test_models.py

import os
import sys
import time
from openai import OpenAI

# ----------------------------------------------------------------------
#  GitHub Model API test (unchanged)
# ----------------------------------------------------------------------
def test_github_model_api():
    """Test GitHub‑hosted OpenAI model API."""
    print("=== Testing GitHub Model API ===")
    
    try:
        import requests
        print("✓ Requests package imported successfully")
    except ImportError as e:
        print(f"✗ Failed to import requests: {e}")
        return False
    
    # GitHub configuration (replace with your own token)
    token = "ghp_**************************"   # <-- keep your token private
    endpoint = "https://models.github.ai/inference"
    model = "openai/gpt-4.1-nano"

    if not token:
        print("✗ No GitHub token provided")
        return False
    
    print(f"Token format check: {token[:10]}...{token[-10:]}")
    print(f"Endpoint: {endpoint}")
    print(f"Model: {model}")
    
    try:
        client = OpenAI(base_url=endpoint, api_key=token)
        print(f"Testing GitHub model: {model}")
        
        response = client.chat.completions.create(
            model=model,
            messages=[
                {"role": "system", "content": "You are a helpful assistant."},
                {"role": "user",   "content": "Most valuable person in the world is?"}
            ],
            max_tokens=50,
            temperature=0.7,
            top_p=0.7,
        )
        
        if response.choices:
            content = response.choices[0].message.content
            print("✓ GitHub Model API test successful!")
            print(f"Response: {content}")
            return True
        else:
            print("✗ No choices in response")
            return False
            
    except Exception as e:
        print(f"✗ Unexpected error: {e}")
        print(f"Error type: {type(e).__name__}")
        import traceback
        print(f"Full traceback: {traceback.format_exc()}")
        return False


# ----------------------------------------------------------------------
#  HuggingFace Embeddings test (unchanged)
# ----------------------------------------------------------------------
def test_huggingface_embeddings():
    """Test HuggingFaceEmbeddings model."""
    print("\n=== Testing HuggingFaceEmbeddings ===")
    
    try:
        from langchain.embeddings import HuggingFaceEmbeddings
        print("✓ HuggingFaceEmbeddings imported successfully")
    except ImportError as e:
        print(f"✗ Failed to import HuggingFaceEmbeddings: {e}")
        return False
    
    try:
        model_name = "l3cube-pune/bengali-sentence-similarity-sbert"
        print(f"Loading model: {model_name}")
        embeddings = HuggingFaceEmbeddings(model_name=model_name)
        print("✓ Model loaded successfully")
        
        test_texts = [
            "প্রাথমিক শিক্ষার গুরুত্ব কি?",
            "স্বাস্থ্য ভালো রাখার উপায়",
            "প্রযুক্তির ব্যবহার"
        ]
        print("Testing embedding generation...")
        for i, txt in enumerate(test_texts, 1):
            emb = embeddings.embed_query(txt)
            print(f"✓ Text {i} embedded (dim: {len(emb)})")
        return True
        
    except Exception as e:
        print(f"✗ HuggingFaceEmbeddings test failed: {e}")
        return False


# ----------------------------------------------------------------------
#  Chroma vector‑store test (replaces FAISS test)
# ----------------------------------------------------------------------
def test_chroma():
    """Test that the Chroma vector store can be instantiated and queried."""
    print("\n=== Testing Chroma Vector Store ===")
    
    # ----‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑
    # Try the modern import first; fall back to the older API path if needed.
    try:
        from chromadb import PersistentClient          # Modern import (v0.4+)
        print("✓ chromadb imported via `from chromadb import PersistentClient`")
    except ImportError:
        try:
            from chromadb.api import PersistentClient  # Older versions
            print("✓ chromadb imported via `from chromadb.api import PersistentClient`")
        except ImportError as e:
            print(f"✗ Failed to import PersistentClient from chromadb: {e}")
            return False
    # ------------------------------------------------------------------
    try:
        # Use the same persistence directory defined in rag_core
        persist_dir = "chroma_db"
        client = PersistentClient(path=persist_dir)
        collection = client.get_or_create_collection(name="bangla_faqs")
        print(f"✓ Chroma collection '{collection.name}' ready (persist dir: {persist_dir})")
        return True
    except Exception as e:
        print(f"✗ Chroma test failed: {e}")
        return False


# ----------------------------------------------------------------------
#  RAG integration test (updated for Chroma)
# ----------------------------------------------------------------------
def test_rag_integration():
    """Test the full RAG pipeline using the new Chroma backend."""
    print("\n=== Testing RAG Integration (Chroma) ===")
    
    try:
        # Import the **correct** core module (it is named rag_core, not chroma_core)
        from chroma_core import (
            initialize_vectorstore,
            RAG_ENABLED,
            retrieve_documents,
            FAQ_DATA,
        )
        print("✓ rag_core imported successfully")
        print(f"✓ FAQ_DATA contains {len(FAQ_DATA)} entries")
    except Exception as e:
        print(f"✗ Failed to import rag_core: {e}")
        return False
    
    # Initialise the Chroma store
    print("Initialising Chroma vector store...")
    init_ok = initialize_vectorstore()
    
    if not init_ok or not RAG_ENABLED:
        print("✗ RAG system initialisation failed")
        return False
    
    print("✓ RAG system initialised")
    
    # Perform a simple retrieval test
    test_query = "শিক্ষা"
    print(f"Retrieving documents for query: '{test_query}' (topic='সব')")
    docs = retrieve_documents(query=test_query, topic="সব", k=2)
    
    if not docs:
        print("✗ No documents retrieved")
        return False
    
    print(f"✓ Retrieved {len(docs)} documents")
    for i, doc in enumerate(docs, 1):
        snippet = doc.page_content[:120].replace("\n", " ")
        print(f"  Doc {i}: {snippet}...")
    
    return True


# ----------------------------------------------------------------------
#  Main driver
# ----------------------------------------------------------------------
def main():
    print("GitHub Model Testing Script")
    print("=" * 50)
    
    github_ok = test_github_model_api()
    hf_ok     = test_huggingface_embeddings()
    chroma_ok = test_chroma()
    rag_ok    = test_rag_integration()
    
    print("\n" + "=" * 50)
    print("TEST RESULTS SUMMARY")
    print("=" * 50)
    print(f"GitHub Model API: {'✓ PASS' if github_ok else '✗ FAIL'}")
    print(f"HuggingFace:      {'✓ PASS' if hf_ok else '✗ FAIL'}")
    print(f"Chroma DB:        {'✓ PASS' if chroma_ok else '✗ FAIL'}")
    print(f"RAG Integration:  {'✓ PASS' if rag_ok else '✗ FAIL'}")
    
    if all([github_ok, hf_ok, chroma_ok, rag_ok]):
        print("\n🎉 All tests passed! Your chatbot is ready.")
    else:
        print("\n❌ Some tests failed. Review the messages above.")
        if not github_ok:
            print("- GitHub Model: verify token & endpoint.")
        if not hf_ok:
            print("- HuggingFace: check internet & disk space.")
        if not chroma_ok:
            print("- Chroma: ensure `chromadb` is installed (`pip install chromadb`).")
        if not rag_ok:
            print("- RAG: confirm `initialize_vectorstore()` succeeds and FAQ data is correct.")
    

if __name__ == "__main__":
    main()