A fast, distributed, high performance gradient boosting (GBT, GBDT, GBRT, GBM or MART) framework based on decision tree algorithms, used for ranking, classification and many other machine learning tasks.
Proposes adding numpydoc-validation and pydoclint, 2 linters that enforce standardization and correctness of Python docstrings. These caught a few issues (like eval_{X,y} missing from the Dask docstrings.
Both are fast enough to not be noticeable.
I think they'll help a lot with consistency and standardization, and relieve a small bit of review burden.
Notes for Reviewers
I tried to choose configurations for these tools that as closely as possible match the project's current conventions. See inline comments below.
I've added this branch (temporarily) to readthedocs configuration.
microsoft/lightgbm