Assemblies

Assemblies load and expose dataset-aligned structures used by the trainer:

  • stories: ordered identifiers

  • get_brain_data(): list of arrays per story

  • get_data_times(), get_tr_times(): timestamp arrays

  • get_split_indices(): train/test folds or concatenation boundaries

Modules

  • encoding.assembly.assemblies: user-facing assembly types

  • encoding.assembly.base_processor: common I/O, validation, contracts

  • encoding.assembly.narratives_processor | lpp_processor | lebel_processor: dataset-specific logic

  • encoding.assembly.assembly_loader: loading helpers

  • encoding.assembly.assembly_generator: composite assembly creation

Typical usage

from encoding.assembly.assemblies import NarrativesAssembly
assembly = NarrativesAssembly(assembly_path=<path_to_narratives.h5>)
print(len(assembly.stories))
print(assembly.get_tr_times()[0].shape)