scripts.data.preprocessing.packageTensors.dumpTrainingTensors

scripts.data.preprocessing.packageTensors.dumpTrainingTensors(in_path, out_path=None, cutoff=1000, save=True, coords_only=False, dummy_terms=None)[source]

Generate features from dTERMen .dat and .red.pdb, and dump the output into a file if requested.

Parameters:

  • in_path (str) – Prefix to .dat and .red.pdb files

  • out_path (str or None) – Prefix to the output .features and .length. Can be None if save=False.

  • cutoff (int, default=1000) – Max number of TERMs to featurize

  • save (bool, default=True) – Whether or not to save the training tensors

  • coords_only (bool, default=False) – Whether to use only backbone-derived features

  • dummy_terms (str or None) – Method by which to incorperate dummy TERMs. Options include 'replace', which means replacing TERM features with those derived from a dummy TERM, or 'include', which includes the dummy TERM into the mined TERM matches.

Returns:

Dictionary of features for TERMinator

Return type:

dict