pachypy - A Python client library for Pachyderm

This package aims to make interactions with Pachyderm more pythonic and user-friendly.

It is primarily intended to be used interactively in a Jupyter notebook environment.

Key features include:

• Show Pachyderm objects (such as repositories, pipelines, jobs, etc.) as nicely formatted tables, backed by pandas DataFrames
• Batch operations using shell-style wildcards on Pachyderm objects
• Create and update pipelines from specifications in YAML format, supporting multiple pipelines per file
• Build and push Docker images before creating and updating pipelines
• Add image digests when creating and updating pipelines to ensure the latest images are used in pipelines without requiring a tag change

pachypy uses the protobufs from the official python_pachyderm package to communicate with Pachyderm.

Requires Python 3.6 or higher

Installation

pip install pachypy

Documentation

• Example notebook
• Example pipeline specifications in YAML format
• API reference

Getting help

Let's chat on Gitter or open an issue on GitHub.