spine.io.parse.larcv.particle

Module that contains all parsers related to LArCV particle data.

Contains the following parsers: - LArCVParticleParser - LArCVNeutrinoParser - LArCVParticlePointParser - LArCVParticleCoordinateParser - LArCVParticleGraphParser - LArCVSingleParticlePIDParser - LArCVSingleParticleEnergyParser

Classes

`LArCVNeutrinoParser`([pixel_coordinates, ...])	Class which loads larcv.Neutrino objects to local Neutrino ones.
`LArCVParticleCoordinateParser`(dtype, **kwargs)	Class that retrieves that end points of particles.
`LArCVParticleGraphParser`([...])	Class that uses larcv.EventParticle to construct edges between particles (i.e. clusters).
`LArCVParticleParser`([pixel_coordinates, ...])	Class which loads larcv.Particle objects to local Particle ones.
`LArCVParticlePointParser`([include_point_tagging])	Class that retrieves the points of interests.
`LArCVSingleParticleEnergyParser`(dtype, **kwargs)	Get the first true particle's kinetic energy.
`LArCVSingleParticlePIDParser`(dtype, **kwargs)	Get the first true particle's species.
`LArCVVertexPointParser`(**kwargs)	Class that retrieves the vertices.

class spine.io.parse.larcv.particle.LArCVParticleParser(pixel_coordinates: bool = True, post_process: bool = True, skip_empty: bool = False, asis: bool = False, **kwargs: Any)[source]

Class which loads larcv.Particle objects to local Particle ones.

Attributes:

overlay

Methods

`__call__`(trees)	Parse one entry.
`get_input_data`(trees)	Build the parser-call input dictionary from loaded tree products.
`process`(particle_event[, sparse_event, ...])	Fetch the list of true particle objects.

name: ClassVar[str | None] = 'particle'

returns: ClassVar[str | None] = 'object_list'

process(particle_event: Any, sparse_event: Any | None = None, cluster_event: Any | None = None, particle_mpv_event: Any | None = None, neutrino_event: Any | None = None) → ParserObjectList[source]

Fetch the list of true particle objects.

Parameters:

particle_event (larcv.EventParticle) – Particle event which contains the list of true particles
sparse_event (larcv.EventSparseTensor3D, optional) – Tensor which contains the metadata needed to convert the positions in voxel coordinates
cluster_event (larcv.EventClusterVoxel3D, optional) – Cluster which contains the metadata needed to convert the positions in voxel coordinates
particle_mpv_event (larcv.EventParticle, optional) – Particle event which contains the list of true MPV particles
neutrino_event (larcv.EventNeutrino, optional) – Neutrino event which contains the list of true neutrinos

Returns:

List of true particle objects

Return type:

List[Particle]

class spine.io.parse.larcv.particle.LArCVNeutrinoParser(pixel_coordinates: bool = True, asis: bool = False, interaction_scheme: int | str = NuInteractionScheme.LARSOFT, **kwargs: Any)[source]

Class which loads larcv.Neutrino objects to local Neutrino ones.

Attributes:

overlay

Methods

`__call__`(trees)	Parse one entry.
`get_input_data`(trees)	Build the parser-call input dictionary from loaded tree products.
`process`(neutrino_event[, sparse_event, ...])	Fetch the list of true neutrino objects.

name: ClassVar[str | None] = 'neutrino'

returns: ClassVar[str | None] = 'object_list'

process(neutrino_event: Any, sparse_event: Any | None = None, cluster_event: Any | None = None) → ParserObjectList[source]

Fetch the list of true neutrino objects.

Parameters:

neutrino_event (larcv.EventNeutrino) – Neutrino event which contains the list of true neutrinos
sparse_event (larcv.EventSparseTensor3D, optional) – Tensor which contains the metadata needed to convert the positions in voxel coordinates
cluster_event (larcv.EventClusterVoxel3D, optional) – Cluster which contains the metadata needed to convert the positions in voxel coordinates

Returns:

List of true neutrino objects

Return type:

List[Neutrino]

class spine.io.parse.larcv.particle.LArCVParticlePointParser(include_point_tagging: bool = True, **kwargs: Any)[source]

Class that retrieves the points of interests.

It provides the coordinates of the end points, types and particle index.

Attributes:

overlay

Methods

`__call__`(trees)	Parse one entry.
`get_input_data`(trees)	Build the parser-call input dictionary from loaded tree products.
`process`(particle_event[, sparse_event, ...])	Fetch the list of label points of interest.

name: ClassVar[str | None] = 'particle_points'

returns: ClassVar[str | None] = 'tensor'

process(particle_event: Any, sparse_event: Any | None = None, cluster_event: Any | None = None) → ParserTensor[source]

Fetch the list of label points of interest.

Parameters:

particle_event (larcv.EventParticle) – Particle event which contains the list of true particles
sparse_event (larcv.EventSparseTensor3D, optional) – Tensor which contains the metadata needed to convert the positions in voxel coordinates
cluster_event (larcv.EventClusterVoxel3D, optional) – Cluster which contains the metadata needed to convert the positions in voxel coordinates

Returns:

np_voxels (np.ndarray) – (N, 3) array of [x, y, z] coordinates
np_features (np.ndarray) – (N, 2/3) array of [point type, particle index(, end point tagging)]
meta (Meta) – Metadata of the parsed image

class spine.io.parse.larcv.particle.LArCVParticleCoordinateParser(dtype: str, **kwargs: Any)[source]

Class that retrieves that end points of particles.

It provides the coordinates of the end points, time and shape.

Attributes:

overlay

Methods

`__call__`(trees)	Parse one entry.
`get_input_data`(trees)	Build the parser-call input dictionary from loaded tree products.
`process`(particle_event[, sparse_event, ...])	Fetch the start/end point and time of each true particle.

name: ClassVar[str | None] = 'particle_coords'

returns: ClassVar[str | None] = 'tensor'

process(particle_event: Any, sparse_event: Any | None = None, cluster_event: Any | None = None) → ParserTensor[source]

Fetch the start/end point and time of each true particle.

Parameters:

particle_event (larcv.EventParticle) – Particle event which contains the list of true particles
sparse_event (larcv.EventSparseTensor3D, optional) – Tensor which contains the metadata needed to convert the positions in voxel coordinates
cluster_event (larcv.EventClusterVoxel3D, optional) – Cluster which contains the metadata needed to convert the positions in voxel coordinates

Returns:

np_voxels (np.ndarray) – (N, 6) array of [x_s, y_s, z_s, x_e, y_e, z_e] start and end point coordinates
np_features (np.ndarray) – (N, 2) array of [first_step_t, shape_id]
meta (Meta) – Metadata of the parsed image

class spine.io.parse.larcv.particle.LArCVVertexPointParser(**kwargs: Any)[source]

Class that retrieves the vertices.

It provides the coordinates of points where multiple particles originate: - If the neutrino_event is provided, it simply uses the coordinates of

the neutrino interaction points.

If the particle_event is provided instead, it looks for ancestor point positions shared by at least two particles.

Attributes:

overlay

Methods

`__call__`(trees)	Parse one entry.
`get_input_data`(trees)	Build the parser-call input dictionary from loaded tree products.
`process`([particle_event, neutrino_event, ...])	Fetch the list of label vertex points.

name: ClassVar[str | None] = 'vertex_points'

returns: ClassVar[str | None] = 'tensor'

process(particle_event: Any | None = None, neutrino_event: Any | None = None, sparse_event: Any | None = None, cluster_event: Any | None = None) → ParserTensor[source]

Fetch the list of label vertex points.

Parameters:

particle_event (larcv.EventParticle) – Particle event which contains the list of true particles
neutrino_event (larcv.EventNeutrino) – Neutrino event which contains the list of true neutrinos
sparse_event (larcv.EventSparseTensor3D, optional) – Tensor which contains the metadata needed to convert the positions in voxel coordinates
cluster_event (larcv.EventClusterVoxel3D, optional) – Cluster which contains the metadata needed to convert the positions in voxel coordinates

Returns:

np_voxels (np.ndarray) – (N, 3) array of [x, y, z] coordinates
np_features (np.ndarray) – (N, 1) array of [vertex ID]
meta (Meta) – Metadata of the parsed image

class spine.io.parse.larcv.particle.LArCVParticleGraphParser(include_fragment_edges: bool = False, **kwargs: Any)[source]

Class that uses larcv.EventParticle to construct edges between particles (i.e. clusters).

Attributes:

overlay

Methods

`__call__`(trees)	Parse one entry.
`get_input_data`(trees)	Build the parser-call input dictionary from loaded tree products.
`process`(particle_event[, cluster_event])	Fetch the parentage connections from the true particle list.

name: ClassVar[str | None] = 'particle_graph'

returns: ClassVar[str | None] = 'tensor'

process(particle_event: Any, cluster_event: Any | None = None) → ParserEdgeIndex[source]

Fetch the parentage connections from the true particle list.

Configuration:

particle_event (larcv.EventParticle) – Particle event which contains the list of true particles
cluster_event (larcv.EventClusterVoxel3D, optional) – Cluster used to check if particles have 0 pixel in the image. If so, the edges to those clusters are removed and the broken parantage is subsequently patched.

Returns:

Parsed edge-index payload with shape (2, E) and the number of particles stored as the global shift span.

Return type:

ParserEdgeIndex

class spine.io.parse.larcv.particle.LArCVSingleParticlePIDParser(dtype: str, **kwargs: Any)[source]

Get the first true particle’s species.

Methods

`__call__`(trees)	Parse one entry.
`get_input_data`(trees)	Build the parser-call input dictionary from loaded tree products.
`process`(particle_event)	Fetch the species of the first particle.

name: ClassVar[str | None] = 'single_particle_pid'

returns: ClassVar[str | None] = 'scalar'

overlay: ClassVar[str | None] = 'cat'

process(particle_event: Any) → int[source]

Fetch the species of the first particle.

Configuration:: particle_event (larcv.EventParticle) – Particle event which contains the list of true particles
Returns:: Species of the first particle
Return type:: int

class spine.io.parse.larcv.particle.LArCVSingleParticleEnergyParser(dtype: str, **kwargs: Any)[source]

Get the first true particle’s kinetic energy.

Methods

`__call__`(trees)	Parse one entry.
`get_input_data`(trees)	Build the parser-call input dictionary from loaded tree products.
`process`(particle_event)	Fetch the kinetic energy of the first particle.

name: ClassVar[str | None] = 'single_particle_energy'

returns: ClassVar[str | None] = 'scalar'

overlay: ClassVar[str | None] = 'cat'

process(particle_event: Any) → float[source]

Fetch the kinetic energy of the first particle.

Configuration:: particle_event (larcv.EventParticle) – Particle event which contains the list of true particles
Returns:: Kinetic energy of the first particle
Return type:: float