thematic_search.SampleQuery

class thematic_search.SampleQuery(db, indices: ndarray)

A query object carrying a set of document indices. Supports navigation to topics or retrieval of document data.

Typical usage:

topicdb.q.search(“jazz music”).nearby().topics().theme()

__init__(db, indices: ndarray)

Methods

__init__(db, indices)
embeddings()	Return the embedding vectors for these indices.
metadata()	Return the document metadata rows for these indices.
neighbours([k])	Average the embeddings of this SampleQuery's indices, then find the k nearest neighbours.
recursive_theme([z_threshold, ...])	Recursively find a theme formula for these document indices.
strengths(expr)	Return the inclusion strengths of these documents for a cluster expression.
theme()	Find the most surprising topic for these document indices relative to the global corpus, using a KL-divergence-style surprise score.
to_fuzzy()	Use this SampleQuery's indices to filter a FuzzyQuery
topics([threshold, logic])	Return the topics containing these document indices.
unwrap()
where(query)	Filter documents by metadata column values.