spine.io.parse.data

Data structures used as canonical outputs of IO parsers.

Classes

ParserEdgeIndex(features, span)	Container describing one edge-index payload.
ParserIndex(features, span)	Container describing one flat index payload.
ParserIndexList(features, span[, single_counts])	Container describing one jagged index-list payload.
ParserObjectList(object_list, default[, ...])	Object list with index shifting instructions.
ParserTensor(features[, coords, meta, ...])	Container describing a parsed tensor-like payload.

class spine.io.parse.data.ParserTensor(features: ndarray, coords: ndarray | None = None, meta: ImageMeta3D | None = None, index_shifts: ndarray | None = None, index_cols: ndarray | None = None, remove_duplicates: bool = False, sum_cols: ndarray | None = None, avg_cols: ndarray | None = None, prec_col: int | None = None, precedence: ndarray | None = None, feats_only: bool = False, overlay_reference: str | None = None)

Container describing a parsed tensor-like payload.

features

Feature matrix associated with the parsed tensor.

Type:

np.ndarray

coords

Sparse tensor coordinates, typically with shape (N, 3).

Type:

np.ndarray, optional

meta

Geometry metadata used to convert voxel indices into detector coordinates.

Type:

Meta, optional

index_shifts

Shifts applied to index-bearing feature columns during batching.

Type:

np.ndarray, optional

index_cols

Feature columns that store indices.

Type:

np.ndarray, optional

remove_duplicates

If True, drop duplicate coordinates during collation.

Type:

bool, default False

sum_cols

Feature columns that should be summed when duplicates are merged.

Type:

np.ndarray, optional

avg_cols

Feature columns that should be averaged when duplicates are merged.

Type:

np.ndarray, optional

prec_col

Feature column used to break duplicate-coordinate ties.

Type:

int, optional

precedence

Precedence ordering used with prec_col.

Type:

np.ndarray, optional

feats_only

If True, the payload is feature-only and has no associated coordinate tensor.

Type:

bool, default False

overlay_reference

Product key whose duplicate-cleaning row selection should be applied to this tensor during overlay.

Type:

str, optional

Attributes:

avg_cols

coords

feat_avg_cols

Return duplicate-averaged columns in feature-only coordinates.

feat_index_cols

Return index-bearing columns expressed in feature-only coordinates.

feat_prec_col

Return the precedence column in feature-only coordinates.

feat_sum_cols

Return duplicate-summed columns in feature-only coordinates.

index_cols

index_shifts

meta

overlay_reference

prec_col

precedence

sum_cols

features: ndarray

coords: ndarray | None = None

meta: ImageMeta3D | None = None

index_shifts: ndarray | None = None

index_cols: ndarray | None = None

remove_duplicates: bool = False

sum_cols: ndarray | None = None

avg_cols: ndarray | None = None

prec_col: int | None = None

precedence: ndarray | None = None

feats_only: bool = False

overlay_reference: str | None = None

property feat_index_cols: ndarray | None

Return index-bearing columns expressed in feature-only coordinates.

Returns:

Feature-column indices corresponding to index_cols.

Return type:

np.ndarray, optional

property feat_sum_cols: ndarray | None

Return duplicate-summed columns in feature-only coordinates.

Returns:

Feature-column indices corresponding to sum_cols.

Return type:

np.ndarray, optional

property feat_avg_cols: ndarray | None

Return duplicate-averaged columns in feature-only coordinates.

Returns:

Feature-column indices corresponding to avg_cols.

Return type:

np.ndarray, optional

property feat_prec_col: int | None

Return the precedence column in feature-only coordinates.

Returns:

Feature-column index corresponding to prec_col.

Return type:

int, optional

class spine.io.parse.data.ParserIndex(features: ndarray, span: int)

Container describing one flat index payload.

features

One-dimensional index array.

Type:

np.ndarray

span

Parent-entry span used when batching entries.

Type:

int

features: ndarray

span: int

class spine.io.parse.data.ParserIndexList(features: list[ndarray], span: int, single_counts: ndarray | None = None)

Container describing one jagged index-list payload.

features

List of one-dimensional index arrays.

Type:

list[np.ndarray]

span

Parent-entry span used when batching entries.

Type:

int

single_counts

Per-index sizes used to restore jagged list structure after batching.

Type:

np.ndarray, optional

Attributes:

single_counts

features: list[ndarray]

span: int

single_counts: ndarray | None = None

class spine.io.parse.data.ParserEdgeIndex(features: ndarray, span: int)

Container describing one edge-index payload.

features

Two-dimensional edge-index array with shape (2, E).

Type:

np.ndarray

span

Parent-entry node span used when batching entries.

Type:

int

features: ndarray

span: int

class spine.io.parse.data.ParserObjectList(object_list: list[Any], default: Any, index_shifts: int | dict[str, int] | None = None)

Object list with index shifting instructions.

index_shifts

Shift(s) to apply to object index attributes during collation.

Type:

int or dict[str, int]

Attributes:

to_object_list

Drop parser-specific batching metadata and return a plain ObjectList.

Methods

append(object, /)	Append object to the end of the list.
clear(/)	Remove all items from list.
copy(/)	Return a shallow copy of the list.
count(value, /)	Return number of occurrences of value.
extend(iterable, /)	Extend list by appending elements from the iterable.
index(value[, start, stop])	Return first index of value.
insert(index, object, /)	Insert object before index.
pop([index])	Remove and return item at index (default last).
remove(value, /)	Remove first occurrence of value.
reverse(/)	Reverse IN PLACE.
sort(*[, key, reverse])	Sort the list in ascending order and return None.

property to_object_list: ObjectList

Drop parser-specific batching metadata and return a plain ObjectList.

Returns:

Underlying object list without index_shifts metadata.

Return type:

ObjectList