Installation¶

The package¶

Requires Python 3.12 or newer.

git clone https://github.com/talmolab/STRIDE.git
cd STRIDE
pip install -e .            # or: uv pip install -e .
pip install -e ".[dev]"     # adds pytest, needed for the mock-data quick starts

This installs the stride command and the stage modules. It is everything the analysis needs: decisions, gait, scale calibration, and the optional statistics run on a plain CPU.

The inference runtime (only when STRIDE runs your models)¶

STRIDE runs your SLEAP models through sleap-nn (PyTorch; a GPU in practice). Install it separately, following its own instructions, in the same environment as STRIDE. It is needed only for the two inference steps (pose, ROI); the analysis of existing pose and ROI files does not use it. sleap-io, which is installed with STRIDE, only reads and writes .slp files.

Trained models are not bundled

STRIDE ships no trained models. Point it at your own SLEAP pose and ROI models (--pose-model, --roi-model, or the config). What the models must provide is listed in Inputs and file layout. The trained weights of the original study are to be deposited (DOI to be added).

OpenCV version¶

The package pins opencv-python >= 4.8, < 4.14. From 4.14 on (and in 5.x) the ChArUco detector reads the same board about 0.1 % lower, so the pin keeps the scale measurement the same for everyone (Scale calibration).

Operating systems¶

	Linux	macOS	Windows
Analysis on synthetic data (the test suite, run for every pull request)	tested	tested	tested
Inference with SLEAP models	run on NVIDIA GPUs	not run	not run

The code uses no shell commands and no links, writes UTF-8 files, matches video extensions case-insensitively and starts its workers under spawn.

Check the install¶

stride --version
stride --help
pytest tests            # synthetic data only; no videos, models or GPU needed

The authors' lab layer (optional)¶

pip install -e lab/ adds the stride-lab command described in Authors' lab workflow. It is not needed to use STRIDE.