Datasets registry

Dataset classes buildable under data.train / data.val / data.test. See ../../data/index.

Build any of these from a config with data.train = dict(type=...), using the type string in the first column. Generated from the live registry - 7 classes (7 names including aliases), grouped by role.

Generic

Directory / JSON-split datasets of pre-arranged numpy assets.

Name	Summary
ConcatDataset	Concatenate several configured datasets into a single index space.
DefaultDataset	Generic point-cloud dataset over preprocessed per-event .npy assets.

pimm.datasets.defaults.ConcatDataset	Concatenate several configured datasets into a single index space.
pimm.datasets.defaults.DefaultDataset	Generic point-cloud dataset over preprocessed per-event .npy assets.

LArTPC point clouds

Liquid-argon TPC detector datasets (PILArNet-M, JAXTPC, MicroBooNE).

Name	Summary
JAXTPCDataset	Multimodal LArTPC simulation dataset over co-indexed JAXTPC HDF5 files.
PILArNetH5Dataset	PILArNet-M LArTPC dataset read directly from clustered HDF5 shards.

pimm.datasets.jaxtpc_dataset.JAXTPCDataset	Multimodal LArTPC simulation dataset over co-indexed JAXTPC HDF5 files.
pimm.datasets.pilarnet.h5.PILArNetH5Dataset	PILArNet-M LArTPC dataset read directly from clustered HDF5 shards.

Water Cherenkov (LUCiD)

Water-Cherenkov sensor/segment datasets and SSL/ring-panoptic variants.

Name	Summary
LUCiDDataset	Water Cherenkov detector dataset over co-indexed LUCiD HDF5 files.

pimm.datasets.lucid_dataset.LUCiDDataset

Water Cherenkov detector dataset over co-indexed LUCiD HDF5 files.

Other

Registered classes not yet assigned to a section above (update CATEGORIES in docs/gen_api.py).

Name	Summary
PILArNetParquetDataset	Map-style PILArNet reader backed by parquet (Arrow-mmap).
PILArNetParquetIterableDataset	Streaming PILArNet reader for large-scale pretraining.

pimm.datasets.pilarnet.parquet.PILArNetParquetDataset	Map-style PILArNet reader backed by parquet (Arrow-mmap).
pimm.datasets.pilarnet.parquet.PILArNetParquetIterableDataset	Streaming PILArNet reader for large-scale pretraining.