loki2.mil.prampt_generate.blca_metastasis_prampt

Generate PRAMPT data for BLCA metastasis analysis.

This script loads cell embeddings from .pt files, samples a subset of cells, and saves the merged embeddings as a parquet file for MIL training.

Module Contents

loki2.mil.prampt_generate.blca_metastasis_prampt.setup_dummy_modules() → None

Sets up dummy modules for ‘loki2’ and its submodules.

This is required to simulate the module structure that the torch model expects when unpickling objects that reference e.g. ‘loki2.data.dataclass.cell_graph’.

loki2.mil.prampt_generate.blca_metastasis_prampt.load_tensor(file_path: str | pathlib.Path) → Any

Load torch object from .pt file with dummy loki2 module support.

Sets up dummy loki2 modules before loading to handle unpickling of objects that reference loki2.data.dataclass.cell_graph classes.

Parameters:

file_path – Path to the .pt file to load.

Returns:

Loaded torch object (typically containing cell embeddings and positions).

loki2.mil.prampt_generate.blca_metastasis_prampt.directory_path = './data/TCGA/BLCA'

loki2.mil.prampt_generate.blca_metastasis_prampt.meta_path = './data/loki2_MIL/data/downsamples/blca_metastasis/BLCA_META.txt'

loki2.mil.prampt_generate.blca_metastasis_prampt.sample_ratio = 0.1

loki2.mil.prampt_generate.blca_metastasis_prampt.meta_df

loki2.mil.prampt_generate.blca_metastasis_prampt.patient_embeddings

loki2.mil.prampt_generate.blca_metastasis_prampt.patient_id

loki2.mil.prampt_generate.blca_metastasis_prampt.records = []

loki2.mil.prampt_generate.blca_metastasis_prampt.combined_df

loki2.mil.prampt_generate.blca_metastasis_prampt.merged_df