EmbeddingsClusteringFilter#

class langchain_community.document_transformers.embeddings_redundant_filter.EmbeddingsClusteringFilter[source]#

Bases: BaseDocumentTransformer, BaseModel

Perform K-means clustering on document vectors. Returns an arbitrary number of documents closest to center.

Create a new model by parsing and validating input data from keyword arguments.

Raises ValidationError if the input data cannot be parsed to form a valid model.

param embeddings: Embeddings [Required]#

Embeddings to use for embedding document contents.

param num_closest: int = 1#

The number of closest vectors to return for each cluster center.

param num_clusters: int = 5#

Number of clusters. Groups of documents with similar meaning.

param random_state: int = 42#

Controls the random number generator used to initialize the cluster centroids. If you set the random_state parameter to None, the KMeans algorithm will use a random number generator that is seeded with the current time. This means that the results of the KMeans algorithm will be different each time you run it.

param remove_duplicates: bool = False#

By default duplicated results are skipped and replaced by the next closest vector in the cluster. If remove_duplicates is true no replacement will be done: This could dramatically reduce results when there is a lot of overlap between clusters.

param sorted: bool = False#

By default results are re-ordered β€œgrouping” them by cluster, if sorted is true result will be ordered by the original position from the retriever

async atransform_documents(documents: Sequence[Document], **kwargs: Any) β†’ Sequence[Document]#

Asynchronously transform a list of documents.

Parameters:
  • documents (Sequence[Document]) – A sequence of Documents to be transformed.

  • kwargs (Any) –

Returns:

A sequence of transformed Documents.

Return type:

Sequence[Document]

transform_documents(documents: Sequence[Document], **kwargs: Any) β†’ Sequence[Document][source]#

Filter down documents.

Parameters:
  • documents (Sequence[Document]) –

  • kwargs (Any) –

Return type:

Sequence[Document]

Examples using EmbeddingsClusteringFilter