derive_generation#

scikitplot.corpus.derive_generation(documents, *, backend=None, schema_version=None)[source]#

Derive the generation identity for a set of documents.

Parameters:

documentsIterable[Any]

iterable of CorpusDocument

backendstr or None, optional

Name of the built index backend.

schema_versionstr or None, optional

Overrides the package’s current serialisation version, for tests and for loading a persisted index written by another build.

Returns:

IndexGeneration

Raises:

ValueError

If the documents span more than one embedding manifest. That condition is already refused at build time; deriving a single generation for incompatible vectors would silently paper over it.

Parameters:

• documents (Iterable[Any])

• backend (str | None)

• schema_version (str | None)

Return type:

IndexGeneration

Notes

Developer. embedding_manifest_id is taken from the documents rather than from an engine, so a persisted index can have its generation recomputed from what it contains – which is the whole point of moving off a counter.