# from embedding_model import ONNXDistilUSEModel
from services.embedding_models.MiniLM_L12_v2_model import ONNXMiniLMModel
# from services.embedding_models.embedding_e5_model import ONNXE5Model
from services.vector_db.similarity_model import VectorDB  # Changed to absolute import
import traceback
import os
import time

# Set OpenMP environment variable to avoid runtime conflicts
os.environ['KMP_DUPLICATE_LIB_OK'] = 'TRUE'

def main():
    data = [
        {
            "id": "1",
            "location": "Cairo",
            "description": "Enjoy the cultural festivals in Cairo this weekend!",
            "itineraries": [
                {
                    "dayNumber": 1,
                    "activities": [
                        "Egyptian Museum guided tour",
                        "Traditional music performance",
                        "Nile dinner cruise"
                    ]
                }
            ]
        },
        {
            "id": "2",
            "location": "Alexandria",
            "description": "Visit the stunning beaches of Alexandria",
            "itineraries": [
                {
                    "dayNumber": 1,
                    "activities": [
                        "Bibliotheca Alexandrina tour",
                        "Mediterranean beach visit",
                        "Seafood dining experience"
                    ]
                }
            ]
        },
        {
            "id": "3",
            "location": "Cairo",
            "description": "Music and dance night at Cairo Opera House",
            "itineraries": [
                {
                    "dayNumber": 1,
                    "activities": [
                        "Opera house tour",
                        "Classical concert",
                        "Downtown Cairo nightlife"
                    ]
                }
            ]
        }
    ]



    updated_data = [{
        "id": "1",
        "location": "Cairo",
        "description": "Updated festival description with new exiting sessions!",
        "itineraries": [
            {
                "dayNumber": 1,
                "activities": [
                    "New activity 1",
                    "New activity 2",
                    "Additional activity"
                ]
            }
        ]
    }]


    try:
        print("Step 1: Starting model initialization...")
        model = ONNXMiniLMModel()
        print("Model initialization completed successfully")
        
        print("\nStep 2: Starting vector database initialization...")
        vector_db = VectorDB()
        print("Vector database initialization completed successfully")
        
        # Correct way to get embeddings (using encode() instead of calling model directly)
        print("\nTesting model encoding...")
        embeddings = model.encode([item["description"] for item in data])
        print("Embeddings shape:", embeddings.shape)  # Should be (3, 384) for MiniLM
        print("Testing completed successfully")

        print("\nStep 3: Starting embeddings storage...")
        vector_db.store_embeddings(data, model)
        print("Initial data stored successfully")

        print("\nStep 4: Updating embeddings...")
        vector_db.update_embeddings(updated_data, model)
        print("Updates completed successfully")
            
        print("\nStep 5: Starting recommendation test...")
        start_time = time.time()
        response = vector_db.get_similar_by_ids(["1"], top_k=2)
        end_time = time.time()
        print(f"Recommendation test completed in {end_time - start_time:.4f} seconds")
        
        if response is None:
            raise Exception("No response received from recommendation query")
        
        print("\nResults:")
        print("Query:", response.get("query", "No query found"))
        print("\nRecommendations:")
        for item in response.get("recommendations", []):
            print("-" * 50)
            print(f"ID: {item['id']}")
            print(f"Similarity Score: {1 - item['distance']:.4f}")
        
        print("\nStep 6: Testing delete functionality...")
        try:
            vector_db.delete_items(["2"])
            print("Delete operation completed successfully")
            
            print("\nGetting recommendations after deletion...")
            response = vector_db.get_similar_by_ids(["1"], top_k=1)
            print("\nUpdated Results:")
            for item in response.get("recommendations", []):
                print("-" * 50)
                print(f"ID: {item['id']}")
                print(f"Similarity Score: {1 - item['distance']:.4f}")
        except Exception as e:
            print(f"Error during deletion test: {str(e)}")
        
    except Exception as e:
        print(f"\nError occurred: {str(e)}")
        traceback.print_exc()

if __name__ == "__main__":
    main()