Elasticsearch

pip install pineflow-vector-stores-elasticsearch

class ElasticsearchVectorStore

Provides functionality to interact with Elasticsearch for storing and querying document embeddings.

Parameters:

• index_name (str) – Name of the Elasticsearch index.

• url (str) – Elasticsearch instance URL.

• embed_model (BaseEmbedding) – Embedding model used to compute vectors.

• user (str, optional) – Elasticsearch username.

• password (str, optional) – Elasticsearch password.

• batch_size (int, optional) – Batch size for bulk operations. Defaults to 200.

• ssl (bool, optional) – Whether to use SSL. Defaults to False.

• distance_strategy (str, optional) – Distance strategy for similarity search. Currently supports “cosine”, “dot_product”, and “l2_norm”. Defaults to cosine.

• text_field (str, optional) – Name of the field containing text. Defaults to text.

• vector_field (str, optional) – Name of the field containing vector embeddings. Defaults to embedding.

Example

from pineflow.embeddings.huggingface import HuggingFaceEmbedding
from pineflow.vector_stores.elasticsearch import ElasticsearchVectorStore

embedding = HuggingFaceEmbedding()
es_vector_store = ElasticsearchVectorStore(
    index_name="pineflow-index",
    url="http://localhost:9200",
    embed_model=embedding,
)

add_documents(documents, create_index_if_not_exists=True)

Add documents to the Elasticsearch index.

Parameters:

• documents (List[Document]) – List of documents to add to the index.

• create_index_if_not_exists (bool, optional) – Whether to create the index if it doesn’t exist. Defaults to True.

delete_documents(ids)

Delete documents from the Elasticsearch index.

Parameters:

ids (List[str]) – List of documents IDs to delete.

get_all_documents(include_fields=[])

Get all documents from vector store.

search_documents(query, top_k=4)

Performs a similarity search for the top-k most similar documents.

Parameters:

• query (str) – Query text.

• top_k (int, optional) – Number of top results to return. Defaults to 4.

Returns:

List of the most similar documents.

Return type:

List[DocumentWithScore]