Bonoboopen sourcearchived

A lightweight extract-transform-load framework for Python, data-processing for humans.

Bonobo describes data pipelines as plain Python functions chained into a directed graph, then runs them in parallel. You write small, pure transformations that are easy to unit-test, and the framework streams your data through them. No heavy runtime, no cluster to babysit.

It began as a rewrite of an older Python 2.7 tool that ran millions of transformations a day in production. It never reached 1.0 and is no longer maintained; the package and these notes stay online for the people still running it. There are more notes in the wiki.