr//I674MSCGRPDVCCOLj4xX7+9OnT9OmTRu8vLz0XlMFCqeUikQlMzMTCwsLRVT+vvLszz/h4EED9u9Xc/KkmoLniE0XiYharX4qSxh90auXljFjyo+QCCH4+eefmTJlCjt27KB169ZPdM7nn3/Oxo0bCQ8Pf2wc5UmC7Tk5OYSGhhIcHExERARt2rTBz8+P3r17lzjnSYLt+fn5bN26leDgYHbs2IGLiwt+fn74+voWS6YMDw/H09OTWbNmMXz48GLXyM3NJT4+vti2SZMmkZWVxYIFC3BwcHhuM9SyRgqJ5LkQQnDjxg1FVI4dO0aLFi3w8vLCy8tLrzVVoPCPtEhUMjIyMDMzUxIg//6wz86GQ4cKYyrHjqn5a0XpE1HoinwPA4PyISIAffpoGTmyfAiJEIKVK1cyceJEtm7dWmLF1aMYNmwYa9euZfPmzTg6Oirbzc3Nld9zUlISSUlJREVFERgYSGRkJKamptSuXfuxQfmEhARWrVrFqlWrMDU15eTJk8X2P+2qrbt377JmzRqCg4NJSEggLS0NKJzO8vT0ZNSoUYwcOVI5/nFLiQcNGvREq7aEEC98EcezIIVEUmoIIbh9+7ZSU+XQoUM0adJEsb+vV6+eXv8o8vPzFVFJT0/HxMSE6tWrP9T/KzcXjh79n1VLTs6jr6vTFY5EDAwMMDYuP1X7+vXTMmJE2QuJEIL/+7//Y9y4cfz++++89957T3zuo74PK1asYNCgQQB8++23TJky5bHH/BOxsbElAv5PKySPut6gQYMIDg4ucUzHjh0JDw9/6PlPIiRarRYDAwMKCgrIysp65pVsLwIpJBK9UFRTZePGjUpNFWdnZ0VUHBwc9CoqBQUF3L17l+TkZNLT0zE2NlZExcTEpIRVS1SUmoiIwtHKg1YthSKSg6GhYbkr/dq/v5Zhw8pWSIQQrF+/ns8//5zQ0FDc3NzKtD2vChqNBkNDQzIzM/n888+pXLkygwcPfuKY04tGColE7wghSEtLU2qq7N27lwYNGij292+99ZZenYo1Gg2pqakkJyeTmppKxYoVFVH5+3JmrRZOn1YTHq5m/364fft+uRQRAD8/LZ9+WrZCEhYWxqeffsr69evp1q1bmbblVUGn06FWq8nOzqZt27bk5eXRr18/vvrqq3IbS5FCoifCw8MZO3YsFy5coFatWkyaNOmJh+GvMg/WVAkNDWXXrl3Url1bsb9v3LixXkVFq9WSlpamiIqhoWExq5YiUcnJyeH48Siys+24dq0ukZGGJCXprVnPhL+/lsDAshOSrVu3MnjwYNasWfNK+UaVBzQaDW5ubqSkpPDDDz/QoUOHh5a8LhKdskYKiR64du0aLi4uDB06lCFDhrB3715Gjx7Ntm3b5ND/b2RmZio1VXbu3Im1tbUiKu+8845e/0h0Oh1paWmkpKRw9+5dVCoV1tbWmJqaEh8fT82aNalfv74iLhcvFvp/RUSouXmz7AOgAQEaAgK0ZXLvnTt34ufnx8qVK+ndu3eZtOFV5uDBg/Tt25dvvvmGwMBAVCoVSUlJ3LhxgwMHDlClShUCAgKA8hGQl0LyFzt37mTatGmcP38eAwMDWrduzYIFC55pfff48ePZtm0b58+fV7b169ePjIwMdu7cWZrNfqXIyclhx44dhIaGsn37dszNzenRowfe3t60bNlSrzVVdDodGRkZ3Lx5k5SUFNRqNTY2NlhbW1O1atUSgnb1qorw8EIL/CtXyuaPODBQg7//ixeSffv20a9fP5YuXUr//v3L/CH2KnLgwAG6dOnCwoULCQwMJDw8nHnz5imu3UII+vXrx9q1a8u6qYD02lLIyclh7NixREVFsXfvXtRqNT4+PkrikLOzMyYmJo/88fDwUK515MgROnfuXOz6bm5uHDly5IX26WWjcuXKfPjhh6xbt46kpCR+/PFHsrKy6NOnD46OjooNuUYP5RXVajUGBgakp6djb2/PO++8g6GhIbGxsURERHDu3DmSk5PRagsf3HZ2goAALcHB+axbl8/QoRoaNnyx72Rl8fyOjIzE19eXH3/8UYpIKfGwd/lq1arh6OjIv//9b9q1a0eXLl1IS0tjxowZJCYmMnToUP74449/9Px6UUiLlL/o1atXsc/Lly+nWrVqxMTE4OLiwvbt2yl4TDbbg7kFSUlJVK9evdj+6tWrk5mZyf3798tNHkJ5xtjYWMlFyc/PZ8+ePYSFheHn54dKpcLT0xNvb286dOhQKgHIjIwMTp8+Tf369RV7cAsLCxwcHMjMzCQlJYX4+HjOnz+v+H9Vq1YNQ0NDatUSfPSRlo8+0pKUhGJ/f+6cqlSz6v+OHgdoD+Xw4cP06dOHuXPnMmjQICkipUDREt+ifwshMDAwoGHDhsydO5fly5dz7do1vvvuO9zc3BS7mbfeegudToepqWlZNl9BCslfXL58mcmTJ3Ps2DFSU1OVkUhCQgIuLi7UqVOnjFv4+lKhQgW6du1K165dWbJkiVJT5ZNPPiE/P79YTZWKFSs+9fXT09OJjo7GwcGBmjVrFtv3oP+Xvb29YtVy/fp1xf+revXqiv+XjQ307aulb18taWlw4IAB4eGFVi3aUp6FepEx1uPHj/Phhx8yY8YMZc5e8nw8KCJTpkzh9OnTZGVlUbduXaZMmUKnTp3o0KEDGo2m2MvnpUuXOHz4MA0aNJBCUt7o3r07derUISgoiBo1aqDT6XBxcVGK4Dg7O3Pjxo1Hnt++fXt27NgBgI2NDcnJycX2JycnY2ZmJkcjz4mRkRGdOnWiU6dOLFq0iIMHDxISEsLo0aPJysrC3d0db29vOnfu/ESGiqmpqZw9e5aGDRtSo0aNxx6rUqkwNTXF1NSU+vXrk5OTQ0pKCgkJCcTExGBpaamISsWKFalaFby9tXh7a8nMLLRqCQ9XExWlpjSKU74oITl16hQ+Pj5MnjyZ4cOHSxEpBYpGHgAeHh4cPXqUJk2aoFar+eOPP9i8eTO//fYb77//frHVWpGRkSxdupRNmzYRGhr6j9/ZF4UUEiAtLY24uDiCgoJo3749QAkH0qeZ2mrdunWJ+ge7d+9+Iu8hyZNjYGBAx44d6dixIwsWLODo0aOEhITw1VdfERgYSJcuXfD29sbNzQ0TE5MS59+9e5ezZ8/i5OSEra3tU9+/cuXK1KtXj3r16nH//n1SUlK4c+cOsbGxmJubK7kqlSpVwswMunbV0rWrlnv34PDhwqz6Y8fU3L//rP1/tvOehrNnz+Ll5cX48eMZM2aMFJFSouj3OHnyZKKjo/nll1/w8vLC0NCQ6dOn8/XXX3Py5Mliy37Hjx/P6tWrqVSpEuvWraNr165l2YViyFVbFK7Ysba2xsPDg2+++YaEhAQmTJjAiRMnlCI5T0PR8t/hw4cTEBDAvn37GDlypFz++4LQ6XScPHlSqf5469YtOnfujLe3Nx4eHpibm7Nq1Sru3buHj49PiXjW85Kbm6tk1WdkZGBqaqqIyt9HSXl5hVYtkZEGHDqk5mnqiI0dq6FnT/2t2oqJicHDw4MRI0YwefJkKSJ6oHv37hgaGrJixQosLCwIDw/H3d2dgIAAvvrqq2JTrUeOHGHnzp14enqWuwx3KSR/sWfPHkaOHMnVq1dxdHRk4cKFvPvuu88kJFCYkDhmzBhiYmKoWbMmX3/9tUxILAN0Oh1nz55VTCXj4+Np2bIlJ06cYO7cufj7+79Q/6/KlSsXs2p5kIICOHmyMKv+wAED/ioW+Ei+/FKDl5d+hCQuLg4PDw8+/vhjpk2bJkWkFCgoKCg2TZWZmUnjxo358MMPmTNnDhEREXh4eNC7d29mz56tvOD88MMP9O7dmzfffLNYXKU8IYVE8toghGDGjBlMnToVFxcXzp07V6ymipWV1Qvx/0pJSSEtLQ1jY2Mlq97U1LTYvXU6iI4uFJXISANSU0teb/x4Dd27l76QxMfH4+Hhga+vL7Nnzy4XmdMvM2lpaZiZmSkismXLFlq3bo2VlRU+Pj5oNBqGDBnCgAED8PLy4vvvv1diH/v378ff35+5c+eW68RP+Q2RvDYEBQUpBZeioqK4cOEC77//PsHBwdjb29OtWzeWLl1KYmLiQ9f2Py9GRkbUqFGDJk2a0LFjR+zs7Lh37x5RUVEcOnSIS5cukZGRgRACtRpcXXWMHath48Y8lizJp18/Lba2/2uXPp7v169fx9PTk169ekkRKQWOHz/OqFGj2LZtGwADBgzgX//6l7IYx8PDg23bttGzZ086d+7MDz/8oIjI7du3+fXXX7G0tHyiUsVliRyRSF4biqrntW3btth2IQTXr19Xpr+OHz9Oy5YtlTyWmjVr6nWkUuT/VWTVYmBgoIxUHlbOOC6u0KqlSRMdLVqUXqLKzZs3cXNzw93dncWLF0sRKQUuXbpE27ZtMTMzw87OjsOHD/Of//wHb29vZenuhAkTmD17NoGBgXz22Wc0adKEs2fP8ttvvzFnzhwWLVrEkCFDyrgnj0cKiUTyAEIIbt26Vaymiqurq2J/X7duXb2Kik6nIz09XYmrqFQqqlWrRvXq1bG0tNTbwz0xMRE3Nzc6dOhAUFBQuZyHf1lJSUmhXr16aDQa/P39mT17NhYWFopVfHJyMrNnz2b+/PlYWVlRo0YNkpOTKSgoYPz48Xz55Zdl3YV/RAqJRPIIhBAkJSUpNVUiIiJwcXFRRKVBgwZ6F5WMjAxFVLRarSIqVapUKbWHfXJyMh4eHjRv3pyVK1dKESlFhBAcPHiQjh07Ym5uTuXKlZkzZw49e/Ys4ciwceNG9u/fT3x8PK1ataJFixa4u7uXUcufDikkEskT8GBNlZCQEPbt24eDg0Oxmir6FJUi+/2UlBTlbdXKyorq1atjZWX1zA//1NRUunbtirOzM2vWrMHQUKaW6YPY2Fjy8vLo3bs32dnZzJgxA19fXypWrFhiJVZ5sYZ/GqSQSCRPiRCCjIwMpabK7t27qVOnjmJ/36hRI70+CIQQZGVlKaKSm5urWLVYWVk9tG7Fw0hPT6dbt27Y2dmxfv36Jz5P8mj+SQQSEhJwc3Pjv//9L9OnT6d///4YGxuTmZlJbGws1tbW1K1b98U1uJSQQiKRPCeZmZls3bpVqaliY2OjiIqrq6veRSUnJ4fk5GRSUlLIycmhSpUqilXLowwtMzIy6N69O7a2toSFhZXbynsvEw+KyO7duzl37hxZWVm0bNmSzp07K6O927dv4+7uTlJSEjNmzKBTp06Eh4czadIkFi9e/FIWCZNCIpGUItnZ2cVqqlhaWtKjRw+8vLz0XlMF4N69e4qoZGVlYWlpqawAKzK0zMzMxNvbG3NzczZv3lwuywi/bDxYXGru3LlMnjwZQ0ND7t+/j5WVFe+99x6//PKLYqWUkpKCu7s70dHR2NnZcefOHfr27cuKFSvKshvPjBQSiURP3L9/nz/++IOwsDC2bNmCsbGxUqirTZs2eo9H3L9/X7FquX79OvPmzeODDz7gyJEjVKxYkW3btkkT0VImODiY4cOHM2rUKPr27Uvjxo0ZPnw4S5cuxd3dnbVr12JmZgYUJqiOHj2a+/fv4+TkxLhx48q49c+OFBKJ5AWQm5vL3r17CQsLY/PmzajVarp3767UVNF3fOLu3bv88ssvrF+/nsuXL9O0aVN69+5Nr169aNCggV7v/bpw7do1BgwYQLNmzRg7dix169ZFp9PRuHFj0tPTycvL45133mHDhg2Ym5sr572MwfW/83K3XiJ5SahUqRLdunVj2bJlJCYmsnbtWoyMjAgMDMTOzo7PPvuMnTt3kpeXp5f7m5qacvToUapWrcrVq1cZMWIEBw8exMXFhVOnTunlnq86f6/UqVaruXPnDs2bN1cC5i1btgRg06ZN+Pn5sWfPHvr27cvdu3cB/nIxePkfw3JEIpGUIRqNRqmpsmnTJrKzs/Hw8FBqqpTG1FNeXh4DBgwgNTWVXbt2YWFhoezLzMzExMTklXiYvSgyMjIwNzdXYiKbN2/G0dGRhg0bcuPGDaUI3sCBAwkPD2fdunW0bduWrKwsGjduzL179zAzM+PEiRPF/i9eZuS3RyIpQwwNDXn33Xf56aefuHHjBtu2bcPGxoYJEyZQt25dBg4cSFhYGNlP4y//APn5+fj7+5OYmMjOnTtLPLjMzMykiDwFmZmZTJw4kQEDBgCFSYQ+Pj5s2bIFjUZDrVq1ALhy5QonT55k4MCBiuV7cnIymZmZdO7cmR49erwyIgJSSCSScoOBgQFt27Zl/vz5XLlyhb1792JnZ8eUKVOoW7cuvr6+/Prrr2RmZj7R9YpcZa9du8auXbuoUqWKnnvw6qNWq1Gr1YSEhNC6dWt69erFuHHj6Nu3L4aGhsVE+fr166jVaipUqEBubi5Hjx7l7bffZvbs2cydO7cMe1H6yKktiaSco9PpOHPmjGIqefXqVTp16kSPHj3w9PTEwsKiRFa9RqPh008/5cyZM+zfv7/Ui3e9jjy4xNfT05Pt27dTt25ddu3ahb29fbGgeUpKCoMGDSI6OprAwEAKCgpYuXIlrVq1IiwsrCy7oRekkEgkLxFCCGJiYggJCSEsLIyLFy/y7rvv4uXlhaenJ1ZWVuh0OkaMGMGRI0cIDw8vN3W9X3aKhCQ7O5v27dsjhCA+Pp6OHTuyYcMG3njjjWJ2J/v27WPu3LmEh4djZmZGu3bt2LBhQxn3Qj9IIZFIXlKEEFy+fFkRlTNnztCmTRs0Gg137twhIiKC2rVrl3UzXwmKRhsxMTEkJiaSlZVFkyZN+PHHH1m2bBnNmzcnJCQEc3Nz8vLylOTPtLQ0EhMTyc3NpVmzZmXcC/0hhUQieQUQQnDt2jVWr17NggULOHDgAM7OzmXdrFeCIhE5ePAgAwYMwNzcnKlTp+Lt7U1eXh6TJk0iKCgIV1dXfv31V6ytrcnNzeXs2bMAtGjRoox7oH+kkEgkEsk/cOrUKd59913c3d0ZNGgQXbt2Vfbl5eUxdepUFi1aRKNGjQgODubGjRsMHTqUN998k3379pVhy18MUkgkEonkMWRnZ+Pr60t2djYLFy6kUaNGyr5bt25hYWGBiYkJ33//PdOmTSMvLw9TU1Nq1KjBiRMnXgtDTFl84AUihFCGyfqsXSGRSEoPAwMDpdhUkYjs2LGDbdu2sWrVKqytrfH19eXf//43Dg4OHDt2DI1Gw9SpU18LEQE5InlhFBQUFPNTKvq1S0GRSMovQghSU1Pp0qULJiYmjBgxglOnTrFmzRqMjIzo0aMHZ8+eVbZ5enqWdZPLBDkieQHExsYSFBREYmIib7/9Nv3791cyYCUSSflFpVJRrVo1ZsyYQZ8+ffD39wdgzJgx+Pj40KJFC+Lj43FycuLKlStl3NqyQ2a26xkhBGfPniUyMpKCggKWLFmCk5MTEyZM+Mfznme/pDiJiYn0798fBwcH1Go1o0ePLusmSV4iPDw8OHnyJDt37uTy5ctMnz6dFi1aoNVqiYuLUwqJva7IEYmeUalU9OnTh969e5OUlERycjKLFi1iwYIFtGrV6pHV0IqmvLRabYmYyqtgO/2iycvLo1q1akyaNIn58+eXdXMkLyEODg44ODgU2xYZGcmcOXOwsLDAw8OjjFpW9sin0Qvgzz//RKVSYWtrS5MmTViwYAHVqlVj9+7dQGH8pIjU1FS2b99ObGwsUBjoe1BEEhMT+e677zh37pyyrWh0kpub+yK688LYuXMn7dq1w8LCgqpVq+Lp6fnM0wd169ZlwYIFDBw4sFgtCInkWcjJyWHYsGFMmDCBa9eusXXrViwtLcu6WWWGFBI9otVqWb58Of7+/jg5OeHr66tYKRgaGiKE4P79+xgZGZGVlcXnn3+Oo6Mj/v7+NG3aFDs7O+bNm0dGRoZyzSNHjvDtt99y8+bNEgH7SZMm0a5dO5KSksqiu6VOTk4OY8eOJSoqir1796JWq/Hx8UGn0wHg7OyMiYnJI39e5zdEiX65cOECmzZtok6dOkRGRirW8a8rcmpLj0RERDBs2DDefvtt3NzciI6OJiAggAEDBmBkZISNjQ3GxsZoNBpGjBjB6tWrCQwMpG3btlSoUIFDhw4xa9YsHB0d6dChAytWrCAsLAwHBwfat2+PSqVS/H+ysrK4cuUKFy9exMbGpqy7Xir06tWr2Ofly5dTrVo1YmJicHFxYfv27cVGc39HlpGV6IsWLVpw+vRpKleujImJSVk3p8yRQqJHpk+fzgcffMCqVauwtLQkJSUFQ0ND/Pz8iImJUSwstm7dytq1axk0aBBLly5Vzv/ggw+oU6cOtra25OTksGHDBg4dOoSxsTFDhw5lyJAhvPfee0BhBbzQ0FAuXboEUMw87mXl8uXLTJ48mWPHjpGamqqMRBISEnBxcXnt3wIlZYt0VP4fcmpLT2g0GhISEhBCKHOn1tbWVKlShYsXL1KrVi2l4M2yZcuoXr06Q4YMAQqLEQFUrVqVcePG4erqio2NDb6+vqjVajw9Pbl8+TKdOnViyZIlwP9qHzRs2BAhxEsvIgDdu3cnPT2doKAgjh07xrFjx4D//X7k1NbTo1KpHvvz7bffPvU1g4KCaN++PZaWllhaWtK5c2eOHz9e7JiwsDC6dOlC1apVUalUREdHl06HJOUCOSLRE4aGhvTs2ZOlS5eycOFCXF1dSU1NJTQ0lOvXr9O7d29q166NTqfj9OnTNGrUiMaNGwMo2bBCiGI1nY8ePYq9vT2TJk3CxsaG+Ph46tevT0ZGBr6+vgghOHr0qBIzeXB1l1arBXhpBCYtLY24uDjlIQVw8ODBYsfIqa2nJzExUfn3b7/9xuTJk4mLi1O2Pek0zc2bN5VcqPDwcHx9fWnTpg2VKlVi1qxZdOnShQsXLvDmm28ChfGudu3a0adPHwIDA0uxR5LygBQSPRIQEEB8fDxffvkltra2ODk5cfjwYZycnGjZsiVQOH1jZGSkvEVrNBoMDQv/W4reEgFu377Nrl278Pb2xt7eHmNjY6ysrFCpVISHh3P37l3FSK4obqJWq7l8+TINGjR4aQSkCEtLS6pWrcrPP/+Mra0tCQkJJXJvnnZqq+gtODs7m7t37xIdHU2FChVwcnIqrWaXex6MnxXVHX/SmFpCQgKrV68mODgYZ2dnNm7cCMCaNWuKHffLL78QGhrK3r17GThwIAB+fn5A4chZ8uohp7b0iKOjI6Ghody6dYsff/yRjRs30qBBA1QqFU2bNgWgfv362NjYkJ6eDqCIiBBCGUUAHD9+nJSUFN5///0Sb9qHDh0iPT2dDz/8ECgUoNDQULp27YqbmxtVq1alZ8+eytRQUayhiAfvU15Qq9X8+uuvnDx5EhcXF8aMGcP333//XNds2rQpTZs25eTJk6xdu5amTZsWc3GVlCQnJ4fVq1fTuXNn6tWrx/bt2/niiy9Yvnz5I8+5d+8eBQUFsrTv64SQvFD+/PNPcfz48WLbpk6dKlQqlfj222/FnTt3RH5+frH9Op1O+Pv7i/r164sLFy4IIYTQarVCCCEyMjJEjx49RMOGDZXjf//9d1GlShVRq1YtMXLkSDF16lTx1ltvCVdXV3Hp0qVi5/8djUZTan2VlG9WrFghzM3NH7ovPDxcDB48WJiYmAg7OzsxefJkER8f/0TX/eyzz4SdnZ24f/9+iX3Xrl0TgDh9+vRztFxS3pBCUg7Iz88X/fv3F+bm5sLd3V1MnDhRBAQEiJs3bwohhEhPTxe1atUSgYGBIiMjQwghREFBgRBCiIiICGFnZyeGDh0qhBDi0qVLolWrVsLGxkZcuXJFuf6ePXtEhQoVxGeffSY0Go0iGBMnThTbtm0r0SadTqf3fkvKlscJCSCMjY3Ff/7zn6e65nfffScsLS3FmTNnHrpfCsmriZzaKgcYGRmxcuVKli5dikqlYuvWrahUKmUKKz09nVu3bmFvb69kZRdNgR05coT09HR8fHwA2L17N2fPnmX06NHY2dkprsOdOnWiVatW7N+/HwMDAwwMDMjPz2fmzJksWLCALVu2MG7cOJYuXUp2drZ0JX7N2bJlC926dWPUqFG4uroyf/78f0x0nTNnDjNnzmTXrl3KwhHJa0JZK5mkJFqtVvz3v/9VPsfExIjatWuLzp07i6ioKJGSkiKEECIzM1P4+PgIBwcH5djBgwcLU1NTERcXp1yraPTh7e0tmjVrJpKTk4UQQmzevFmoVCrh6OgoHBwchLe3tzAzMxPOzs7i8uXLxdpUNELR6XSioKBAjlheAR43IikiPT1dLFq0SLRo0UIYGBgId3d3sWbNGpGTk1PsuFmzZgkzMzNx5MiRx15PjkheTeSIpByiVquxsLBQPjdo0IAJEyYQHR1N8+bNmTdvHlBo0xAbG6skJaanp5OWloalpSUODg7K8t+iFVuHDx+mdu3aymhj2bJlAHzxxRecOXOGZcuWsXDhQmJiYpR9RahUKu7du4dKpcLQ0FDJqpe82lhaWjJs2DCOHTvG+fPnefvtt/nXv/6lrMICmDVrFl9//TXLly+nbt26JCUlkZSURHZ2tnJMeno60dHRxMTEABAXF0d0dPQrY+fz2lPWSiZ5OjZt2iRCQ0OFEEJMmzZNqFQqsWPHDmV/nz59hLOzs8jNzS123q5du4RKpRIzZ84UQhSOVIyMjETPnj1FVlZWsWOdnJzEoEGDlM8xMTFi6tSpon379qJNmzZixowZ4s6dO/rqouQF8SQjkoeh1WqVEa8QQtSpU0cAJX6++eabYvf6p2MkLy9SSF5izp8/L7766qtiK7AmT54sVCqVWLx4sSImFy9eFC1bthROTk7i8OHDQggh1q9fL1QqlRJMLZr+KigoEBYWFmLkyJHKqpsuXboIMzMz4e3tLfr37y/q168vatasKVavXv3I1V8SieT1QSYkvsQ4Ozszffr0YtvGjx/P0aNHmTlzppKxHBYWxu3bt9mwYQPNmjUDYPXq1dSsWVOxaSny5tqzZw8A9vb2VKpUidu3b7N7924mTpzI9OnTSU9PJzY2loULF3Lr1i3y8vJkBrlE8pojheQVQgjBG2+8wZIlS/jhhx8IDQ2lYsWKuLq6MnfuXHr27Kkcu3v3bry9vXF0dARQrFRCQ0OpVq0arq6uQGGNE0dHRyUrvEqVKrRp04Y2bdqQnJwsRUQikaASQkZMX2Xu3LlDlSpVqFSpkmK/8vvvv+Pt7c3ixYsZOnSoYqkChVXgGjVqRHBwMCYmJmi1WpYsWcLo0aNp0qQJ48ePp0ePHlSsWLGMeyaRSMoLctXWK06NGjWoVKkS8L/ck927d/PGG2/QqFEj4H8VGvft28fdu3dp2bKlYt5nYGDAiBEjCAkJwdjYmHHjxikeSxKJRAJyRPLakpiYSNWqValQoYKStBgQEMCBAwdYtmwZHTp0ICsrC1NTU+WcGzduMGzYMMLDwzlx4sRrZXYokUgejRyRvKbY2toqdvVGRkZAoXW7nZ0dLi4uQGFhrYkTJ3Lnzh0KCgqoU6cOAQEB5Obmlqg3IZFIXl9ksF2isHnzZmUUcu/ePd566y3WrVtHXFwcXl5eqFQqfvrpJ2xtbRWxkUgkEjm1JSmBeKAO/G+//cbPP//MxYsXsbS0xNzcnFGjRinVHCUSiUQKieSJSE1N5dKlS9jb22NtbV3WzZFIJOUIKSQSiUQieS5ksF0ikUgkz4UUEolEIpE8F/8PniXsHrL/DZIAAAAASUVORK5CYII=",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "import matplotlib.pyplot as plt\n",
+ "\n",
+ "acts = ['a=0','a=1']\n",
+ "stas = ['T<21', '2124']\n",
+ "colors = ['b', 'g', 'r']\n",
+ "\n",
+ "fig = plt.figure()\n",
+ "ax = fig.add_subplot(111, projection='3d')\n",
+ "ax.set_xlabel('actions', labelpad=6, fontsize=12)\n",
+ "ax.set_ylabel('states', labelpad=10, fontsize=12)\n",
+ "ax.set_zlabel('$\\mathbf{q(s,a)}$', labelpad=0, fontsize=15)\n",
+ "plt.xticks(ticks=range(len(acts)), labels=acts)\n",
+ "plt.yticks(ticks=range(len(stas)), labels=stas)\n",
+ "\n",
+ "for i, s in enumerate(stas):\n",
+ " x = np.arange(len(acts))\n",
+ " h = model.q[i,:]\n",
+ "\n",
+ " # Set color\n",
+ " color = [colors[i]]*len(acts)\n",
+ "\n",
+ " # Plot the 3D bar graph\n",
+ " ax.bar(x, h, zs=i, zdir='y', color=color, alpha=0.8)\n",
+ "\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "V0AIl-HeVyqs"
+ },
+ "source": [
+ "Does it sound familiar? this is actually the [q-function that we had conceptually introduced before](#qFunctionConcept), but for our specific case!\n",
+ "\n",
+ "We observe that the state with the highest value is the one in the middle (green bars 🟢👌, `2124`), there is more value on `a=0`, so there is a preference for the agent to turn heating off.\n",
+ "\n",
+ "Sometimes it is useful to know what is the value of being on a specific state, independently of the action to be taken. This is represented by the so-called state-value function, which relates to the action-value function as follows:\n",
+ "\n",
+ "\\begin{align}\n",
+ " v(\\pmb{s}) = \\max_{\\pmb{a}} q(\\pmb{s},\\pmb{a})\n",
+ "\\end{align}\n",
+ "\n",
+ "At this point we can easily compute and plot the value function for our case:\n",
+ "\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 462
+ },
+ "id": "urJOkjSNoa-h",
+ "outputId": "8819e8b8-00b8-4177-898d-e81f283e2571"
+ },
+ "outputs": [
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "# Compute the state-value function\n",
+ "v = np.amax(model.q, axis=1)\n",
+ "\n",
+ "# Plot state-value function\n",
+ "fig = plt.figure()\n",
+ "\n",
+ "ax = fig.add_subplot(111)\n",
+ "ax.set_xlabel('states', labelpad=10, fontsize=12)\n",
+ "ax.set_ylabel('$\\mathbf{v(s)}$', labelpad=0, fontsize=15)\n",
+ "plt.xticks(ticks=range(len(stas)), labels=stas)\n",
+ "x = np.arange(len(stas))\n",
+ "ax.bar(x, v, color=colors, alpha=0.8)\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "clFn8dd7obRI"
+ },
+ "source": [
+ "Notice that we have trained our agent following an off-policy method: the actions were driven by a policy different than that one that our agent would follow. This is because the agent was using an epsilon-greedy policy to explore more rewarding actions. If we conclude we are happy with the learned policy, we can test it by setting `deterministic=True` with the `predict` method. For example, let's test our learned agent for the first day of February:\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "WuYEBf9nsmH6",
+ "outputId": "661787af-457d-4c58-e710-bcb35c156eee"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "-------------------------------------------------------------------\n",
+ "State [Bin #] = 1\n",
+ "Action [ - ] = 0\n",
+ "-------------------------------------------------------------------\n"
+ ]
+ }
+ ],
+ "source": [
+ "env.stop()\n",
+ "env = BoptestGymEnvCustomReward(url = url,\n",
+ " testcase = 'bestest_hydronic_heat_pump',\n",
+ " actions = ['oveHeaPumY_u'],\n",
+ " observations = {'reaTZon_y':(lower_setp,upper_setp)},\n",
+ " random_start_time = False,\n",
+ " start_time = 31*24*3600,\n",
+ " max_episode_length = 24*3600,\n",
+ " warmup_period = 24*3600,\n",
+ " step_period = 3600)\n",
+ "env = DiscretizedActionWrapper(env, n_bins_act=1)\n",
+ "env = DiscretizedObservationWrapper(env, n_bins_obs=3, outs_are_bins=True)\n",
+ "\n",
+ "done = False\n",
+ "obs, _ = env.reset()\n",
+ "\n",
+ "from IPython.display import clear_output\n",
+ "while not done:\n",
+ " # Clear the display output at each step\n",
+ " clear_output(wait=True)\n",
+ " # Compute control signal\n",
+ " action = model.predict(obs, deterministic=True)\n",
+ " # Print the current operative temperature and decided action\n",
+ " print('-------------------------------------------------------------------')\n",
+ " print('State [Bin #] = {:.0f}'.format(obs))\n",
+ " print('Action [ - ] = {:.0f}'.format(action))\n",
+ " print('-------------------------------------------------------------------')\n",
+ " # Implement action\n",
+ " obs,reward,terminated,truncated,info = env.step(action) # send the action to the environment\n",
+ " done = (terminated or truncated)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "sLBD3joyxe9Z"
+ },
+ "source": [
+ "Now there is no randomness involved. The agent exploits its policy by ALWAYS picking action `a=1` when `s=0` because it has learned that that is the action with the highest value in that state.\n",
+ "\n",
+ "We can now evaluate our learned policy by calculating the core KPIs with BOPTEST:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "eLzZaaNzyeZv",
+ "outputId": "7af5fe45-51ea-4b2b-899d-cfd2db3feee7"
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "{'tdis_tot': 1.7415794961384694,\n",
+ " 'idis_tot': 0,\n",
+ " 'ener_tot': 0.17501300879744733,\n",
+ " 'cost_tot': 0.044365797730152895,\n",
+ " 'emis_tot': 0.029227172469173696,\n",
+ " 'pele_tot': 0.01990768126278055,\n",
+ " 'pgas_tot': None,\n",
+ " 'pdih_tot': None,\n",
+ " 'time_rat': 0.0002186830550576178}"
+ ]
+ },
+ "execution_count": 24,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "env.get_kpis()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "WbDAlStV2Tvx"
+ },
+ "source": [
+ "This prepares the ground for different RL configurations to be evaluated and compared between each other and to other types of controls like classical rule based controllers or more advanced model predictive control. Recall that there are specific [scenario periods for each test case in BOPTEST](https://github.com/ibpsa/project1-boptest/tree/master/testcases#test-cases) that are set for these comparisons."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "FBL289bfsmcJ"
+ },
+ "source": [
+ "**Food for thought: 🤔**\n",
+ "- If the agent never receives a reward when the temperature is out of the comfort bounds (states 0 🔵 and 2 🔴), why is the q-function not 0 for those states?\n",
+ "- Could you think of measures to improve learning?\n",
+ "- Could you think of measures to improve performance?\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "eKqvn5yb_mqJ"
+ },
+ "source": [
+ "# **Gearing up** 💪"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "X1sdwYm5b66G"
+ },
+ "source": [
+ "The previously stylished example had a very limited representation of the state space. It was useful to illustrate how we can configure and train a RL agent without needing too many interactions with the environment (our building). However, using RL for solving this environment may feel like overkilling the problem. Our `SimpleController` was already enough to decide when to turn on heating based on indoor temperature readings. You should note, however, that you have developed a general agent capable of learning from any environment and the potential to infer way more complex relationships between environment observations and actions. Examples of what this RL agent could infer for building control are the following:\n",
+ "- Dynamic energy pricing\n",
+ "- A heating schedule based on user inputs.\n",
+ "- A heating curve based on ambient temperature.\n",
+ "- The variable heat pump COP based on condenser, evaporator, and ambient temperature reaadings.\n",
+ "\n",
+ "We could for example extend our reward function as to minimize the building energy use or the greenhouse gas emissions while keeping comfort.\n",
+ "And all this can be inferred without the need of a model that requires domain knowledge. On the downside, learning more complex dynamics from higher dymensional observation spaces requires more training data. This means that more interactions with the environment (the building) are required, which sometimes are unavailable. For this reason, sample-efficiency is key in RL and there exist several tricks to expedite learning.\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "V3ZF29MChF4F"
+ },
+ "source": [
+ "To finalize, we are going to instantiate a more complete building environment by extending the observation space with the time of the week as well as information about the ambient temeprature, solar irradiation, internal gains, electricity pricing, or temperature setpoints. With BOPTEST-Gym we can also establish a predictive and a regressive period that include predictions of the boundary condition data and past observations of the measured data, respectively.\n",
+ "\n",
+ "Because of its high dimensional state-action space, an agent will probably require many more interactions to solve this environment. Luckily, there are readily available state-of-the-art RL algorithms that use the learning principle you have learned above while implement all sort of tricks to expedite and stabilize learning. For example, we can access the advanced Deep Q-Network (DQN) algorithm from Stable-Baselines3 to learn this more complex environment. We set here our agent to learn for `10` steps to show how this learning process would be initiated."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 28,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "Bdx3qCDvhFSX",
+ "outputId": "d3fdf2ca-64d5-4e1d-ac9f-b999c7a3ff09"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Using cpu device\n",
+ "Wrapping the env with a `Monitor` wrapper\n",
+ "Wrapping the env in a DummyVecEnv.\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "{'tdis_tot': 8.370911955136352,\n",
+ " 'idis_tot': 0,\n",
+ " 'ener_tot': 0.3965284774903316,\n",
+ " 'cost_tot': 0.10051996904379909,\n",
+ " 'emis_tot': 0.06622025574088537,\n",
+ " 'pele_tot': 0.021041629229404186,\n",
+ " 'pgas_tot': None,\n",
+ " 'pdih_tot': None,\n",
+ " 'time_rat': 0.0004946192215990137}"
+ ]
+ },
+ "execution_count": 28,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "# env.stop()\n",
+ "\n",
+ "from boptestGymEnv import BoptestGymEnv, NormalizedObservationWrapper, DiscretizedActionWrapper\n",
+ "from stable_baselines3 import DQN\n",
+ "\n",
+ "# url for the BOPTEST service\n",
+ "url = 'https://api.boptest.net'\n",
+ "\n",
+ "# Decide the state-action space of your test case\n",
+ "env = BoptestGymEnv(\n",
+ " url = url,\n",
+ " testcase = 'bestest_hydronic_heat_pump',\n",
+ " actions = ['oveHeaPumY_u'],\n",
+ " observations = {'time':(0,604800),\n",
+ " 'reaTZon_y':(280.,310.),\n",
+ " 'TDryBul':(265,303),\n",
+ " 'HDirNor':(0,862),\n",
+ " 'InternalGainsRad[1]':(0,219),\n",
+ " 'PriceElectricPowerHighlyDynamic':(-0.4,0.4),\n",
+ " 'LowerSetp[1]':(280.,310.),\n",
+ " 'UpperSetp[1]':(280.,310.)},\n",
+ " predictive_period = 24*3600,\n",
+ " regressive_period = 6*3600,\n",
+ " random_start_time = True,\n",
+ " max_episode_length = 24*3600,\n",
+ " warmup_period = 24*3600,\n",
+ " step_period = 3600)\n",
+ "\n",
+ "# Normalize observations and discretize action space\n",
+ "env = NormalizedObservationWrapper(env)\n",
+ "env = DiscretizedActionWrapper(env,n_bins_act=10)\n",
+ "\n",
+ "# Instantiate an RL agent\n",
+ "model = DQN('MlpPolicy', env, verbose=1, gamma=0.99,\n",
+ " learning_rate=5e-4, batch_size=24, seed=123456,\n",
+ " buffer_size=365*24, learning_starts=24, train_freq=1)\n",
+ "\n",
+ "# Main training loop\n",
+ "model.learn(total_timesteps=10)\n",
+ "\n",
+ "# Loop for one episode of experience (one day as set in max_episode_length)\n",
+ "done = False\n",
+ "obs, _ = env.reset()\n",
+ "while not done:\n",
+ " action, _ = model.predict(obs, deterministic=True)\n",
+ " obs,reward,terminated,truncated,info = env.step(action)\n",
+ " done = (terminated or truncated)\n",
+ "\n",
+ "# Obtain KPIs for evaluation\n",
+ "env.get_kpis()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "vudahvPN_ZaA"
+ },
+ "source": [
+ "Learning for 10 interaction steps is clearly not enough and leads to poor performance. This new environment has a way higher dimensional state-action space than the ones we treated before:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 26,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "GSg90XCe-26Q",
+ "outputId": "8f6d8dcd-3f11-420f-d6a7-918f1616cffb"
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Observation space of the building environment (dimension):\n",
+ "(158,)\n",
+ "Action space of the building environment:\n",
+ "Discrete(11)\n"
+ ]
+ }
+ ],
+ "source": [
+ "print('Observation space of the building environment (dimension):')\n",
+ "print(env.observation_space.shape)\n",
+ "print('Action space of the building environment:')\n",
+ "print(env.action_space)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "sOs93H9X_ZaA"
+ },
+ "source": [
+ "Solving an environment of these dimensions requires millions of steps or other tricks to accelerate learning. Could you think of any?"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 27,
+ "metadata": {
+ "id": "EDJHCuQ2NFN6"
+ },
+ "outputs": [],
+ "source": [
+ "env.stop()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "F-_f2qRTB0Nw"
+ },
+ "source": [
+ "# **Further resources** 📚"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "p54OK_TGrtfp"
+ },
+ "source": [
+ "- For RL, check out the resources page from Stable-Baselines 3 [here](https://stable-baselines3.readthedocs.io/en/master/guide/rl.html) and the [open access book of Richard S. Sutton and Andrew G. Barto](http://incompleteideas.net/book/the-book-2nd.html)\n",
+ "- For BOPTEST, check out the websites of the [BOPTEST framework](https://ibpsa.github.io/project1-boptest/), its [GitHub repository](https://ibpsa.github.io/project1-boptest/), and its overarching project: [IBPSA Project 1](https://ibpsa.github.io/project1/)."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Jblq_C7CHQHj"
+ },
+ "source": [
+ "# **Feedback** 💬"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "jJ9lmUndHLMq"
+ },
+ "source": [
+ "Please help us improve by filling out [this form](https://forms.gle/JdprK6tgxQtwvhFV8). It'll only take a couple of minutes!"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "WNB5MoRmOWc9"
+ },
+ "source": [
+ "#**Annex I: Formal Reinforcement Learning theory** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "G-ZId2TdCngy"
+ },
+ "source": [
+ "In RL we aim to derive an optimal control policy from the direct interaction of an agent (the RL algorithm) and an environment (the process to be optimized).\n",
+ "A policy is a mapping from environment states to actions that the agent \"decides\" to take.\n",
+ "This control method is based on the principle of dynamic programming. Unlike\n",
+ "classical dynamic programming, RL does not assume the existence of a perfect\n",
+ "system model and uses function approximations to build a policy from samples\n",
+ "of historical data. Hence, the agent performs empirical learning and decides on\n",
+ "actions to drive the environment towards favorable trajectories according to a reward function that the environment delivers every control step.\n",
+ "\n",
+ "The process of the RL agent interacting with the environment is a sequential decision-making problem formalized as a **Markov Decission Process (MDP)**. A diagram summarizing the RL approach is shown in the following figure:\n",
+ "\n",
+ "\n",
+ "\n",
+ "*Figure: Diagram of the RL approach. The RL agent decides an action. After the action is implemented, the environment returns the new state $\\pmb{S}_{k+1}$ and associated reward $R_{k+1}$.*\n",
+ "\n",
+ "In an MDP, the agent and the environment interact during a sequence of discrete-time steps indexed here as $k=0,1,2,...,K$, with $K$ being the terminal sample that could be $K=\\infty$.\n",
+ "Every time step $k$ the agent receives a representation of the environment named state: $\\pmb{S}_k \\in \\pmb{\\mathcal{S}}$, where $\\pmb{\\mathcal{S}}$ is the state space.\n",
+ "Note that the agent's observation of the state-space may or may not fully characterize the environment state.\n",
+ "In the latter case where the agent can only see a partial observation of the environment's state-space, we refer to **partially observable Markov decision processes (POMDPs)**.\n",
+ "\n",
+ "Upon receiving the state representation, the agent computes its control logic and in turn sends back to the environment a control action $\\pmb{A}_k \\in \\pmb{\\mathcal{A}}$, where $\\pmb{A}_k$ is the most appropriate action chosen from the action space $\\pmb{\\mathcal{A}}$.\n",
+ "One time step later, the agent observes a new state from the environment $\\pmb{S}_{k+1}$ along with a scalar value indicating its reward $R_{k+1} \\in \\mathcal{R} \\subset{\\mathbb{R}}$. Notice that the reward $R_{k+1}$ is an indicator of the agent's performance when taking action $\\pmb{A}_k$ from state $\\pmb{S}_k$.\n",
+ "\n",
+ "The environment $\\mathcal{E}_{\\pmb{f}}$ is governed by the natural laws of the system dynamics $\\pmb{f}$ and it is defined by $\\mathcal{E}_{\\pmb{f}}:\\pmb{\\mathcal{S}}\\times \\pmb{\\mathcal{A}} \\rightarrow \\pmb{\\mathcal{S}}\\times \\mathcal{R}$.\n",
+ "The goal of RL is to infer an **optimal control policy** $\\pi_{*}:\\pmb{\\mathcal{S}} \\rightarrow \\pmb{\\mathcal{A}}$ that maximizes the **expected cumulative return** $G$ when the agent acts according to it.\n",
+ "The cumulative return is defined as some function of the rewards sequence, and a typical definition is to discount the rewards with a **discount factor** $\\gamma \\in [0,1]$ as shown in the following equation:\n",
+ "\n",
+ "\\begin{align}\n",
+ " G_k = R_{k+1} + \\gamma R_{k+2} + \\gamma^2 R_{k+3} + ... = \\sum_{i=0}^\\infty \\gamma^i R_{k+i+1}\n",
+ "\\end{align}\n",
+ "\n",
+ "The **action-value function** $q(\\pmb{S},\\pmb{A})$ estimates the expected return when being in a specific state $\\pmb{S}$ and taking an action $\\pmb{A}$.\n",
+ "The **state-value function** $v(\\pmb{S})$ directly estimates the expected return for being in state $\\pmb{S}$.\n",
+ "Frequently, the policy and value functions are approximated by **function approximations** to cope with high-dimensional state-action spaces.\n",
+ "Examples of commonly used regressors are neural networks or randomized trees.\n",
+ "\n",
+ "\n",
+ "\n",
+ "A **trajectory** of an MDP is defined as a sequence of states, actions and rewards.\n",
+ "Most of the RL algorithms learn from finite trajectories of experience called **episodes**.\n",
+ "Sometimes, the trajectories are broken down into tuples of the form $(\\pmb{s}_k,\\pmb{a}_k,r_k,\\pmb{s}_{k+1})$ and stored in a **replay memory** $\\pmb{\\mathcal{D}}$.\n",
+ "Using a replay memory allows to serve the historical data in random batches of tuples to preserve as much as possible the independent and identically distributed assumption that is typically taken to parametrize policies and value functions.\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "_RIO07aKaQHG"
+ },
+ "source": [
+ "# **References** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "cQM-3Ra5BYM7"
+ },
+ "source": [
+ "\n",
+ "- **[1]** *Blum, D., Arroyo, J., Huang, S., Drgona, J., Jorissen, F., Taxt Walnum, H., Yan, C., Benne, K., Vrabie, D., Wetter, M., and Helsen,\n",
+ "L. Building Optimization Testing Framework (BOPTEST) for Simulation-\n",
+ "Based Benchmarking of Control Strategies in Buildings. Journal of Building\n",
+ "Performance Simulation 14, 5 (2021), 586–610. https://doi.org/10.1080/19401493.2021.1986574*\n",
+ "\n",
+ "- **[2]** *Arroyo, J., Manna, C., Spiessens, F., and Helsen, L. An OpenAI-Gym\n",
+ "environment for the Building Optimization Testing (BOPTEST) framework.\n",
+ "In Proceedings of the 17th IBPSA Conference (Bruges, Belgium, September 2021) [https://doi.org/10.26868/25222708.2021.30380](https://www.conftool.pro/bs2021/index.php/30380_Arroyo_Javier.pdf?page=downloadPaper&filename=30380_Arroyo_Javier.pdf&form_id=30380)*\n",
+ "\n",
+ "- **[3]** *Drgona, J., Arroyo, J., Cupeiro Figueroa, I., Blum, D., Arendt, K., Kim, D.,Ollé, E. P., Oravec, J., Wetter, M., Vrabie, D. L., and Helsen, L. All you need to know about model predictive control for buildings. Annual Reviews in Control 50 (2020), 190–232. https://doi.org/10.1016/j.arcontrol.2020.09.001*\n",
+ "\n",
+ "- **[4]** *Vázquez-Canteli, J. R., and Nagy, Z. Reinforcement learning\n",
+ "for demand response: A review of algorithms and modeling techniques.\n",
+ "Applied energy 235 (2019), 1072–1089. https://doi.org/10.1016/j.apenergy.2018.11.002*\n",
+ "\n",
+ "- **[5]** *Chen, B., Cai, Z., and Bergés, M. Gnu-RL: A Practical and Scalable Reinforcement Learning Solution for Building HVAC Control Using a Differentiable MPC Policy. Frontiers in Built Environment 6 (2020). https://doi.org/10.3389/fbuil.2020.562239*\n",
+ "\n",
+ "- **[6]** *Sutton, R. S., and Barto, A. G. Reinforcement Learning: An Introduction, second ed. The MIT Press, 2018.*\n"
+ ]
+ }
+ ],
+ "metadata": {
+ "colab": {
+ "include_colab_link": true,
+ "provenance": []
+ },
+ "kernelspec": {
+ "display_name": "Python 3 (ipykernel)",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.10.11"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 0
+}
diff --git a/docs/tutorials/CCAI Summer School 2022/README.md b/docs/tutorials/CCAI_Summer_School_2022/README.md
similarity index 100%
rename from docs/tutorials/CCAI Summer School 2022/README.md
rename to docs/tutorials/CCAI_Summer_School_2022/README.md
diff --git a/examples/test_and_plot.py b/examples/test_and_plot.py
index 952759e..b6c43ca 100644
--- a/examples/test_and_plot.py
+++ b/examples/test_and_plot.py
@@ -68,7 +68,8 @@ def test_agent(env, model, start_time, episode_length, warmup_period,
return observations, actions, rewards, kpis
-def plot_results(env, rewards, points=['reaTZon_y','oveHeaPumY_u'],
+def plot_results(env, rewards, points=['reaTZon_y','reaTSetHea_y','reaTSetCoo_y','oveHeaPumY_u',
+ 'weaSta_reaWeaTDryBul_y', 'weaSta_reaWeaHDirNor_y'],
log_dir=os.getcwd(), model_name='last_model', save_to_file=False):
@@ -79,48 +80,23 @@ def plot_results(env, rewards, points=['reaTZon_y','oveHeaPumY_u'],
# Retrieve all simulation data
# We use env.start_time+1 to ensure that we don't return the last
# point from the initialization period to don't confuse it with
- # actions taken by the agent
+ # actions taken by the agent in a previous episode.
res = requests.put('{0}/results'.format(env.url),
data={'point_names':points,
- 'start_time':env.start_time+1,
- 'final_time':3.1536e7}).json()['payload']
+ 'start_time':env.start_time+1,
+ 'final_time':3.1536e7}).json()['payload']
- df_res = pd.DataFrame(res).set_index('time')
-
- # Retrieve boundary condition data.
- # Only way we have is through the forecast request.
+ df = pd.DataFrame(res)
+ df = create_datetime_index(df)
+ df.dropna(axis=0, inplace=True)
scenario = env.scenario
- requests.put('{0}/initialize'.format(env.url),
- data={'start_time':df_res['time'].iloc[0],
- 'warmup_period':0}).json()['payload']
-
- # Store original forecast parameters
- forecast_parameters_original = requests.get('{0}/forecast_parameters'.format(env.url)).json()['payload']
- # Set forecast parameters for test. Take 10 points per step.
- forecast_parameters = {'horizon':env.max_episode_length,
- 'interval':env.step_period/10}
- requests.put('{0}/forecast_parameters'.format(env.url),
- data=forecast_parameters)
- forecast = requests.get('{0}/forecast'.format(env.url)).json()['payload']
- # Back to original parameters, just in case we're testing during training
- requests.put('{0}/forecast_parameters'.format(env.url),
- data=forecast_parameters_original)
-
- df_for = pd.DataFrame(forecast)
- df_for = reindex(df_for)
- df_for.drop('time', axis=1, inplace=True)
-
- df = pd.concat((df_res,df_for), axis=1)
- df = create_datetime(df)
-
- df.dropna(axis=0, inplace=True)
-
if save_to_file:
df.to_csv(os.path.join(log_dir, 'results_tests_'+model_name+'_'+scenario['electricity_price'],
'results_sim_{}.csv'.format(str(int(res['time'][0]/3600/24)))))
-
- rewards_time_days = np.arange(df_res['time'].iloc[0],
+
+ # Project rewards into results index
+ rewards_time_days = np.arange(df['time'][0],
env.start_time+env.max_episode_length,
env.step_period)/3600./24.
f = interpolate.interp1d(rewards_time_days, rewards, kind='zero',
@@ -129,43 +105,34 @@ def plot_results(env, rewards, points=['reaTZon_y','oveHeaPumY_u'],
rewards_reindexed = f(res_time_days)
if not plt.get_fignums():
- # no window(s) open
- # fig = plt.figure(figsize=(10,8))
+ # no window(s) are open, so open a new window.
_, axs = plt.subplots(4, sharex=True, figsize=(8,6))
else:
- # get current figure. Combine this with plt.ion(), plt.figure()
+ # There is a window open, so get current figure.
+ # Combine this with plt.ion(), plt.figure()
fig = plt.gcf()
axs = fig.subplots(nrows=4, ncols=1, sharex=True)
x_time = df.index.to_pydatetime()
axs[0].plot(x_time, df['reaTZon_y'] -273.15, color='darkorange', linestyle='-', linewidth=1, label='_nolegend_')
- axs[0].plot(x_time, df['LowerSetp[1]'] -273.15, color='gray', linewidth=1, label='Comfort setp.')
- axs[0].plot(x_time, df['UpperSetp[1]'] -273.15, color='gray', linewidth=1, label='_nolegend_')
+ axs[0].plot(x_time, df['reaTSetHea_y'] -273.15, color='gray', linewidth=1, label='Comfort setp.')
+ axs[0].plot(x_time, df['reaTSetCoo_y'] -273.15, color='gray', linewidth=1, label='_nolegend_')
axs[0].set_yticks(np.arange(15, 31, 5))
axs[0].set_ylabel('Operative\ntemperature\n($^\circ$C)')
- axt = axs[0].twinx()
- axt.plot(x_time, df['PriceElectricPowerHighlyDynamic'], color='dimgray', linestyle='dotted', linewidth=1, label='Price')
- axs[0].plot([],[], color='dimgray', linestyle='-', linewidth=1, label='Price')
-
- axt.set_ylim(0,0.3)
- axt.set_yticks(np.arange(0, 0.31, 0.1))
- axt.set_ylabel('(EUR/kWh)')
- axt.set_ylabel('Price\n(EUR/kWh)')
-
axs[1].plot(x_time, df['oveHeaPumY_u'], color='darkorange', linestyle='-', linewidth=1, label='_nolegend_')
axs[1].set_ylabel('Heat pump\nmodulation\nsignal\n( - )')
axs[2].plot(x_time, rewards_reindexed, 'b', linewidth=1, label='rewards')
axs[2].set_ylabel('Rewards\n(-)')
- axs[3].plot(x_time, df['TDryBul'] - 273.15, color='royalblue', linestyle='-', linewidth=1, label='_nolegend_')
+ axs[3].plot(x_time, df['weaSta_reaWeaTDryBul_y'] - 273.15, color='royalblue', linestyle='-', linewidth=1, label='_nolegend_')
axs[3].set_ylabel('Ambient\ntemperature\n($^\circ$C)')
axs[3].set_yticks(np.arange(-5, 16, 5))
axt = axs[3].twinx()
- axt.plot(x_time, df['HDirNor'], color='gold', linestyle='-', linewidth=1, label='$\dot{Q}_rad$')
+ axt.plot(x_time, df['weaSta_reaWeaHDirNor_y'], color='gold', linestyle='-', linewidth=1, label='$\dot{Q}_rad$')
axt.set_ylabel('Solar\nirradiation\n($W$)')
axs[3].plot([],[], color='darkorange', linestyle='-', linewidth=1, label='RL')
@@ -179,9 +146,10 @@ def plot_results(env, rewards, points=['reaTZon_y','oveHeaPumY_u'],
plt.tight_layout()
if save_to_file:
- plt.savefig(os.path.join(log_dir, 'results_tests_'+model_name+'_'+scenario['electricity_price'],
- 'results_sim_{}.pdf'.format(str(int(res['time'][0]/3600/24)))),
- bbox_inches='tight')
+ dir_name = os.path.join(log_dir, 'results_tests_'+model_name+'_'+scenario['electricity_price'])
+ fil_name = os.path.join(dir_name,'results_sim_{}.pdf'.format(str(int(res['time'][0]/3600/24))))
+ os.makedirs(dir_name, exist_ok=True)
+ plt.savefig(fil_name, bbox_inches='tight')
if not save_to_file:
# showing and saving to file are incompatible
@@ -199,9 +167,9 @@ def reindex(df, interval=60, start=None, stop=None):
'''
if start is None:
- start = df['time'][df.index[0]]
+ start = df['time'][0]
if stop is None:
- stop = df['time'][df.index[-1]]
+ stop = df['time'][-1]
index = np.arange(start,stop+0.1,interval).astype(int)
df_reindexed = df.reindex(index)
@@ -218,7 +186,7 @@ def reindex(df, interval=60, start=None, stop=None):
return df_reindexed
-def create_datetime(df):
+def create_datetime_index(df):
'''
Create a datetime index for the data
@@ -226,7 +194,7 @@ def create_datetime(df):
datetime = []
for t in df['time']:
- datetime.append(pd.Timestamp('2020/1/1') + pd.Timedelta(t,'s'))
+ datetime.append(pd.Timestamp('2023/1/1') + pd.Timedelta(t,'s'))
df['datetime'] = datetime
df.set_index('datetime', inplace=True)
diff --git a/testing/references/tutorial_output_get_name.json b/testing/references/tutorial_output_get_name.json
new file mode 100644
index 0000000..1b51e94
--- /dev/null
+++ b/testing/references/tutorial_output_get_name.json
@@ -0,0 +1 @@
+{"name": "bestest_hydronic_heat_pump"}
\ No newline at end of file
diff --git a/testing/references/tutorial_output_kpis_DQN_alg.json b/testing/references/tutorial_output_kpis_DQN_alg.json
new file mode 100644
index 0000000..4556eb1
--- /dev/null
+++ b/testing/references/tutorial_output_kpis_DQN_alg.json
@@ -0,0 +1 @@
+{"tdis_tot": 8.370911955136352, "idis_tot": 0, "ener_tot": 0.3965284774903316, "cost_tot": 0.10051996904379909, "emis_tot": 0.06622025574088537, "pele_tot": 0.021041629229404186, "pgas_tot": null, "pdih_tot": null}
\ No newline at end of file
diff --git a/testing/references/tutorial_output_kpis_Q_alg.json b/testing/references/tutorial_output_kpis_Q_alg.json
new file mode 100644
index 0000000..f0faa5f
--- /dev/null
+++ b/testing/references/tutorial_output_kpis_Q_alg.json
@@ -0,0 +1 @@
+{"tdis_tot": 1.7415794961384694, "idis_tot": 0, "ener_tot": 0.17501300879744733, "cost_tot": 0.044365797730152895, "emis_tot": 0.029227172469173696, "pele_tot": 0.01990768126278055, "pgas_tot": null, "pdih_tot": null}
\ No newline at end of file
diff --git a/testing/test_boptestGymEnv.py b/testing/test_boptestGymEnv.py
index e104932..0b32c68 100644
--- a/testing/test_boptestGymEnv.py
+++ b/testing/test_boptestGymEnv.py
@@ -421,6 +421,89 @@ def check_obs_act_rew_kpi(self, obs=None, act=None, rew=None, kpi=None,
df.dropna(inplace=True)
ref_filepath = os.path.join(utilities.get_root_path(), 'testing', 'references', 'kpis_{}.csv'.format(label))
self.compare_ref_values_df(df, ref_filepath)
+
+ def test_tutorial(self):
+ '''
+ Test the tutorial in the `docs`. The tutorial is written as
+ an ipython notebook so the `nbconvert` package is used to convert
+ the notebook to plain python to execute the test by comparing the
+ outputs of some of the notebook cells with references.
+ Note that the notebook actually uses the `boptest-gym-service`
+ branch, which should be even with the `master` branch but uses
+ BOPTEST-Service. Therefore, this is a check for the
+ `boptest-gym-service` branch and, contrarily to other tests,
+ this one could be parallelized. The last section of the tutorial
+ (Gearing Up) is using the DQN algorithm from stable-baselines3
+ and is used as such in the Quick Start example in the README.md
+ of this repository. Therefore, this is also testing the
+ Quick Start example.
+
+ '''
+
+ from nbconvert.preprocessors import ExecutePreprocessor
+ import nbformat
+
+ # Get root directory
+ root_dir = utilities.get_root_path()
+
+ # Change working dir to tutorial directory
+ run_path = os.chdir(os.path.join(root_dir, 'docs', 'tutorials', 'CCAI_Summer_School_2022'))
+
+ # Path to the notebook file
+ notebook_path = os.path.join(root_dir, 'docs', 'tutorials', 'CCAI_Summer_School_2022',
+ 'Building_Control_with_RL_using_BOPTEST.ipynb')
+
+ # Read the notebook file
+ with open(notebook_path, 'r', encoding='utf-8') as f:
+ notebook_content = f.read()
+
+ # Execute the notebook cells
+ executor = ExecutePreprocessor(timeout=-1, resources={'metadata': {'path': run_path}})
+ executed_notebook, _ = executor.preprocess(nbformat.reads(notebook_content, as_version=4),
+ resources={'metadata': {'path': run_path}})
+
+ # Test output when requesting test case name
+ out_get_name = executed_notebook.cells[41].outputs[0]['text']
+ self.check_from_cell_output(out_get_name, 'get_name')
+
+ # Check KPIs when testing our Q-algorithm
+ out_kpis_Q_alg = executed_notebook.cells[119].outputs[0]['data']['text/plain']
+ self.check_from_cell_output(out_kpis_Q_alg, 'kpis_Q_alg')
+
+ # Check KPIs when testing DQN algorithm from stable-baselines3
+ out_kpis_DQN_alg = executed_notebook.cells[125].outputs[2]['data']['text/plain']
+ self.check_from_cell_output(out_kpis_DQN_alg, 'kpis_DQN_alg')
+
+ def check_from_cell_output(self, cell_output, str_output):
+ '''Compares a cell output to a reference file.
+ Parameters
+ ----------
+ cell_output: str
+ Content of the cell output that is
+ reformatted in this method to become json
+ str_ouput: str
+ Tag to identify the reference file of the output
+
+ '''
+
+ import json
+
+ # Conform to the json syntax rules to transform to json
+ out = cell_output.replace("\n","").replace("'","\"").replace("None","null")
+
+ # Convert string to json
+ out_json = json.loads(out)
+
+ # Drop time ratio if it is in output
+ if 'time_rat' in out_json:
+ del out_json['time_rat']
+
+ # Assign files
+ file_ref = os.path.join(utilities.get_root_path(), 'testing', 'references',
+ 'tutorial_output_{}.json'.format(str_output))
+ # Check results
+ self.compare_ref_json(out_json, file_ref)
+
if __name__ == '__main__':
utilities.run_tests(os.path.basename(__file__))
![\"Open](\"https://colab.research.google.com/assets/colab-badge.svg\"){kind=icon}