import openai

from IPython.display import display, HTML

openai.api_key = OPENAI_API_KEY

def get_embedding(text):
    """Generate an embedding for the given text using OpenAI's API."""

    # Check for valid input
    if not text or not isinstance(text, str):
        return None

    try:
        # Call OpenAI API to get the embedding
        embedding = openai.embeddings.create(
            input=text,
            model="text-embedding-3-small", dimensions=1536).data[0].embedding
        return embedding
    except Exception as e:
        print(f"Error in get_embedding: {e}")
        return None

def handle_user_query(query, db, collection):
    # Assuming vector_search returns a list of dictionaries with keys 'title' and 'plot'
    get_knowledge = vector_search(query, db, collection)

    # Check if there are any results
    if not get_knowledge:
        return "No results found.", "No source information available."
        
     # Convert search results into a list of SearchResultItem models
    search_results_models = [
        SearchResultItem(**result)
        for result in get_knowledge
    ]

    # Convert search results into a DataFrame for better rendering in Jupyter
    search_results_df = pd.DataFrame([item.dict() for item in search_results_models])

    # Generate system response using OpenAI's completion
    completion = openai.chat.completions.create(
        model="gpt-3.5-turbo",
        messages=[
            {
                "role": "system", 
                "content": "You are a airbnb listing recommendation system."},
            {
                "role": "user", 
                "content": f"Answer this user query: {query} with the following context:\n{search_results_df}"
            }
        ]
    )

    system_response = completion.choices[0].message.content

    # Print User Question, System Response, and Source Information
    print(f"- User Question:\n{query}\n")
    print(f"- System Response:\n{system_response}\n")

    # Display the DataFrame as an HTML table
    display(HTML(search_results_df.to_html()))

    # Return structured response and source info as a string
    return system_response