Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
70 lines (48 loc) · 1.94 KB

CONTRIBUTING.md

File metadata and controls

70 lines (48 loc) · 1.94 KB

Contributing

Thanks for helping improve lerobot-go.

Development setup

git clone https://github.com/ioai-tech/lerobot-go.git
cd lerobot-go
go test ./... -count=1

Requires Go 1.26+ (see go.mod).

Before opening a PR

  1. Run tests: make test
  2. Run linter: make lint — install the official binary v2.9.0+ (must be built with Go ≥ go.mod; do not use go install)
  3. Keep changes focused; match existing naming and package layout
  4. Public API changes belong in lerobot/; CLI wiring in internal/cli/

Commit messages

Use clear, imperative subjects. Examples:

  • fix merge stats aggregation for v3 episode meta
  • docs: add create command to CLI reference

End-to-end tests

Heavy dataset tests are optional locally:

make e2e

CI runs unit tests only.

Official LeRobot compatibility smoke checks

For higher-confidence stability validation against the official Python loader:

  1. Install the official Python runtime into an isolated prefix (one workable approach on this VM was pip --prefix, since uv was unavailable).
  2. Download authoritative sample datasets such as:
    • lerobot/pusht
    • lerobot/pusht_image
    • optionally nvidia/LIBERO_LeRobot_v3 (for example libero_10/**)
  3. Run:
make official-stability

The script scripts/run_official_stability_checks.sh:

  • loads official datasets with the official lerobot package,
  • exercises Go convert v3→v2.1→v3 round-trips,
  • exercises Go merge on official sample datasets,
  • re-loads the resulting v3 outputs with the official lerobot package.

Useful environment overrides:

export LEROBOT_OFFICIAL_PY_PREFIX=/path/to/prefix/local
export LEROBOT_OFFICIAL_DATA_ROOT=/path/to/downloaded/datasets
export LEROBOT_OFFICIAL_WORK_ROOT=/tmp/lerobot-go-official-checks

Questions

Open a GitHub issue for bugs or feature requests.