redis_openai_agents.RedisVectorStore#

class RedisVectorStore(name, redis_url='redis://localhost:6379', vector_dims=384, distance_metric='COSINE', embeddings_cache=None)[source]#

Vector store for document storage and semantic search.

Uses Redis with RedisVL for high-performance vector similarity search.

Example

>>> store = RedisVectorStore(name="docs", redis_url="redis://localhost:6379")
>>> store.add_documents([{"content": "Hello world", "metadata": {"source": "test"}}])
>>> results = store.search(query="greeting", k=5)
Parameters:

  • name (str) – Index name in Redis

  • redis_url (str) – Redis connection URL

  • vector_dims (int) – Dimension of embedding vectors (default 384 for all-MiniLM-L6-v2)

  • distance_metric (str) – Distance metric (COSINE, L2, IP)

  • embeddings_cache (EmbeddingsCache | None)

Initialize the vector store.

Parameters:

  • name (str) – Index name in Redis

  • redis_url (str) – Redis connection URL

  • vector_dims (int) – Dimension of embedding vectors (384 for all-MiniLM-L6-v2)

  • distance_metric (str) – Distance metric (COSINE, L2, IP)

  • embeddings_cache (EmbeddingsCache | None) – Optional RedisVL EmbeddingsCache. When provided, repeated embeddings of identical content are served from the cache rather than re-invoking the vectorizer.

__init__(name, redis_url='redis://localhost:6379', vector_dims=384, distance_metric='COSINE', embeddings_cache=None)[source]#

Initialize the vector store.

Parameters:

  • name (str) – Index name in Redis

  • redis_url (str) – Redis connection URL

  • vector_dims (int) – Dimension of embedding vectors (384 for all-MiniLM-L6-v2)

  • distance_metric (str) – Distance metric (COSINE, L2, IP)

  • embeddings_cache (EmbeddingsCache | None) – Optional RedisVL EmbeddingsCache. When provided, repeated embeddings of identical content are served from the cache rather than re-invoking the vectorizer.

Return type:

None

Methods

__init__(name[, redis_url, vector_dims, ...])

Initialize the vector store.

aadd_documents(documents)

Async version of add_documents().

add_documents(documents)

Add documents to the vector store.

ahybrid_search(query[, k, text_weight, ...])

Async version of hybrid_search().

asearch(query[, k, filter])

Async version of search().

count()

Count documents in the store.

delete(ids)

Delete documents by ID.

delete_all()

Delete all documents and the index.

hybrid_search(query[, k, text_weight, ...])

Hybrid search combining vector similarity and BM25 text search.

search(query[, k, filter])

Search for similar documents.

Attributes

name

Index name in Redis.

vectorizer_model

Name of the underlying embedding model.