development_logs/legacy_src/chat_app.py

import sys
from loguru import logger
from langchain_qdrant import QdrantVectorStore
from langchain_huggingface import HuggingFaceEmbeddings
from qdrant_client import QdrantClient

# Configuration
DB_PATH = "data/qdrant_db"
COLLECTION_NAME = "satellites"
MODEL_NAME = "BAAI/bge-small-en-v1.5"

# Setup logging
logger.remove()
logger.add(sys.stderr, format="<green>{time:HH:mm:ss}</green> | <level>{level: <8}</level> | <cyan>{message}</cyan>", level="INFO")

def get_vector_store():
    logger.info(f"Loading embedding model: {MODEL_NAME}...")
    embeddings = HuggingFaceEmbeddings(
        model_name=MODEL_NAME,
        model_kwargs={'device': 'cpu'},
        encode_kwargs={'normalize_embeddings': True}
    )
    
    logger.info(f"Connecting to Qdrant at {DB_PATH}...")
    client = QdrantClient(path=DB_PATH)
    
    vector_store = QdrantVectorStore(
        client=client,
        collection_name=COLLECTION_NAME,
        embedding=embeddings,
    )
    return vector_store

def chat_loop():
    args = sys.argv[1:]
    if args:
        query = " ".join(args)
        logger.info(f"Retrieving context for: '{query}'")
        try:
            vector_store = get_vector_store()
            results = vector_store.similarity_search_with_score(query, k=3)
            
            print(f"\n--- Context Retrieved ({len(results)} chunks) ---")
            for i, (doc, score) in enumerate(results, 1):
                meta = doc.metadata
                print(f"[{i}] {meta.get('name')} (Score: {score:.4f})")
                print(f"    Source: {meta.get('url')}")
                print(f"    Content Snippet: {doc.page_content[:150]}...")
        except Exception as e:
            logger.error(f"Error during retrieval: {e}")
        return

    logger.info("Initializing RAG Retrieval System...")
    try:
        vector_store = get_vector_store()
        logger.info("System Ready. Type 'exit' to quit.")
    except Exception as e:
        logger.error(f"Failed to initialize: {e}")
        return

    while True:
        try:
            query = input("\nUse Query: ")
            if query.lower() in ['exit', 'quit']:
                break
            
            logger.info(f"Retrieving context for: '{query}'")
            
            # Simple similarity search
            results = vector_store.similarity_search_with_score(query, k=3)
            
            context_text = "\n\n".join([doc.page_content for doc, _score in results])
            
            print(f"\n--- Context Retrieved ({len(results)} chunks) ---")
            for i, (doc, score) in enumerate(results, 1):
                meta = doc.metadata
                print(f"[{i}] {meta.get('name')} (Score: {score:.4f})")
            
            # Placeholder for Generation
            print("\n--- Answer Generation (Requires Ollama) ---")
            print("Note: Install Ollama (https://ollama.com/) and run `ollama pull llama3`")
            # In a real app, we'd call:
            # from langchain_community.llms import Ollama
            # llm = Ollama(model="llama3")
            # response = llm.invoke(f"Context: {context_text}\n\nQuestion: {query}")
            # print(response)
            
        except KeyboardInterrupt:
            break
        except Exception as e:
            logger.error(f"Error during retrieval: {e}")

if __name__ == "__main__":
    chat_loop()