tabular-trees is a package for making analysis on tree-based models easier.
Tree based models (specifically GBMs) from xgboost, lightgbm or scikit-learn can be exported to TabularTrees objects for further analysis.
The explain and validate modules contain functions that operate on TabularTrees objects.
See the documentation for more information.
The easiest way to get tabular-trees is to install directly from pypi:
pip install tabular_trees
tabular-trees works with GBMs from xgboost, lightgbm or scikit-learn. These packages must be installed to use the relevant functionality from tabular-trees.
[lightgbm, sklearn, xgboost] are optional depedencies that can be specified for tabular-trees. They can be installed along with tabular-trees as follows:
pip install tabular_trees[lightgbm, sklearn]
tabular-trees uses poetry as the environment management and package build tool. Follow the instructions here to install.
To install the package locally, for development purposes along with the development dependencies run:
poetry install --with dev
dev is an optional dependency group, the other one is docs which is only required if building the documentation.
To install all the optional, development dependencies as well as all the extras for the package run:
poetry install --extras "lightgbm xgboost" --with dev,docs