Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
41 lines (25 loc) · 1.54 KB

CONTRIBUTING.md

File metadata and controls

41 lines (25 loc) · 1.54 KB

Contributing

We want this project to stay trustworthy research OSS: small, reviewable diffs; runnable evaluation hooks; and honest docs. Improvements that tighten reproducibility (tests, check_regression_gate, build_leaderboard) or clarify behavior usually beat large refactors with no measurement.

Quick start

  1. Build (with tests):

    cmake -S . -B build -G Ninja -DBUILD_TESTS=ON
cmake --build build
ctest --test-dir build --output-on-failure

  2. Style: match existing C++ (snake_case files, PascalCase classes, English comments where practical). Keep diffs focused on one concern.

  3. Regression checks (requires local TUM data under data/tum/ and evo_ape on PATH):

    python3 scripts/check_regression_gate.py --all-gates --quiet

    Comparison baseline (single preset, fast sanity vs external OSS prep):

    bash scripts/verify_comparison_benchmark.sh xyz_depth

    This runs every entry in eval/regression_baselines.json (several TUM windows; full run is often ~10+ minutes locally). Adjust ceilings in that file only when a change intentionally shifts accuracy (document why in the commit message).

What not to commit

  • data/ (datasets), build/, eval_results/, large artifacts — see .gitignore.

Maintainers

Release and version bumps: RELEASING.md. User-facing changes: CHANGELOG.md.

License

By contributing, you agree your contributions are under the same BSD-2-Clause terms as LICENSE.