umpirix/CONTRIBUTING.md
caelum0x 4847e35451 OpenJev: open-source typed decision service
Typed decision/classification service with an in-process gte-small ONNX
embedding engine and lexical fallback, JS/Python SDKs, adapters, and the
otsone model family. Apache-2.0 (sub-packages keep their MIT licenses).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-25 00:12:36 +03:00

2.8 KiB

Contributing to OpenJev

Thanks for your interest in OpenJev. Bug reports, fixes, new public datasets for the leaderboard, engine improvements, docs and SDKs for more languages are all welcome.

Before you start

  • For anything larger than a small fix, open an issue first so we can agree on the approach.
  • Security problems: do not open a public issue. Follow SECURITY.md.
  • By participating you agree to the Code of Conduct.

Development setup

Requirements: Node.js 22+, npm. Python 3.11+ and uv only if you work on otsone/, sdk/python/ or adapters/python/. Docker is optional.

git clone https://github.com/caelum0x/openjev && cd openjev
npm install
npm run embed-download      # optional: the gte-small model for the embedding engine
npm run dev                 # http://localhost:8080

No API keys, database or model weights are needed to run the server or the tests. Copy .env.example to .env only for the optional integrations it documents.

Checks to run before opening a pull request

Area you changed Command
server / console (src/) npm test && npm run gold && npm run typecheck && npm run lint
native engine (engine/) make -C engine
sdk/js cd sdk/js && npm install && npm test
sdk/js-lite cd sdk/js-lite && npm test
sdk/python cd sdk/python && uv sync && uv run pytest
adapters/python cd adapters/python && uv sync && uv run pytest
skills cd skills && npm test
otsone cd otsone && uv sync --extra dev && uv run pytest

If you change the engine's scoring, also run npm run leaderboard and include the before/after numbers in the pull request. Accuracy claims in docs must come from a reproducible run, never estimates.

Guidelines

  • Keep pull requests focused; one logical change per PR.
  • Add or update tests for behaviour changes.
  • Use Conventional Commits style messages (feat:, fix:, docs:, refactor:, test:, chore:, perf:).
  • Never commit secrets, .env files, model weights (*.onnx, *.safetensors, ...) or downloaded datasets; .gitignore covers the usual paths. See docs/WEIGHTS.md for how weights are fetched.
  • Code copied from other projects must be under a compatible license (Apache-2.0, MIT, BSD) and attributed in the relevant NOTICE file.

Licensing of contributions

Unless you state otherwise, contributions to Apache-2.0 parts of the repository are licensed under Apache-2.0, and contributions to the MIT-licensed directories (sdk/js, sdk/python, adapters/python, skills) under MIT, as described in section 5 of the Apache License. See NOTICE.