Edit of sklearndf git_url so conda build works on windows #39
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
mgelsm
approved these changes
Oct 7, 2020
joerg-schneider
approved these changes
Oct 7, 2020
konst-int-i
added a commit
that referenced
this pull request
Jan 12, 2021
* improve to check for transfomer coverage * improve coverage test * paramterize coverage test * find pipeline classes on the basis of _BaseComposition * define UNSUPPORTED_SKLEARN_PACKAGES; skip cov. tests * add KNNImputerDF * add simple test for imputers * improved imputer support + add a test case * correctly support NaN fields in IterativeImputer * remove old "not for client use" comments * make sklearndf tests run again with scikit-learn 0.21 * implement LGBMClassifierDF alongside LGBMRegressorDF * update introductory documentation for sklearndf module * validate that all BaseEstimatorDF implementations subclass BaseEstimator * make definition of Module type more specific * Replace references to _BaseEstimatorWrapperDF with BaseEstimatorDF * Blend properties root_estimator & delegate_estimator to native_estimator * Improve type variable definition for T_DelegateLearner * invoke super().__init__() in base estimator wrapper * tweak BaseEstimatorDF API and update the docstrings * update type hinting in df_estimator * remove "Base" prefix from the names of most abstract sklearndf classes * update introductory documentation of module sklearndf * update docstrings and type hints for LearnerDF * Update wording in class docstrings of sklearndf base classes * unify terminology for inputs, outputs, and observations * update docstrings for TransformerDF * document return value for LearnerDF.predict * update docstrings for RegressorDF * update API and docstrings for ClassifierDF * update docstrings in _wrapper.py * update docstring for LearnerDF.score() * set version to 1.3.0rc1 * make sklearndf imputers work again for scikit-learn v0.21 * rename BaseEstimatorDF.root_estimator to native_estimator * use TypeVar to annotate missing _BaseImputer (avoids exception) * fix reference to scikit-learn version in a comment * move PyCharm settings to separate settings repository * rename gamma.sklearndf to sklearndf: all tests pass * no longer export T_Self in package common.fit * rename string/comment references to gamma.… to new facet counterparts * update setup.py and set version to 1.0 * move LGBMClassifier to sklearndf.classification.extra … from sklearndf.regression.extra * set to 1.0.0 * remove punch * un-ignore local PyCharm settings in git project repo * add PyCharm module definitions * flatten structure of 'pytools' module, eliminating 'common' sub-package * remove references to custom test data * add first version of Makefile & recipe for sklearndf * small bugfix in Makefile variable * mark /dist as excluded to prevent scanning * CONDA: update conda build path to replace spaces * added __all__ * added __all__ * added __all__ * added __all__ * added __all__ * added __all__ * added __all__ * added __all__ * start sphinx doc * start sphinx doc * start sphinx doc * start sphinx doc * Set up CI with Azure Pipelines [skip ci] * CI/CD: adds local pip installation to account for pytools dependency * CI/CD: adds pytest.ini and setup.cfg for coverage reports * CI/CD: update meta.yaml to take local path * CI/CD: udpates azure-pipelines.yml to build conda package * CI/CD: add coverage reports to azure-pipelines.yml * CI/CD: fix indent error * CI/CD: builds conda package dependency instead of pip install * CI/CD: clean up azure pipeline * DOCS: remove --implicit-namespace flag from Makefile * DOCS: add BCG copyright * CI/CD: divide azure-pipelines.yml into two dependent stages * CI/CD: update conda-build dependency * CI/CD: update minimum FACET_V_PYTHON_BUILD to v3.6 * CI/CD: build package only for sklearndf * CI/CD: update sklearn versions for sklearndf package build * CI/CD: update sklearn versions for sklearndf package build * CI/CD: update sklearn versions for sklearndf package build * CI/CD: update meta.yaml for new pytools version * CI/CD: update meta.yaml for new pytools version * CI/CD: update azure-pipelines.yml * CI/CD: update PYTHONPATHS in azure-pipelines.yml * CI/CD: update PYTHONPATHS in azure-pipelines.yml * CI/CD: checkout pytools * CI/CD: update test matrix * CI/CD: add less strict sklearn minimum dependency * CI/CD: lower max dependencies * CI/CD: lower max dependencies * using rst file instead * updates to doc layout * updates to doc layout * updates to doc layout * updates to doc layout * new readme using rst * place for images * API layout * contributions layout * tutorial layout * faq layout * getting started layout * script for notebook transforms * updating API layout * adding tutorial to examples layout * revised sklearn tutorial * BUILD: change to readme.rst in setup.py * removed m2r * removed text * removing m2r from extensions list * DOCS: update heading to sklearndf only * DOCS: add extensive contribution guide * DOCS: update matrix strategy in contribution_guide.rst * ENV: adds environment.yml for sklearndf * DOCS: update contribution_guide.rst * ENV: add dev-setup.sh for pre-commit hooks * DOCS: update env name in contribution_guide.rst * DOCS: remove redundant sphinx setup instructions * ENV: activate env in shell script * minor corrections and revisions * ENV: updates shebang line to use /sh instead of /bash * inherit learner pipeline docstrings from superclasses * document return value for method LearnerDF.fit_predict() * CI/CD: update repositories to checkout * use more meaningful names for the target variables used for unit tests (#16) * add convenience property `classes` to ClassifierDF (#17) * fix a typo in azure-pipelines.yml (#18) * Prevent sample weights from being passed to the preprocessing step of a learner pipeline (#19) * pass sample weights only to the final estimator of a 2-step pipeline * add docstring to ˋ_BaseEstimatorPipelineDF.fit()ˋ * clarify a docstring for new sample_weight behaviour Co-authored-by: Jason Bentley <Bentley.Jason@bcg.com> * update PyCharm settings (#20) * update python interpreter setting * ignore pytest cache folder * SPHINX: deactivate option show-inheritance; we generate this ourselves * SPHINX: no longer include native sklearn documentation in sklearndf docs … add references to sklearn docs instead * SPHINX: remove redundant __all__ statements from __init__.py files * DOC: tweak sklearndf class hierarchy to generate clean documentation * SPHINX: stop auto=generating API doc in the sphinx Makefiles * SPHINX: mark sphinx/build as excluded in PyCharm * SPHINX: use event callbacks from new pytools.sphinx package * DOC: apply corrections to docstrings * WRAPPER: fix typo in local variable name * fix return statement format in docstring * clarify scikit-learn error messages (#23) * catch and enhance sklearn exceptions to help locate errors in pipelines * clarify exception message for invalid ColumnTransformerDF parameters * rename classes whose names start with BaseEstimator… to Estimator… (#24) * rename classes whose names start with BaseEstimator… to Estimator… * adapt some comments to renamed class Co-authored-by: Jörg Schneider <schneider.joerg@bcg.com> * Docs/use py script for docs (#25) * used in worked example * used in worked example * used in worked example * used in worked example * so logo shows in readme * added cleanout of jupyter output to pre-commit * added getting started notebook * script for doc build * added metadata to cells for removal during transform * added metadata to cells for removal during transform * replaced by make.py * replaced my make.py * decided to retain NB output as it is minimal * also pass getting started NB to notebooks folder * updated with output * change reference to make.py * load Ames housing dataset from OpenML, rather than using a local copy Co-authored-by: Jan Ittner <ittner.jan@bcg.com> * append '_' to names of properties available only after fitting (#26) * Create LICENSE * Bugfix: rename _EstimatorPipelineDF.features_out to features_out_ (#28) * rename _EstimatorPipelineDF.features_out to features_out_ * make _EstimatorPipelineDF.fit kwargs non-positional * Bugfix: prevent duplicate occurrences of parameters in API documentation generated by sphinx (#27) * deactivate autodoc typehints option always_document_param_types * exclude notebooks folder from git: content is generated during build (#31) * reduce scope of tests in azure pipeline for develop branch (#32) * reduce scope of tests in azure pipeline for develop branch * update environment.yml * update package versions in conda build recipe * Share code for Sphinx make.py and conf.py across all repos (#30) This commit: * adds custom css and js for custom HTML styling and external link handling * replaces make.py with forwarder to pytools/sphinx/make.py (see shared config_base in pytools) and removes duplicate transform_notebook.py * updates package dependencies and ensures conda develop environment is created * This commit corrects a typo in a line of code in the tutorial notebook (#33) * bugfix: set sphinx parameters passed to set_config() (#35) * add custom `css` and `js` for custom HTML styling and external link handling * move `gamma.js` to `_static/js` * open all links in a new frame that are not known to be internal * replace make.py with forwarder to `pytools/sphinx/make.py` * remove the now duplicate `transform_notebook.py` from the sphinx scripts * move bulk of sphinx config to shared `config_base` in pytools * refactor `make.py` proxy script * add custom `css` and `js` files from within conf_base script * update package dependencies; ensure conda develop environment is created * replace `path.dirname(__file__)` with `path.dirname(path.realpath(__file__))` * bugfix: set sphinx parameters passed to `set_config()` * update conditional build logic depending on target branch and scheduling (#34) * replace wildcard imports with explicit import statements; upgrade black to v20.8b1 (#36) * replace wildcard imports with explicit import statements * upgrade black to v20.8b1 * Add code checks to the CI pipeline: isort, black, flake (#37) * add isort to pre-commit config * set CI checks (black, isort, flake8) * remove PEP8 errors, W293, E302, W291, F811, E121, W291, E126, E503, E122, E501 * clarify names of pipeline steps and jobs * add test folder to known local folders Co-authored-by: Jan Ittner <ittner.jan@bcg.com> * Update CI pipeline configuration (avoid black/flake8 conflicts; report failed unit test) (#38) * update errors ignored by flake8 to avoid conflicts with black * force pipeline to stop when a unit test fails * Remove forward slash at end of sklearndf git_url so build works on windows (#39) * Strengthen pre-commit checks (#40) * add flake8 config in pre-commit * add large file detector in pre-commit * add yaml files checker in pre-commit * update sklearndf sphinx build for new build process (#41) * update develop environment to ensure error-free documentation build * move conda_base and make to new directory sphinx/base * update sklearndf documentation for new build process * fix indentation of admonition in auto-generated rst docstring for scikit-learn estimators * align PyCharm folder exclusions with .gitignore * re-enable tutorial build * update nbsphinx to v0.7 for compatibility with nbconvert 6 * commit cell output from the jupyter notebook tutorial used in documentation * Update flake8 pre-commit to use local configs (#43) * fix a comment in function test_regression_pipeline_df (#47) * add initial gitbadges faq and correct typos (#45) This commit * adds initial gitbadge placeholders to the readme file and those needing to be updated upon release have been left blank to ensure they are noticeable for correction * adds initial material to the FAQ section of the documentation * Rename attributes 'feature_…_' to 'feature_names_…_' (#46) This commit * renames iris_target to iris_target_name * renames attributes 'feature_…_' to 'feature_names_…_' * adjusts property names of fittable classes to scikit-learn conventions * minor updates to tutorials and readme for api changes (#48) * replace deprecated imports from pandas.util.testing with pandas.testing (#50) * Add FACET team email in the sklearndf FAQ (#49) * Update PyCharm configuration, incl. newly introduced PyCharm features (#52) * Move to Python make script and link version strings (#51) * build: setup.py reference package's __version__ * build: move meta.yaml to condabuild/; reference setup.py version * build: add make.py for sklearndf package calling pytools' make script * build: remove bsd/gnu Makefile, use make.py in azure-pipelines.yml * build: add x flag to make.py * build: bugfix reference in setup.py * build: move package version to _version.py; to avoid ext. dependencies * build: name gamma_sklearndf_version __version__; load w/o ext. deps. * build: join path to version module platform independently * Fix the typo in "preprocessing" in the ReadMe (#53) * Apply ReadMe extensions from facet repo (#55) * simplify the logo banners on documentation landing pages (#56) * Feature/build/rework sklearndf build (#54) * change distribution/package name to sklearndf (#57) * update channel for conda install (#58) * minor readme updates to align formatting with facet readme (#59) * remove breaks in bullet list for proper rendering (#60) * build: set same pipeline triggers as used for pytools (#61) * update pypi and conda package build metadata (#62) * make sklearndf italic and remove index from doc build (#63) * make sklearndf italic * remove genindex * no longer include index * Bugfix/remove sgd regressor warning (#64) * bugfix add missing requirement packaging (#66) * add package badges to readme (#68) * add trove classifiers to pypi metadata (#65) * Add automatic docs deployment to GitHub Pages (#67) * add azure badge links and update conda install instructions (#69) * update notebook links for tutorials page in docs (#70) * remove local path setting from tutorial notebooks (#74) * implement flexible get started (#75) * add nb download icon (#76) * replace raw html (#78) * revise doc build documentation (#77) * GitHub release pipeline (#73) * CI/CD: Adapt release pipeline from pytools * CI/CD: Add atrifact storage to publish in release * CI/CD: Fix path for conda/tox builds * CI/CD: Use fork as target for testing release * CI/CD: Use existing target in test repo for release test (instead of triggering repo) * CI/CD: Fix path of artifacts to upload with release * CI/CD: Remove test settings * CI/CD: Reset single build to test 'max' dependencies * CI/CD: Fix format in config file * CI/CD: Use OAuth service connection to GitHub as required by Azure Pipelines * CI/CD: Let release pipeline fail when line command fails * CI/CD: Use bcg_gamma channel for conda upgrade instructions in release notes * reformat api preamble (#79) * add API landing rst to documentation guide (#80) * updates to development guidelines (#81) * fix docstrings (#82) * remove trailing backslashes in docstrings: not needed in canonical ReST * fix spelling * remove trailing backslashes in docstrings: not needed in canonical ReST (#83) * remove trailing backslashes in docstrings: not needed in canonical ReST * fix spelling * minor code tweaks to eliminate IDE warnings (#84) * Doc/tabs to spaces (#85) * Update issue templates (#86) * doc cleanup (#88) * rename type variables for methods' "self" arguments to T_Self (#89) * add 'return' after referencing an estimator method from a DF class (#90) * fix sklearn links (#91) * Bugfix/pipeline fixes (#94) * restore unconstr build (#95) * Documentation improvements (#92) * add missing documentation for sklearndf class attributes * return two redundant sub-methods that mess up the sphinx documentation * introduce @inheritdoc for class OutlierRemoverDF * improve code and docs of class OutlierRemoverDF (no functional changes) * Improve docstrings for COL_FEATURE_IN and COL_FEATURE_OUT * introduce AllTracker across all of sklearndf * remove badges * assign wrapper's module when wrapping a native sklearn estimator * assign wrapper's module when wrapping a native sklearn estimator (bugfix) * assign wrapper's module when wrapping native sklearn estimator (bugfix) * make detection of sklearn class tag lines more robust * fix incomplete cross-references * update doc_url * updates for pypi and conda package pages * updates for pypi and conda package pages * point docs to index * bump initial release version * update docstring Co-authored-by: Jason Bentley <Bentley.Jason@bcg.com> * Bugfix/python3.9 compatibility (#96) * add missing documentation for sklearndf class attributes * return two redundant sub-methods that mess up the sphinx documentation * introduce @inheritdoc for class OutlierRemoverDF * improve code and docs of class OutlierRemoverDF (no functional changes) * Improve docstrings for COL_FEATURE_IN and COL_FEATURE_OUT * introduce AllTracker across all of sklearndf * remove badges * assign wrapper's module when wrapping a native sklearn estimator * assign wrapper's module when wrapping a native sklearn estimator (bugfix) * assign wrapper's module when wrapping native sklearn estimator (bugfix) * make detection of sklearn class tag lines more robust * fix incomplete cross-references * do not pass arguments on to object.__new__() Co-authored-by: Jason Bentley <Bentley.Jason@bcg.com> * import from numpy.testing instead of sklearn.utils.testing (#98) * sklearn 0.24 currently not supported (#99) * disallow scipy >=1.6, allow pandas 1.2.*, prefer joblib 1.0.* (#100) * disallow scipy >=1.6, allow pandas 1.2.*, prefer joblib 1.0.* * set pandas to 1.1.0 in 'max' matrix test for compatibility with pytools * CI/CD: roll back joblib test dependency to 0.16 Co-authored-by: Konstantin Hemker <hemker.konstantin@bcg.com> Co-authored-by: Jörg Schneider <schneider.joerg@bcg.com> Co-authored-by: Jan Ittner <ittner.jan@bcg.com> Co-authored-by: Jörg Schneider <46053259+joerg-schneider@users.noreply.github.com> Co-authored-by: Konstantin Hemker <hemker.konstantin@bcg.com> Co-authored-by: Jason Bentley <Bentley.Jason@bcg.com> Co-authored-by: Jan Ittner <GitHub.com@ittner.de> Co-authored-by: Kennedy Ricardo <Kennedy.Ricardo@bcg.com> Co-authored-by: Jason <jasonpben@gmail.com> Co-authored-by: mgelsm <malo.grisard@pm.me> Co-authored-by: mgelsm <grisard.malo@bcg.com> Co-authored-by: sithankanna <sithankanna@gmail.com>
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.
PR
This PR removes the forward slash at the end of the git_url in the
conda-build/meta.ymlso that the build via make.py will work as intended on a windows machine.