wojciechmigda / pytsetlini

Scikit-learn-compatible python3 wrapper for Tsetlini - a Tsetlin Machine learning model

Based on cython setup template from https://github.com/Technologicat/setup-template-cython

Usage as usual with setuptools:

python3 setup.py build_ext
python3 setup.py build
python3 setup.py install
python3 setup.py sdist
python3 setup.py pytest

For details, see http://setuptools.readthedocs.io/en/latest/setuptools.html#command-reference or

python3 setup.py --help
python3 setup.py --help-commands
python3 setup.py --help bdist_wheel  # or any command

pytsetlini's Issues

Release v0.0.2

Pull in new Tsetlini release.

Rename sklearn.py to avoid name clash with sklearn package

Re-enable python3.8 in CI Travis pipeline

Enable continuous integration with Travis

Release v0.0.4

Pull in v0.0.4 pre-release of Tsetlini.

Install pandas in Travis pipeline

scikit-learn warns that pandas is not installed and not all estimator checks are executed.
This is not required by the estimator per se, but will improve CI verification.

Relax scikit-learn version restriction in requirements.txt

Write proper README

Add ability to instantaneously kill running estimator by spawning it in a separate job

Release v0.0.3

Pull in v.0.0.3 pre-release of Tsetlini.

Add Tsetlin regressor

Release v0.0.5

Pull in Tsetlini release v0.0.5

Create release script for embedding libtsetlini code

Add estimator check for NaNs and Infs in input X matrix

With newer scikit-learn estimator checks assert that the estimator will throw ValueError if there are NaNs or Infs in input X matrix.
Once this check is added a version of scikit-learn in requirements.txt should be relaxed.

Create Regressor target transformer

and replace MinMaxScaler.

Remove scikit-learn warning

  /usr/local/lib/python3.6/dist-packages/sklearn/utils/estimator_checks.py:488: FutureWarning: Passing a class is deprecated since version 0.23 and won't be supported in 0.24.Please pass an instance instead.
    warnings.warn(msg, FutureWarning)