Skip to content

quintoandar/butterfree

Repository files navigation

Butterfree

A tool for building feature stores. Transform your raw data into beautiful features.

Release Python Version License Code style: black

Source Downloads Page Installation Command
PyPi PyPi Downloads Link pip install butterfree

Build status

Develop Stable Documentation Sonar
Test Publish Documentation Status Quality Gate Status

Made with ❤️ by the MLOps team from QuintoAndar

This library supports Python version 3.7+ and meant to provide tools for building ETL pipelines for Feature Stores using Apache Spark.

The library is centered on the following concetps:

  • ETL: central framework to create data pipelines. Spark-based Extract, Transform and Load modules ready to use.
  • Declarative Feature Engineering: care about what you want to compute and not how to code it.
  • Feature Store Modeling: the library easily provides everything you need to process and load data to your Feature Store.

To understand the main concepts of Feature Store modeling and library main features you can check Butterfree's Documentation, which is hosted by Read the Docs.

To learn how to use Butterfree in practice, see Butterfree's notebook examples

Requirements and Installation

Butterfree depends on Python 3.7+ and it is Spark 3.0 ready ✔️

PyPI hosts reference to a pip-installable module of this library, using it is as straightforward as including it on your project's requirements.

pip install butterfree

Or after listing butterfree in your requirements.txt file:

pip install -r requirements.txt

Dev Package are available for testing using the .devN versions of the Butterfree on PyPi.

License

Apache License 2.0

Contributing

All contributions are welcome! Feel free to open Pull Requests. Check the development and contributing guidelines described here.