Provide Spotify specific TensorFlow helpers.

• tf.data integration with Featran
• common Dataset API to read:
  • TFRecord datasets as tf.Tensor
  • TFRecord datasets as Pandas DataFrame
  • TFRecord datasets as python dict
• tf.Graph freezing utilities
• TensorFlow integration with Luigi

Check examples in the examples directory, currently they include:

• sklearn training via Pandas DataFrame dataset
• XGBoost training via tf.Example dataset
• TensorFlow training via tf.Example dataset

To run the examples:

$ pip install -e .[examples]
$ bin/run-examples

This project uses tox.

$ pip install tox

To see all testenv's:

$ tox -l
mypy
lint
examples-py27
examples-py35
test-py27
test-py35
upload-coverage
license

To run the tests:

tox -e test

To release:

git commit --allow-empty -m "Release x.y.z"
git tag x.y.z
git push --tags  origin master

Then upload to pypi:

python setup.py sdist upload -r pypi