Vector Store: Weaviate

Weaviate is a cloud-native, open-source vector search engine. It is the default vector database configuration utilized in the AI-Horizon platform for storing semantic embedding chunks generated by the RAG pipelines.

1. Configuration Parameters

To connect to Weaviate, pass your configuration within the vector_store_params dictionary:

Parameter	Type	Default	Description
`vector_store_type`	`string`	`"WeaviateVectorStore"`	Name of the Weaviate wrapper class.
`url`	`string`	`"http://localhost:8080"`	Endpoint URL of your Weaviate instance.
`api_key`	`string`	`None`	Optional API key authentication header.
`index_name`	`string`	`None`	Class index name in Weaviate (defaults to a generic session id class name).

2. Configuration Example

Here is how you specify Weaviate parameters during agent initialization:

from aih import ChatBot
 
weaviate_config = {
    "vector_store_type": "WeaviateVectorStore",
    "url": "https://your-weaviate-instance.weaviate.network",
    "api_key": "your_weaviate_wcs_key",
    "index_name": "DocumentationIndex"
}
 
chatbot = ChatBot.pdf_chat(
    input_files=["./handbook.pdf"],
    vector_store_params=weaviate_config
)

Search Agent Qdrant