Python Pants - A build system for large codebases with multiple dependencies
summary
Pants is presented as a build system for large monorepo Python codebases that produces immutable PEX files containing all Python deps, demonstrated by building a Flask service that uses the third‑party textblob translator and running the single PEX binary in Docker. It enforces single-version dependency lists, runs tests across dependent targets, uses small build files beside source, and requires OS packages for native libs (so PEX does not bundle system C libraries). Companies cited: Twitter and Viki.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:13Outline
- 2:06Code Organization
- 9:12Benefits
- 10:08Pants history
- 11:07Pants basics
- 15:28Examples
- 16:10Example
- 19:03Running Pants
- 23:18Twitter Commons
related talks
24mDjango in Production with PEX
07mLightning Talk – Python Dependencies the Right Way
16mMigrating Twitter's monorepo from Pants to Bazel
36mPython Packaging from Init to Deploy
30mPython Containers: Best Practices
30mreliable testing with pip constraints
28mTesting Complex Python Dependencies With Docker Containers
23mReproducible Environments for Reproducible Results