build_decaf_model_from_h5

soda_analyzer.build_decaf_model_from_h5(h5_lib, embed_str=None, cluster_str=None, bandwidth_estimate=None)

Reconstruct a DECAF model from the persisted descriptor data.

Parameters:

  • h5_lib (h5py.File) – Open SODA HDF5 file. Keep it open while using views derived from it.

  • embed_str (str, optional) – DECAF embedding method; None uses persisted model metadata. Default: None.

  • cluster_str (str, optional) – DECAF clustering method; None uses persisted model metadata. Default: None.

  • bandwidth_estimate (str, optional) – DECAF bandwidth estimator; None uses persisted model metadata. Default: None.

Returns:

object – DECAF embed_cluster instance. Requires the optional decaf package; missing training rows raise ValueError.

Notes

This function only depends on: - h5py handle (h5_lib) - decaf package - data persisted in the HDF5 file