lisbet.hub#

Functions

fetch_dataset(dataset_id, download_path)

Download and preprocess keypoints datasets from remote repositories.

fetch_model(model_id[, download_path])

Fetch a model from the HF Hub.
lisbet.hub.fetch_dataset(dataset_id, download_path)[source]#

Download and preprocess keypoints datasets from remote repositories.

Downloads the specified dataset, processes raw data (e.g., keypoints, annotations), and stores them in a standardized format for analysis.

Parameters:

  • dataset_id (str) – Identifier for the dataset to fetch. Currently supported datasets: - “CalMS21_Task1”: Mouse behavior classification dataset - “CalMS21_Unlabeled”: Unlabeled mouse behavior videos - “SampleData”: Sample dataset for testing Additional datasets may be supported in future versions.

  • download_path (str) – Base directory path where the dataset will be stored. The function creates subdirectories for cache and processed data.

Returns:

Data is saved to disk in standardized format.

Return type:

None

Raises:

ValueError – If dataset_id is not one of the supported options.

Notes

The function handles downloads with checksums and caching using pooch. Downloaded data is temporarily stored in a cache directory before being processed into the final standardized format.

lisbet.hub.fetch_model(model_id, download_path=PosixPath('.'))[source]#

Fetch a model from the HF Hub.