Installation

Requirements

• Python ≥ 3.10
• FFmpeg and ffprobe on your PATH — used for decoding and sampling frames. Verify with ffmpeg -version. Install via your platform's package manager (brew install ffmpeg, apt install ffmpeg, conda install -c conda-forge ffmpeg).
• A working uv is recommended (Astral's package manager). Plain pip also works.

Install with uv

From a clone of the repository:

git clone https://github.com/fmi-basel/gfelsenb-anytrack.git
cd gfelsenb-anytrack
uv venv
uv pip install -e .

The console scripts (anytrack, anytrack-run, …) are installed into the venv. Activate it (source .venv/bin/activate) or prefix commands with uv run, e.g. uv run anytrack run --video ....

Optional extras

Extra	Install	Adds
pose	uv pip install -e '.[pose]'	sleap-io — the SLEAP label/export bridge (Milestone B).
sam	uv pip install -e '.[sam]'	segment-anything — the optional SAM odor-port backend.
docs	uv pip install -e '.[docs]'	mkdocs-material — build/preview this documentation.

Pose backend

The pose training/inference backend (sleap-nn + torch) is not declared as a dependency: it is prerelease and needs Python ≥ 3.11, which would constrain the base install. Install it directly into your environment when you need it:

uv pip install --prerelease=allow 'sleap-nn[torch]'

Verify

uv run anytrack            # lists subcommands
uv run anytrack-run --help # options for the main pipeline
uv run python -m pytest    # run the test suite (use python -m pytest, not pytest)