BUILD: Release facet 1.1.0rc0 #269
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
rename branch 1.1 to 1.1.x
* FIX: drop support for joblib v0.13 * streamline package requirements * update version numbers
# Conflicts: # pyproject.toml # sphinx/source/faqs.rst # src/facet/__init__.py
# Conflicts: # pyproject.toml
# Conflicts: # pyproject.toml
* increase minimum pytools and sklearndf version to 1.1
# Conflicts: # pyproject.toml
) * add 'legacy' parameter to LearnerInspector * replace SHAP vector decomposition with SHAP vector projection * add inspector attribute 'orthogonalize' * allow switching from/to legacy global explanations for experimentation * allow switching from/to orthogonalization for experimentation * cluster/order affinity matrices based on symmetrical, relative affinity * pass feature importance to class AffinityMatrices, not the SHAP variance * avoid NaN results for zero-variance SHAP vectors * use relative (not absolute) feature importance to get absolute affinity * test inspectors for legacy and non-legacy algorithms Co-authored-by: mgelsm <grisard.malo@bcg.com>
* FIX: make documentation test work again * move docstrings tests to pytools * run docstring tests using the pytools DocValidator * update pytools API and API docs so the doc tests pass * set working directory for tox pytest to project root * set working directory for conda pytest to project root * rename validate_docstrings() to validate_doc()
# Conflicts: # azure-pipelines.yml # pyproject.toml # src/facet/__init__.py
* move facet.simulation global constants into class definitions * no longer export the type aliases defined in facet.crossfit * require sklearndf v1.1.0rc0 or later
# Conflicts: # pyproject.toml # src/facet/__init__.py
* calculate synergy first, only then redundancy as r^2 of autonomy vectors * do not average SHAP values across splits for increased accuracy * update expected results in learner inspector unit tests * scale redundancy to remaining variance not already explained by synergy * remove LearnerInspector parameter 'orthogonalize': always orthogonalize the interaction vector * do not calculate SHAP legacy global explanations unless legacy=True * factor out SHAP context creation before global explanation calculations * store affinity matrices as one ndarray, not as nested tuples * prevent division by zero for relative symmetry of 0-importance features * calculate S/R/A matrices separately per split, then get mean and std * add release notes for FACET 1.1.0
…263) * API: move package facet.simulation.partition to facet.data.partition * update imports in notebooks * update imports in README.rst/; sphinx/aux Co-authored-by: Jörg Schneider <schneider.joerg@bcg.com>
* API: add support for shap v0.39 * relax shap version requirement for max build to >=0.37,<0.40 * remove test for supported shap versions: this is already covered by the package requirements
|
Docs issue addressed with: BCG-X-Official/pytools#215 |
joerg-schneider
approved these changes
Mar 25, 2021
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.