vector_store.py

import chromadb
from sentence_transformers import SentenceTransformer
import os

# --- Constants ---
MODEL_NAME = "all-MiniLM-L6-v2"
COLLECTION_NAME = "aura_mind_knowledge"
KNOWLEDGE_BASE_DIR = "knowledge_base_data"

# --- Initialize ChromaDB and Model ---
client = chromadb.PersistentClient(path="chroma_db")
model = SentenceTransformer(MODEL_NAME)
collection = client.get_or_create_collection(name=COLLECTION_NAME)

def embed_and_store_documents():
    """
    Reads documents from the knowledge base directory, generates embeddings,
    and stores them in ChromaDB.
    """
    if collection.count() > 0:
        print("✅ Knowledge base is already loaded into ChromaDB.")
        return

    print("Embedding and storing documents in ChromaDB...")
    documents = []
    ids = []
    for filename in os.listdir(KNOWLEDGE_BASE_DIR):
        if filename.endswith(".txt"):
            with open(os.path.join(KNOWLEDGE_BASE_DIR, filename), "r") as f:
                documents.append(f.read())
                ids.append(filename)

    if documents:
        embeddings = model.encode(documents).tolist()
        collection.add(
            embeddings=embeddings,
            documents=documents,
            ids=ids
        )
        print(f"✅ Successfully stored {len(documents)} documents in ChromaDB.")

def search_documents(query: str, n_results: int = 1) -> list:
    """
    Searches for relevant documents in ChromaDB based on a query.

    Args:
        query: The search query.
        n_results: The number of results to return.

    Returns:
        A list of relevant documents.
    """
    if not query:
        return []

    query_embedding = model.encode([query]).tolist()
    results = collection.query(
        query_embeddings=query_embedding,
        n_results=n_results,
    )
    return results['documents'][0] if results['documents'] else []