Release v24.9.1

.github/workflows/scheduled_tests.yaml

@@ -6,9 +6,9 @@ on:
     branches:
       - main
 
-  # runs every day at 09:00 UTC
+  # runs every day at 03:00 UTC
   schedule:
-    - cron: '0 9 * * *'
+    - cron: '0 3 * * *'
 
 # Check noxfile.py for associated environment variables
 env:

CHANGELOG.md

@@ -6,6 +6,17 @@
 
 ## Breaking Changes
 
+# [v24.9.1](https://github.com/pybop-team/PyBOP/tree/v24.9.0) - 2024-09-16
+
+
+## Features
+
+## Bug Fixes
+
+- [#495](https://github.com/pybop-team/PyBOP/pull/495) - Bugfixes for Transformation class, adds `apply_transform` optional arg to `BaseCost` for transformation functionality.
+
+## Breaking Changes
+
 # [v24.9.0](https://github.com/pybop-team/PyBOP/tree/v24.9.0) - 2024-09-10
 
 ## Features

CITATION.cff

@@ -11,5 +11,5 @@ authors:
     family-names: Courtier
   - given-names: David
     family-names: Howey
-version: "24.9.0" # Update this when you release a new version
+version: "24.9.1" # Update this when you release a new version
 repository-code: 'https://www.github.com/pybop-team/pybop'

README.md

@@ -15,12 +15,12 @@
   [![Open Issues](https://img.shields.io/github/issues/pybop-team/PyBOP)](https://github.com/pybop-team/PyBOP/issues/)
   [![License](https://img.shields.io/github/license/pybop-team/PyBOP?color=blue)](https://github.com/pybop-team/PyBOP/blob/develop/LICENSE)
   [![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/pybop-team/PyBOP/blob/develop/)
+  [![nbviewer](https://raw.githubusercontent.com/jupyter/design/master/logos/Badges/nbviewer_badge.svg)](https://nbviewer.org/github/pybop-team/PyBOP/tree/develop/examples/notebooks/)
   [![Static Badge](https://img.shields.io/badge/https%3A%2F%2Fpybop-team.github.io%2Fpybop-bench%2F?label=Benchmarks)](https://pybop-team.github.io/pybop-bench/)
   [![Releases](https://img.shields.io/github/v/release/pybop-team/PyBOP?color=gold)](https://github.com/pybop-team/PyBOP/releases)
 
 </div>
 
-## PyBOP
 PyBOP provides a complete set of tools for parameterisation and optimisation of battery models, using both Bayesian and frequentist approaches, with [example workflows](https://github.com/pybop-team/PyBOP/tree/develop/examples/notebooks) to assist the user. PyBOP can be used to parameterise various battery models, including electrochemical and equivalent circuit models available in [PyBaMM](https://pybamm.org/). PyBOP prioritises clear and informative diagnostics for the user, while also allowing for advanced probabilistic methods.
 
 The diagram below shows the conceptual framework of PyBOP. This package is currently under development, so users can expect the API to evolve with future releases.
@@ -54,44 +54,50 @@ To check that PyBOP is installed correctly, run one of the examples in the follo
 ## Using PyBOP
 PyBOP has two intended uses:
 
-1. Parameter estimation from battery test data.
+1. Parameter inference from battery test data.
 
 2. Design optimisation under battery manufacturing/use constraints.
 
 These include a wide variety of optimisation problems that require careful consideration due to the choice of battery model, data availability and/or the choice of design parameters.
 
-### Notebooks
-PyBOP comes with a number of [example notebooks](https://github.com/pybop-team/PyBOP/blob/develop/examples), which can be found in the examples folder. A few noteworthy ones are listed below.
+### Jupyter Notebooks
 
-- [Gravimetric design optimisation of a single particle model](https://github.com/pybop-team/PyBOP/blob/develop/examples/notebooks/spm_electrode_design.ipynb)
-- [Experimental GITT fitting of an ECM for an LG M50](https://github.com/pybop-team/PyBOP/blob/develop/examples/notebooks/LG_M50_ECM/1-single-pulse-circuit-model.ipynb)
-- [Compare PyBOP optimisers for parameter identification](https://github.com/pybop-team/PyBOP/blob/develop/examples/notebooks/multi_optimiser_identification.ipynb)
-- [Parameter identification for a spatial pouch cell model](https://github.com/pybop-team/PyBOP/blob/develop/examples/notebooks/pouch_cell_identification.ipynb)
+Explore our [example notebooks](https://github.com/pybop-team/PyBOP/blob/develop/examples) for hands-on demonstrations:
 
-### Scripts
-Additional script-based examples can be found in the [examples directory](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/). Some notable scripts are listed below.
+- [Gravimetric design optimisation (SPM)](https://github.com/pybop-team/PyBOP/blob/develop/examples/notebooks/spm_electrode_design.ipynb)
+- [GITT fitting of an ECM for an LG M50](https://github.com/pybop-team/PyBOP/blob/develop/examples/notebooks/LG_M50_ECM/1-single-pulse-circuit-model.ipynb)
+- [Non-linear constrained ECM parameter identification](https://github.com/pybop-team/PyBOP/blob/develop/examples/notebooks/ecm_trust-constr.ipynb)
+- [Optimiser comparison for parameter identification](https://github.com/pybop-team/PyBOP/blob/develop/examples/notebooks/multi_optimiser_identification.ipynb)
+- [Parameter identification for spatial pouch cell model](https://github.com/pybop-team/PyBOP/blob/develop/examples/notebooks/pouch_cell_identification.ipynb)
+
+### Python Scripts
+
+Find additional script-based examples in the [examples directory](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/):
+
+- [UKF parameter identification (SPM)](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/spm_UKF.py)
+- [BPX format parameter import/export](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/BPX_spm.py)
+- [Electrochemical Impendence Spectroscopy (EIS) parameter identification](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/eis_fitting.py)
+- [Maximum a Posteriori parameter identification (SPM)](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/BPX_spm.py)
+- [Gradient-based parameter identification (SPM)](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/spm_AdamW.py)
 
-- [Unscented Kalman filter parameter identification of a SPM](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/spm_UKF.py)
-- [Import and export parameters using Faraday's BPX format](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/BPX_spm.py)
-- [Maximum a posteriori parameter identification of a SPM](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/BPX_spm.py)
-- [Gradient based parameter identification of a SPM](https://github.com/pybop-team/PyBOP/blob/develop/examples/scripts/spm_AdamW.py)
 
 ### Supported Methods
 The table below lists the currently supported [models](https://github.com/pybop-team/PyBOP/tree/develop/pybop/models), [optimisers](https://github.com/pybop-team/PyBOP/tree/develop/pybop/optimisers), and [cost functions](https://github.com/pybop-team/PyBOP/tree/develop/pybop/costs) in PyBOP.
 
 <p align="center">
 
-| Battery Models | Optimization Algorithms | Cost Functions |
-|-----------------------------------------------|-------------------------------------------------------------|------------------------------------------|
-| Single Particle Model (SPM) | Covariance Matrix Adaptation Evolution Strategy (CMA-ES) | Sum of Squared Errors (SSE) <tr></tr> |
-| Single Particle Model with Electrolyte (SPMe) | Particle Swarm Optimization (PSO) | Root Mean Squared Error (RMSE) <tr></tr> |
-| Doyle-Fuller-Newman (DFN) | Exponential Natural Evolution Strategy (xNES)  | Gaussian Log Likelihood <tr></tr> |
-| Many Particle Model (MPM) | Separable Natural Evolution Strategy (sNES) | Gaussian Log Likelihood w/ known variance <tr></tr> |
-| Multi-Species Multi-Reactants (MSMR) | Adaptive Moment Estimation with Weight Decay (AdamW) | Maximum a Posteriori (MAP) <tr></tr> |
-| Equivalent Circuit Models (ECM) | Improved Resilient Backpropagation (iRProp-) | Unscented Kalman Filter (UKF) <tr></tr> |
-| | SciPy Minimize & Differential Evolution | Gravimetric Energy Density <tr></tr> |
-| | Gradient Descent| Volumetric Energy Density<tr></tr> |
-| | Nelder-Mead | <tr></tr> |
+| Battery Models                                | Optimization Algorithms                                  | Cost Functions                           |
+|-----------------------------------------------|----------------------------------------------------------|------------------------------------------|
+| Single Particle Model (SPM)                   | Covariance Matrix Adaptation Evolution Strategy (CMA-ES) | Sum of Squared Errors (SSE) <tr></tr>    |
+| Single Particle Model with Electrolyte (SPMe) | Particle Swarm Optimization (PSO)                        | Root Mean Squared Error (RMSE) <tr></tr> |
+| Doyle-Fuller-Newman (DFN)                     | Exponential Natural Evolution Strategy (xNES)            | Minkowski <tr></tr>                      |
+| Many Particle Model (MPM)                     | Separable Natural Evolution Strategy (sNES)              | Sum of Power <tr></tr>                   |
+| Multi-Species Multi-Reactants (MSMR)          | Adaptive Moment Estimation with Weight Decay (AdamW)     | Gaussian Log Likelihood <tr></tr>        |
+| Weppner-Huggins                               | Improved Resilient Backpropagation (iRProp-)             | Log Posterior <tr></tr>                  |
+| Equivalent Circuit Models (ECM)               | SciPy Minimize & Differential Evolution                  | Unscented Kalman Filter (UKF) <tr></tr>  |
+|                                               | Cuckoo Search                                            | Gravimetric Energy Density <tr></tr>     |
+|                                               | Gradient Descent                                         | Volumetric Energy Density<tr></tr>       |
+|                                               | Nelder-Mead                                              | <tr></tr>                                |
 
 </p>
 
@@ -106,6 +112,10 @@ PyBOP aims to foster a broad consortium of developers and users, building on and
 
 -   User-friendliness (putting user requirements first via user-assistance & workflows)
 
+## License
+
+PyBOP is released under the [BSD 3-Clause License](https://github.com/pybop-team/PyBOP/blob/develop/LICENSE).
+
 ## Contributors ✨
 
 Thanks goes to these wonderful people ([emoji key](https://allcontributors.org/docs/en/emoji-key)):

docs/_static/switcher.json

@@ -4,9 +4,9 @@
       "url": "https://pybop-docs.readthedocs.io/en/latest/"
     },
     {
-      "name": "v24.9.0 (stable)",
-      "version": "v24.9.0",
-      "url": "https://pybop-docs.readthedocs.io/en/v24.9.0/",
+      "name": "v24.9 (stable)",
+      "version": "v24.9",
+      "url": "https://pybop-docs.readthedocs.io/en/v24.9.1/",
       "preferred": true
     },
     {

examples/standalone/cost.py

@@ -36,14 +36,14 @@ def __init__(self, problem=None):
         """
         super().__init__(problem)
 
-        self.parameters = pybop.Parameters(
+        self._parameters = pybop.Parameters(
             pybop.Parameter(
                 "x",
                 initial_value=4.2,
                 bounds=[-1, 10],
             ),
         )
-        self.x0 = self.parameters.initial_value()
+        self.x0 = self._parameters.initial_value()
 
     def compute(
         self, y: dict = None, dy: np.ndarray = None, calculate_grad: bool = False
@@ -59,4 +59,4 @@ def compute(
         float
             The calculated cost value for the given parameter.
         """
-        return self.parameters["x"].value ** 2 + 42
+        return self._parameters["x"].value ** 2 + 42

noxfile.py

@@ -14,7 +14,6 @@
 
 @nox.session
 def unit(session):
-    session.install("setuptools", "wheel")
     session.install("-e", ".[all,dev]", silent=False)
     if PYBOP_SCHEDULED:
         session.run("pip", "install", f"pybamm=={PYBAMM_VERSION}", silent=False)
@@ -23,7 +22,6 @@ def unit(session):
 
 @nox.session
 def coverage(session):
-    session.install("setuptools", "wheel")
     session.install("-e", ".[all,dev]", silent=False)
     session.install("pip")
     if PYBOP_SCHEDULED:
@@ -44,7 +42,6 @@ def coverage(session):
 @nox.session
 def plots(session):
     """Run the tests that generate plots."""
-    session.install("setuptools", "wheel")
     session.install("-e", ".[plot,dev]", silent=False)
     session.install("pip")
     session.run("pytest", "--plots", "-n", "0")
@@ -53,15 +50,13 @@ def plots(session):
 @nox.session
 def integration(session):
     """Run the integration tests."""
-    session.install("setuptools", "wheel")
     session.install("-e", ".[all,dev]", silent=False)
     session.run("pytest", "--integration")
 
 
 @nox.session
 def examples(session):
     """Run the examples and notebooks"""
-    session.install("setuptools", "wheel")
     session.install("-e", ".[all,dev]", silent=False)
     session.run("pytest", "--examples")
     notebooks(session)
@@ -70,7 +65,6 @@ def examples(session):
 @nox.session
 def notebooks(session):
     """Run the Jupyter notebooks."""
-    session.install("setuptools", "wheel")
     session.install("-e", ".[all,dev]", silent=False)
     if PYBOP_SCHEDULED:
         session.run("pip", "install", f"pybamm=={PYBAMM_VERSION}", silent=False)
@@ -86,7 +80,6 @@ def notebooks(session):
 @nox.session(name="notebooks-overwrite")
 def notebooks_overwrite(session):
     """Run the Jupyter notebooks."""
-    session.install("setuptools", "wheel")
     session.install("-e", ".[all,dev]", silent=False)
     if PYBOP_SCHEDULED:
         session.run("pip", "install", f"pybamm=={PYBAMM_VERSION}", silent=False)
@@ -103,7 +96,6 @@ def notebooks_overwrite(session):
 @nox.session(name="tests")
 def run_tests(session):
     """Run all the tests."""
-    session.install("setuptools", "wheel")
     session.install("-e", ".[all,dev]", silent=False)
     if PYBOP_SCHEDULED:
         session.run("pip", "install", f"pybamm=={PYBAMM_VERSION}", silent=False)
@@ -118,7 +110,6 @@ def run_doc_tests(session):
     Checks if the documentation can be built, runs any doctests (currently not
     used).
     """
-    session.install("setuptools", "wheel")
     session.install("-e", ".[plot,docs,dev]", silent=False)
     session.run("pytest", "--docs", "-n", "0")
 
@@ -130,7 +121,6 @@ def lint(session):
 
     Credit: PyBaMM Team
     """
-    session.install("setuptools", "wheel")
     session.install("pre-commit", silent=False)
     session.run("pre-commit", "run", "--all-files")
 
@@ -149,7 +139,6 @@ def run_quick(session):
 @nox.session
 def benchmarks(session):
     """Run the benchmarks."""
-    session.install("setuptools", "wheel")
     session.install("-e", ".[all,dev]", silent=False)
     session.install("asv[virtualenv]")
     session.run("asv", "run", "--show-stderr", "--python=same")
@@ -161,7 +150,6 @@ def docs(session):
     Build the documentation and load it in a browser tab, rebuilding on changes.
     Credit: PyBaMM Team
     """
-    session.install("setuptools", "wheel")
     envbindir = session.bin
     session.install("-e", ".[all,docs]", silent=False)
     session.chdir("docs")

pybop/costs/_likelihoods.py

@@ -111,9 +111,11 @@ def __init__(
         self._dsigma_scale = dsigma_scale
         self._logpi = -0.5 * self.n_data * np.log(2 * np.pi)
 
+        # Add sigma parameter, join with self.parameters, reapply transformations
         self.sigma = Parameters()
         self._add_sigma_parameters(sigma0)
-        self.parameters.join(self.sigma)
+        self.join_parameters(self.sigma)
+        self.transformation = self._parameters.construct_transformation()
 
     def _add_sigma_parameters(self, sigma0):
         sigma0 = [sigma0] if not isinstance(sigma0, list) else sigma0
@@ -235,11 +237,11 @@ def __init__(
 
         # Store the likelihood and prior
         self._log_likelihood = log_likelihood
-        self.parameters = self._log_likelihood.parameters
+        self._parameters = self._log_likelihood.parameters
         self._has_separable_problem = self._log_likelihood.has_separable_problem
 
         if log_prior is None:
-            self._prior = JointLogPrior(*self.parameters.priors())
+            self._prior = JointLogPrior(*self._parameters.priors())
         else:
             self._prior = log_prior
 
@@ -271,16 +273,16 @@ def compute(
 
         if calculate_grad:
             if isinstance(self._prior, BasePrior):
-                log_prior, dp = self._prior.logpdfS1(self.parameters.current_value())
+                log_prior, dp = self._prior.logpdfS1(self._parameters.current_value())
             else:
                 # Compute log prior first
-                log_prior = self._prior.logpdf(self.parameters.current_value())
+                log_prior = self._prior.logpdf(self._parameters.current_value())
 
                 # Compute a finite difference approximation of the gradient of the log prior
-                delta = self.parameters.initial_value() * self.gradient_step
+                delta = self._parameters.initial_value() * self.gradient_step
                 dp = []
 
-                for parameter, step_size in zip(self.parameters, delta):
+                for parameter, step_size in zip(self._parameters, delta):
                     param_value = parameter.value
                     upper_value = param_value * (1 + step_size)
                     lower_value = param_value * (1 - step_size)
@@ -293,7 +295,7 @@ def compute(
                     )
                     dp.append(gradient)
         else:
-            log_prior = self._prior.logpdf(self.parameters.current_value())
+            log_prior = self._prior.logpdf(self._parameters.current_value())
 
         if not np.isfinite(log_prior).any():
             return (-np.inf, -self.grad_fail) if calculate_grad else -np.inf

pybop/costs/_weighted_cost.py

@@ -62,7 +62,7 @@ def __init__(self, *costs, weights: Optional[list[float]] = None):
             super().__init__()
 
         for cost in self.costs:
-            self.parameters.join(cost.parameters)
+            self.join_parameters(cost.parameters)
 
         # Weighted costs do not use this functionality
         self._has_separable_problem = False

pybop/costs/base_cost.py

@@ -34,7 +34,7 @@ class BaseCost:
     """
 
     def __init__(self, problem: Optional[BaseProblem] = None):
-        self.parameters = Parameters()
+        self._parameters = Parameters()
         self.transformation = None
         self.problem = problem
         self.verbose = False
@@ -44,17 +44,17 @@ def __init__(self, problem: Optional[BaseProblem] = None):
         self._de = 1.0
         if isinstance(self.problem, BaseProblem):
             self._target = self.problem.target
-            self.parameters.join(self.problem.parameters)
+            self._parameters.join(self.problem.parameters)
             self.n_outputs = self.problem.n_outputs
             self.signal = self.problem.signal
-            self.transformation = self.parameters.construct_transformation()
+            self.transformation = self._parameters.construct_transformation()
             self._has_separable_problem = True
             self.grad_fail = None
             self.set_fail_gradient()
 
     @property
     def n_parameters(self):
-        return len(self.parameters)
+        return len(self._parameters)
 
     @property
     def has_separable_problem(self):
@@ -95,8 +95,8 @@ def __call__(
         self.has_transform = self.transformation is not None and apply_transform
         if self.has_transform:
             inputs = self.transformation.to_model(inputs)
-        inputs = self.parameters.verify(inputs)
-        self.parameters.update(values=list(inputs.values()))
+        inputs = self._parameters.verify(inputs)
+        self._parameters.update(values=list(inputs.values()))
 
         y, dy = None, None
         if self._has_separable_problem:
@@ -183,3 +183,16 @@ def verify_args(self, dy: ndarray, calculate_grad: bool):
             raise ValueError(
                 "Forward model sensitivities need to be provided alongside `calculate_grad=True` for `cost.compute`."
             )
+
+    def join_parameters(self, parameters):
+        """
+        Setter for joining parameters. This method sets the fail gradient if the join adds parameters.
+        """
+        original_n_params = self.n_parameters
+        self._parameters.join(parameters)
+        if original_n_params != self.n_parameters:
+            self.set_fail_gradient()
+
+    @property
+    def parameters(self):
+        return self._parameters