1.1 KiB
1.1 KiB
Ray
Ray is a unified framework for scaling AI and Python applications.
Building & Linting
Refer to doc/source/ray-contribute/development.md for the most up-to date information. Use the /rebuild skill for guided rebuilds and the /lint skill for linting.
Testing
See per-library .claude/CLAUDE.md for library-specific test paths. Default test timeout: 180s (pytest.ini).
Structure
- src/ray/ — C++ core runtime
- python/ray/ — Python API and AI libraries (data, serve, train, tune)
- rllib/ — RLlib (symlinked from python/ray/rllib)
- doc/source/ — Sphinx documentation
Development Workflow
- After making C++ or Cython changes, use /rebuild to determine the correct build steps
- After making code changes, use /lint to fix lint issues on modified files
- When investigating CI failures, use /fetch-buildkite-logs to fetch and analyze build logs
Contributing
See @CONTRIBUTING.rst and @doc/source/ray-contribute/getting-involved.md