Knowledge Database Interaction with Weaviate

This prototype interacts with a Knowledge Database built in Weaviate, a vector database designed for storing and retrieving interconnected structured objects and their relationships.

Database Structure

The database organizes information into a hierarchical structure as follows:

• QueryEntityInfo (query) → linked to → Documents (via inDocs)
• Documents → linked to → Pages (via hasPage)
• Pages → linked to → Results (via hasResult)

How It Works

The provided code recursively navigates these relationships by:

1. Querying the main entity: Starts by querying the QueryEntityInfo node.
2. Traversing to associated entities: The traversal then moves to related Documents, Pages, and Results.
3. Extracting information: Each entity’s information is extracted and displayed through its UUIDs using weaviate.util.get_valid_uuid.

Purpose

This process allows for a complete retrieval of related knowledge starting from a QueryEntityInfo node, forming a dynamic and flexible knowledge graph traversal.

Example

import weaviate
from weaviate.util import get_valid_uuid

# Initialize Weaviate client
client = weaviate.Client("http://localhost:8080")

# Example recursive function to traverse the graph
def retrieve_entity_info(query_entity_uuid):
    # Query the main entity
    query_entity = client.data_object.get(uuid=query_entity_uuid)
    
    # Traverse related documents, pages, and results
    documents = query_entity['inDocs']
    for doc in documents:
        doc_info = client.data_object.get(uuid=doc['uuid'])
        pages = doc_info['hasPage']
        
        for page in pages:
            page_info = client.data_object.get(uuid=page['uuid'])
            results = page_info['hasResult']
            
            for result in results:
                result_info = client.data_object.get(uuid=result['uuid'])
                print(result_info)

# Start traversal from a given QueryEntityInfo UUID
retrieve_entity_info("query_entity_uuid_here")