test_local.py

import os
from rag_pipeline import load_repo, create_vectorstore, create_qa_chain, query
from dotenv import load_dotenv

# Load local environment variables for OpenAI API Key (if any)
load_dotenv()

def test_local_rag(repo_url: str):
    """
    Test the RAG pipeline with sample questions.
    """
    openai_api_key = os.getenv("OPENAI_API_KEY")
    if not openai_api_key:
        print("❌ Error: Set OPENAI_API_KEY env variable or add to .env file.")
        return

    print(f"🚀 Initializing RAG pipeline for: {repo_url}...")
    
    try:
        # 1. Load Repo
        print("📥 Loading repository...")
        documents = load_repo(repo_url)
        
        # 2. Create Vectorstore
        print("🧠 Creating vector store and processing embeddings...")
        vectorstore = create_vectorstore(documents)
        
        # 3. Create QA Chain
        print("🔗 Setting up QA chain...")
        qa_chain = create_qa_chain(vectorstore, openai_api_key)
        
        # 4. Test Sample Questions
        questions = [
            "How does shared memory tiling work?",
            "What is tree reduction?",
            "How does CUDA optimization work?",
            "Explain the backward pass",
            "What are the main functions?"
        ]
        
        print("\n--- 🧪 Starting local tests ---\n")
        
        for q in questions:
            print(f"❓ QUESTION: {q}")
            answer, sources = query(qa_chain, q)
            print("-" * 30)
            print(f"💡 AI ANSWER:\n{answer}")
            print("-" * 15)
            print(f"📄 Top source chunks retrieved:")
            for i, src in enumerate(sources[:2]): # Show only top 2 chunks
                path = src.metadata.get('file_path', 'Unknown')
                print(f"  - Chunk {i+1} from {path}")
            print("\n" + "="*50 + "\n")

    except Exception as e:
        print(f"❌ An error occurred during testing: {e}")

if __name__ == "__main__":
    # You can change this to any public repo you want to test
    # (e.g., https://github.com/MacroTorch/MacroTorch if it were public)
    # Using a common small repo for demonstration if MacroTorch is unavailable
    DEFAULT_TEST_REPO = "https://github.com/MacroTorch/MacroTorch" 
    # (Note: langchain repo might be too large for a quick test, maybe something smaller)
    
    test_local_rag(DEFAULT_TEST_REPO)