Transform distributions

[NicolaCourtier]

Description

The implementation of parameters, distributions and transformations is over-complicated. The aim here is to simplify the definitions and interactions between these classes to improve maintainability. A new UnboundedDistribution is created to define the behaviour of a parameter without a statistical distribution. The get_transformed_distribution functionality is added to both univariate and multivariate distributions.

Type of change

Please add a line in the relevant section of CHANGELOG.md to document the change (include PR #).

Important checks:

Please confirm the following before marking the PR as ready for review:

• No style issues: $ pre-commit run or $ nox -s pre-commit (see CONTRIBUTING.md for how to set this up to run automatically when committing locally, in just two lines of code)
• All tests pass: nox -s tests
• The documentation builds: nox -s doctest
• Code is commented for hard-to-understand areas
• Tests added that prove fix is effective or that feature works

[codecov]

Codecov Report

❌ Patch coverage is 89.20635% with 34 lines in your changes missing coverage. Please review.
✅ Project coverage is 90.24%. Comparing base (4628a65) to head (183a238).
⚠️ Report is 53 commits behind head on develop.

Files with missing lines	Patch %	Lines
pybop/parameters/distributions.py	88.88%	16 Missing ⚠️
pybop/parameters/multivariate_distributions.py	85.54%	12 Missing ⚠️
pybop/parameters/parameter.py	92.00%	6 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #925      +/-   ##
===========================================
+ Coverage    89.41%   90.24%   +0.83%     
===========================================
  Files           63       73      +10     
  Lines         4857     5692     +835     
===========================================
+ Hits          4343     5137     +794     
- Misses         514      555      +41     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[martinjrobins]

looks good, couple of questions around distribution = None / unbounded

[martinjrobins]

what does distribution=None mean? unbounded? need to add to the docstring

[NicolaCourtier]

To make things clearer, I have added the BaseDistribution class - which can be used as a placeholder for an empty parameter.

[martinjrobins]

again, what is distribution=None mean?

[NicolaCourtier]

Added to the description:

distribution : BaseDistribution, optional
        Probability distribution for the parameter. If None, an empty
        `pybop.BaseDistribution` will return a NotImplementedError for any
        functionality that requires a distribution, such as `rvs`.

[martinjrobins]

but what does it mean from a user perspective? Does it mean that the parameter is unbounded? Or that it can't be used in sampling, or multi-start optimisation?

[NicolaCourtier]

If no distribution, bounds or initial value is specified, an empty Parameter with an empty BaseDistribution will be created. It doesn't really mean anything... but it should return a somewhat helpful error when something is required but not implemented. It has no bounds and yes, it cannot be used for sampling (no mean or cov) or multi-start optimisation (no rvs).