ParentChildChunker#

class scikitplot.corpus.ParentChildChunker(config=None)[source]#

Large parents for context, small children for retrieval, linked.

Parameters:

configParentChildChunkerConfig or None, default=None: None is ParentChildChunkerConfig().

Parameters:

config (ParentChildChunkerConfig | None)

See also

RecursiveChunker: The splitting rule used at both levels.
scikitplot.corpus.children: Traverse the documents it produces.

Notes

User. Each parent is followed by its children. Every chunk’s metadata holds chunk_role ("parent" or "child") and parent_span; a child holds parent_chunk, the position of its parent in the result. A document built from a child gets parent_doc_id (its parent’s doc_id): retrieve children, then give the model the parent. Documents keep chunk_role and parent_span in their metadata, so with emit_parents=False a child still locates its parent’s region. Both levels report the RECURSIVE strategy; chunk_role tells them apart. A parent no longer than child_size has one child with the same text: index children only.

Examples

>>> text = "Alpha one. Beta two.\n\nGamma three. Delta four."
>>> config = ParentChildChunkerConfig(parent_size=25, child_size=12)
>>> result = ParentChildChunker(config).chunk(text)
>>> [
...     (c.metadata["chunk_role"], c.text) for c in result.chunks
... ]
[('parent', 'Alpha one. Beta two.'), ('child', 'Alpha one.'), ('child', 'Beta two.'),
 ('parent', 'Gamma three. Delta four.'), ('child', 'Gamma three.'), ('child', 'Delta four.')]

chunk(text, doc_id=None, extra_metadata=None)[source]#

Split text into parents and the children inside each.

Parameters:

textstr: Raw document text.
doc_idstr or None, default=None: Stored in each chunk’s metadata.
extra_metadatadict or None, default=None: Merged into the result metadata.

Returns:

ChunkResult: Parents and children in source order (each parent first).

Raises:

TypeError: If text is not a str.
ValueError: If text is empty or whitespace-only, or holds a NUL character or a lone surrogate.

Parameters:

text (str)
doc_id (str | None)
extra_metadata (dict[str, Any] | None)

Return type:

ChunkResult

chunk_batch(texts, doc_ids=None, extra_metadata=None)[source]#

Chunk a list of documents.

Parameters:

textslist of str: Input documents.
doc_idslist of str or None, default=None: Parallel document identifiers.
extra_metadatadict or None, default=None: Shared metadata for every result.

Returns:

list of ChunkResult: One result per document.

Raises:

TypeError: If texts is not a list.
ValueError: If doc_ids length mismatches texts.

Parameters:

texts (list[str])
doc_ids (list[str] | None)
extra_metadata (dict[str, Any] | None)

Return type:

list[ChunkResult]

property config: ParentChildChunkerConfig#: The configuration (read-only).

config_class[source]#: alias of ParentChildChunkerConfig

strategy: ClassVar[ChunkingStrategy] = 'recursive'[source]#

text_is_source_slice = True#