Transforms registry

Transform pipeline steps. See ../../data/transforms.

Build any of these from a config with transform=[dict(type=...)], using the type string in the first column. Generated from the live registry - 46 classes (46 names including aliases), grouped by role.

Input & collection

Tensor conversion and the final Collect projection into the model batch.

Name	Summary
Collect	Final projection of a sample dict into the model-facing batch contract.
Copy	Duplicate keys in the sample dict under new names.
ToTensor	Recursively convert numpy arrays and numeric leaves to torch tensors.
Update	Inject constant key-value pairs into the sample dict.

pimm.datasets.transform.base.Collect	Final projection of a sample dict into the model-facing batch contract.
pimm.datasets.transform.base.Copy	Duplicate keys in the sample dict under new names.
pimm.datasets.transform.base.ToTensor	Recursively convert numpy arrays and numeric leaves to torch tensors.
pimm.datasets.transform.base.Update	Inject constant key-value pairs into the sample dict.

Spatial & geometric

Coordinate normalization, grid sampling, crops, rotations, jitter, and warps.

Name	Summary
CenterShift	Center coord on the midpoint of its bounding box per axis.
ClipGaussianJitter	Add isotropic clipped multivariate-Gaussian noise to coord.
ConditionalRandomTransform	Wall-aware random translation that keeps points inside fixed bounds.
CropBoundary	Drop points whose segment label is 0 or 1.
ElasticDistortion	Apply smooth random elastic warping to coord.
GridSample	Voxel-downsample a point cloud onto a hash grid.
HardExampleCrop	Spherical crop biased toward rare (“hard”) segment labels.
MultiplicativeRandomJitter	Multiply one or more keys by clipped Gaussian noise around 1.
NormalizeCoord	Recenter and rescale coord into a normalized frame.
PointClip	Clamp coord to an axis-aligned point-cloud range.
PositiveShift	Translate coord so its minimum corner sits at the origin.
RandomDrop	Randomly overwrite a fraction of one key’s rows with a constant.
RandomDropout	Randomly drop a fraction of points from the sample.
RandomFlip	Randomly mirror the cloud across one or more axes.
RandomJitter	Add clipped Gaussian noise to one or more point-aligned keys.
RandomRotate	Rotate the cloud by a random angle about one axis.
RandomRotateTargetAngle	Rotate the cloud by a random angle drawn from a discrete set.
RandomScale	Scale coord by a random factor.
RandomShift	Translate coord by a uniform random per-axis offset.
SetRandomValue	Overwrite one or more keys with clipped Gaussian noise.
ShufflePoint	Randomly permute the point ordering of the sample.
SphereCrop	Crop to the nearest points around a center to cap the point count.

pimm.datasets.transform.spatial.CenterShift	Center coord on the midpoint of its bounding box per axis.
pimm.datasets.transform.spatial.ClipGaussianJitter	Add isotropic clipped multivariate-Gaussian noise to coord.
pimm.datasets.transform.spatial.ConditionalRandomTransform	Wall-aware random translation that keeps points inside fixed bounds.
pimm.datasets.transform.spatial.CropBoundary	Drop points whose segment label is 0 or 1.
pimm.datasets.transform.spatial.ElasticDistortion	Apply smooth random elastic warping to coord.
pimm.datasets.transform.spatial.GridSample	Voxel-downsample a point cloud onto a hash grid.
pimm.datasets.transform.spatial.HardExampleCrop	Spherical crop biased toward rare ("hard") segment labels.
pimm.datasets.transform.spatial.MultiplicativeRandomJitter	Multiply one or more keys by clipped Gaussian noise around 1.
pimm.datasets.transform.spatial.NormalizeCoord	Recenter and rescale coord into a normalized frame.
pimm.datasets.transform.spatial.PointClip	Clamp coord to an axis-aligned point-cloud range.
pimm.datasets.transform.spatial.PositiveShift	Translate coord so its minimum corner sits at the origin.
pimm.datasets.transform.spatial.RandomDrop	Randomly overwrite a fraction of one key's rows with a constant.
pimm.datasets.transform.spatial.RandomDropout	Randomly drop a fraction of points from the sample.
pimm.datasets.transform.spatial.RandomFlip	Randomly mirror the cloud across one or more axes.
pimm.datasets.transform.spatial.RandomJitter	Add clipped Gaussian noise to one or more point-aligned keys.
pimm.datasets.transform.spatial.RandomRotate	Rotate the cloud by a random angle about one axis.
pimm.datasets.transform.spatial.RandomRotateTargetAngle	Rotate the cloud by a random angle drawn from a discrete set.
pimm.datasets.transform.spatial.RandomScale	Scale coord by a random factor.
pimm.datasets.transform.spatial.RandomShift	Translate coord by a uniform random per-axis offset.
pimm.datasets.transform.spatial.SetRandomValue	Overwrite one or more keys with clipped Gaussian noise.
pimm.datasets.transform.spatial.ShufflePoint	Randomly permute the point ordering of the sample.
pimm.datasets.transform.spatial.SphereCrop	Crop to the nearest points around a center to cap the point count.

Feature, energy & color

Energy/charge log-transforms and feature/color augmentations.

Name	Summary
ChromaticAutoContrast	Randomly auto-contrast the per-point RGB color.
ChromaticJitter	Add independent per-point Gaussian noise to RGB channels.
ChromaticTranslation	Randomly shift all RGB channels by a shared offset.
EnergeticTranslation	Randomly shift per-point energy by a shared scalar offset.
EnergyJitter	Apply multiplicative per-point jitter to energy.
HueSaturationTranslation	Randomly translate hue and scale saturation of point colors.
LogTransform	Compress scalar features (e.g. energy) onto [-1, 1].
MomentumTransform	Log10-compress strictly-positive momentum values.
NormalizeColor	Scale per-point colors from [0, 255] into [0, 1].
RandomColorDrop	Randomly drop (zero out or attenuate) per-point color.
RandomColorGrayScale	Randomly convert per-point RGB color to grayscale.
RandomColorJitter	Randomly jitter brightness, contrast, saturation, and hue of point colors.
RelativeLogNormalize	Per-event relative log normalization (e.g. for hit times).

pimm.datasets.transform.color.ChromaticAutoContrast	Randomly auto-contrast the per-point RGB color.
pimm.datasets.transform.color.ChromaticJitter	Add independent per-point Gaussian noise to RGB channels.
pimm.datasets.transform.color.ChromaticTranslation	Randomly shift all RGB channels by a shared offset.
pimm.datasets.transform.color.EnergeticTranslation	Randomly shift per-point energy by a shared scalar offset.
pimm.datasets.transform.color.EnergyJitter	Apply multiplicative per-point jitter to energy.
pimm.datasets.transform.color.HueSaturationTranslation	Randomly translate hue and scale saturation of point colors.
pimm.datasets.transform.color.LogTransform	Compress scalar features (e.g. energy) onto [-1, 1].
pimm.datasets.transform.color.MomentumTransform	Log10-compress strictly-positive momentum values.
pimm.datasets.transform.color.NormalizeColor	Scale per-point colors from [0, 255] into [0, 1].
pimm.datasets.transform.color.RandomColorDrop	Randomly drop (zero out or attenuate) per-point color.
pimm.datasets.transform.color.RandomColorGrayScale	Randomly convert per-point RGB color to grayscale.
pimm.datasets.transform.color.RandomColorJitter	Randomly jitter brightness, contrast, saturation, and hue of point colors.
pimm.datasets.transform.color.RelativeLogNormalize	Per-event relative log normalization (e.g. for hit times).

Detector-specific

LArTPC label/motif derivations (PDG→semantic, stuff/noise, multi-scale time).

Name	Summary
PDGToSemantic	Fallback semantic labels derived from PDG codes.

pimm.datasets.transform.detector.PDGToSemantic

Fallback semantic labels derived from PDG codes.

Instance

Instance parsing, anchors, and local-covariance features for detection.

Name	Summary
ComputeAnchors	Compute geometric anchors once per event and attach them to the sample.
InstanceParser	Compact instance ids and derive per-instance bounding boxes and centroids.
LocalCovarianceFeatures	Per-point local-neighborhood covariance eigen-features.

pimm.datasets.transform.instance.ComputeAnchors	Compute geometric anchors once per event and attach them to the sample.
pimm.datasets.transform.instance.InstanceParser	Compact instance ids and derive per-instance bounding boxes and centroids.
pimm.datasets.transform.instance.LocalCovarianceFeatures	Per-point local-neighborhood covariance eigen-features.

Multi-view (SSL)

Contrastive / multi-scale view generators for self-supervised pretraining.

Name	Summary
ContrastiveViewsGenerator	Generate two independently-augmented views for contrastive SSL.
MixedScaleGeometryMultiViewGenerator	Multi-view generator with coarse locals plus extra fine local crops.
MultiViewGenerator	Generate multiple global and local crops for DINO/iBOT-style SSL.

pimm.datasets.transform.multiview.ContrastiveViewsGenerator	Generate two independently-augmented views for contrastive SSL.
pimm.datasets.transform.multiview.MixedScaleGeometryMultiViewGenerator	Multi-view generator with coarse locals plus extra fine local crops.
pimm.datasets.transform.multiview.MultiViewGenerator	Generate multiple global and local crops for DINO/iBOT-style SSL.