(v3.6.8) - OU process migration (#457)

* Added ou process function in common module

* Create test for ou process in test_common.py

* Adapt modeling module to apply weather variability with this function

* Weather Forecasting wrapper: Miograted ou process with common module call. Added observation variables, updated tests, improved some implementations aspects

* EnergyCost wrapper: Migrated ou process with common module call. Update tests and improved some implementation aspects

* Wrappers: Improved redifinition of observation spaces, now read from old observation spaces relevant information

* Fix code

* Update Sinergym version from 3.6.7 to 3.6.8

pyproject.toml

@@ -6,7 +6,7 @@
 package-mode = true
 name = "sinergym"
 
-version = "3.6.7"
+version = "3.6.8"
 description = "The goal of sinergym is to create an environment following OpenAI Gym interface for wrapping simulation engines for building control using deep reinforcement learning."
 license = "MIT"
 

sinergym/config/modeling.py

@@ -11,7 +11,8 @@
 from eppy.modeleditor import IDF
 from epw.weather import Weather
 
-from sinergym.utils.common import eppy_element_to_dict, get_delta_seconds
+from sinergym.utils.common import (eppy_element_to_dict, get_delta_seconds,
+                                   ornstein_uhlenbeck_process)
 from sinergym.utils.constants import (CWD, LOG_MODEL_LEVEL, PKG_DATA_PATH,
                                       WEEKDAY_ENCODING, YEAR)
 from sinergym.utils.logger import TerminalLogger
@@ -326,38 +327,16 @@ def apply_weather_variability(
         Returns:
             str: New EPW file path generated in simulator working path in that episode or current EPW path if variation is not defined.
         """
-        # deepcopy for weather_data
-        weather_data_mod = deepcopy(self.weather_data)
+
         filename = self._weather_path.split('/')[-1]
+        weather_data_mod = deepcopy(self.weather_data)
 
         # Apply variation to EPW if exists
         if weather_variability is not None:
 
-            T = 1.  # Total time.
-            # All the columns are going to have the same num of rows since they are
-            # in the same dataframe
-            # get first column of df
-            n = weather_data_mod.dataframe.shape[0]
-            dt = T / n
-            # t = np.linspace(0., T, n)  # Vector of times.
-
-            for variable, variation in weather_variability.items():
-
-                sigma = variation[0]  # Standard deviation.
-                mu = variation[1]  # Mean.
-                tau = variation[2]  # Time constant.
-
-                sigma_bis = sigma * np.sqrt(2. / tau)
-                sqrtdt = np.sqrt(dt)
-
-                # Create noise
-                noise = np.zeros(n)
-                for i in range(n - 1):
-                    noise[i + 1] = noise[i] + dt * (-(noise[i] - mu) / tau) + \
-                        sigma_bis * sqrtdt * np.random.randn()
-
-                # Add noise
-                weather_data_mod.dataframe[variable] += noise
+            weather_data_mod.dataframe = ornstein_uhlenbeck_process(
+                data=self.weather_data.dataframe,
+                variability_config=weather_variability)
 
             self.logger.info(
                 'Weather noise applied in columns: {}'.format(

sinergym/utils/common.py

@@ -1,10 +1,12 @@
 """Common utilities."""
 
+from copy import deepcopy
 from datetime import datetime, timedelta
 from typing import Any, Dict, Optional, Tuple, Type, Union
 
 import gymnasium as gym
 import numpy as np
+import pandas as pd
 import xlsxwriter
 from eppy.modeleditor import IDF
 
@@ -180,6 +182,52 @@ def export_schedulers_to_excel(
         object_num += 1
     workbook.close()
 
+# ------------------------ Ornstein Unhelbeck Process ------------------------ #
+
+
+def ornstein_uhlenbeck_process(
+        data: pd.DataFrame,
+        variability_config: Dict[str, Tuple[float, float, float]]) -> pd.DataFrame:
+    """Add noise to the data using the Ornstein-Uhlenbeck process.
+
+    Args:
+        data (pd.DataFrame): Data to be modified.
+        variability_config (Dict[str, Tuple[float, float, float]]): Dictionary with the variability configuration for each variable (sigma, mu and tau constants).
+
+    Returns:
+        pd.DataFrame: Data with noise added.
+    """
+
+    # deepcopy for weather_data
+    data_mod = deepcopy(data)
+
+    # Total time.
+    T = 1.
+    # get first column of df
+    n = data_mod.shape[0]
+    dt = T / n
+    # t = np.linspace(0., T, n)  # Vector of times.
+
+    for variable, variation in variability_config.items():
+
+        sigma = variation[0]  # Standard deviation.
+        mu = variation[1]  # Mean.
+        tau = variation[2]  # Time constant.
+
+        sigma_bis = sigma * np.sqrt(2. / tau)
+        sqrtdt = np.sqrt(dt)
+
+        # Create noise
+        noise = np.zeros(n)
+        for i in range(n - 1):
+            noise[i + 1] = noise[i] + dt * (-(noise[i] - mu) / tau) + \
+                sigma_bis * sqrtdt * np.random.randn()
+
+        # Add noise
+        data_mod[variable] += noise
+
+    return data_mod
+
 # ------------------ Reading JSON environment configuration ------------------ #
 
 

sinergym/utils/logger.py

@@ -245,7 +245,7 @@ def write(
             # Log all metrics
             wandb.log(metrics_to_log)
 except ImportError:
-    class WandBOutputFormat():
+    class WandBOutputFormat():  # pragma: no cover
         """WandBOutputFormat class for logging in WandB from SB3 logger.
         """