spine.io.parse.hdf5.index

Lightweight parsers for cached HDF5 index products.

Classes

HDF5EdgeIndexParser(dtype, index_event, ...)	Build an edge-index ParserEdgeIndex from cached HDF5 data.
HDF5IndexListParser(dtype, index_event, ...)	Build an index-list ParserIndexList from cached HDF5 data.
HDF5IndexParser(dtype, index_event, count_event)	Build a flat ParserIndex from cached HDF5 data.

class spine.io.parse.hdf5.index.HDF5IndexParser(dtype: str, index_event: str, count_event: str)

Build a flat ParserIndex from cached HDF5 data.

Attributes:

overlay

Methods

__call__(trees)	Parse one cached entry into a flat index parser tensor.
get_input_data(trees)	Build the parser-call input dictionary from loaded tree products.
process(index_event, count_event)	Normalize one cached flat index for collation into an IndexBatch.

name: ClassVar[str | None] = 'index'

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

process(index_event: ndarray, count_event: ndarray) → ParserIndex

Normalize one cached flat index for collation into an IndexBatch.

Parameters:

• index_event (np.ndarray) – Cached flat index array for one event entry.

• count_event (np.ndarray) – Cached tensor or scalar count used to infer the offset span of the indexed parent entry.

Returns:

Parser index containing one normalized 1D index array.

Return type:

ParserIndex

class spine.io.parse.hdf5.index.HDF5IndexListParser(dtype: str, index_event: str, count_event: str)

Build an index-list ParserIndexList from cached HDF5 data.

Attributes:

overlay

Methods

__call__(trees)	Parse one cached entry into a jagged index-list parser tensor.
get_input_data(trees)	Build the parser-call input dictionary from loaded tree products.
process(index_event, count_event)	Normalize cached index lists for collation into an IndexBatch.

name: ClassVar[str | None] = 'index_list'

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

process(index_event: ndarray, count_event: ndarray) → ParserIndexList

Normalize cached index lists for collation into an IndexBatch.

Parameters:

• index_event (np.ndarray) – Object array or nested array containing one index list per element.

• count_event (np.ndarray) – Cached tensor or scalar count used to infer the offset span of the indexed parent entry.

Returns:

Parser index list containing normalized 1D index arrays.

Return type:

ParserIndexList

class spine.io.parse.hdf5.index.HDF5EdgeIndexParser(dtype: str, index_event: str, count_event: str)

Build an edge-index ParserEdgeIndex from cached HDF5 data.

Attributes:

overlay

Methods

__call__(trees)	Parse one cached entry into an edge-index parser tensor.
get_input_data(trees)	Build the parser-call input dictionary from loaded tree products.
process(index_event, count_event)	Normalize cached edge indexes for collation into an EdgeIndexBatch.

name: ClassVar[str | None] = 'edge_index'

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

process(index_event: ndarray, count_event: ndarray) → ParserEdgeIndex

Normalize cached edge indexes for collation into an EdgeIndexBatch.

Parameters:

• index_event (np.ndarray) – Cached edge-index array with shape (2, E) or (E, 2).

• count_event (np.ndarray) – Cached tensor or scalar count used to infer the node span of the indexed parent entry.

Returns:

Parser edge index containing a normalized (2, E) array.

Return type:

ParserEdgeIndex