Doc Store Vector Stores
The Doc Store Vector Stores node is a component designed to search and retrieve documents from a Document Store using vector-based similarity search. It’s part of a larger system for managing and querying document collections efficiently.
Node Details
-
Label: Document Store (Vector)
-
Name: documentStoreVS
-
Version: 1.0
-
Type: DocumentStoreVS
-
Category: Vector Stores
Inputs
| Label | Name | Type | Description |
|---|---|---|---|
| Select Store | selectedStore | asyncOptions | Allows selection of an existing store |
The listStores method is used to dynamically populate the store options.
Outputs
| Label | Name | Base Classes |
|---|---|---|
| Retriever | retriever | BaseRetriever |
| Vector Store | vectorStore | VectorStore |
Functionality
-
Store Selection: Users can select an existing Document Store from a list of available stores.
-
Embedding Configuration: The node retrieves and configures the embedding model based on the selected store’s settings.
-
Vector Store Configuration: It sets up the vector store using the configuration associated with the selected store.
-
Initialization: The node initializes either a Retriever or a VectorStore object based on the configuration and user selection.
Key Methods
-
listStores: Asynchronously fetches and returns a list of available Document Stores. -
init: Initializes the node with the selected store and creates the necessary embedding and vector store objects.
Helper Functions
-
_createEmbeddingsObject: Creates and initializes the embedding object based on the store’s configuration. -
_createVectorStoreNodeData: Prepares the data structure for initializing the vector store. -
_createVectorStoreObject: Creates the actual vector store object based on the prepared data.
Usage
This node is typically used in workflows where efficient document retrieval is required. It’s particularly useful in scenarios involving:
-
Large document collections
-
Semantic search applications
-
Question-answering systems
-
Content recommendation engines
By leveraging vector-based similarity search, it can quickly find and retrieve relevant documents based on the semantic meaning of the query, rather than just keyword matching.
Integration
The DocStore_VectorStores node is designed to work seamlessly with other components in the system, such as various embedding models and vector store implementations. It can be easily incorporated into larger workflows for document processing, analysis, and retrieval tasks.