import os
from openai import OpenAI
import requests
import gradio as gr

# Initialize OpenAI client
client = OpenAI(api_key=os.environ.get("OPENAI_API_KEY"))

# Vector database search function
def search_document(query, k=5):
    url = "http://154.12.226.68:8000/search"
    
    payload = {
        "text": query,
        "k": k
    }
    
    headers = {
        "Content-Type": "application/json"
    }
    
    try:
        response = requests.post(url, json=payload, headers=headers)
        response.raise_for_status()
        return response.json()
    except requests.exceptions.RequestException as e:
        return f"An error occurred: {e}"

# Function to query OpenAI
def query_openai(prompt):
    try:
        response = client.chat.completions.create(
            model="gpt-4o-mini",
            messages=[
                {"role": "system", "content": "You are a helpful assistant. Answer the question based on the provided context."},
                {"role": "user", "content": prompt}
            ]
        )
        return response.choices[0].message.content
    except Exception as e:
        return f"An error occurred while querying OpenAI: {e}"

# Function to perform vector search and format results
def vector_search(query):
    results = search_document(query)
    if isinstance(results, str):  # Error occurred
        return results
    if not isinstance(results, dict) or 'results' not in results:
        return "Unexpected format in vector database response."
    
    formatted_results = ""
    for i, result in enumerate(results['results'], 1):
        content = result['metadata']['content']
        source = f"Source {i}: {result['metadata'].get('source', 'Unknown source')}, page {result['metadata'].get('page', 'Unknown page')}"
        metadata = ", ".join([f"{k}: {v}" for k, v in result['metadata'].items() if k != 'content'])
        formatted_results += f"{source}\nMetadata: {metadata}\nContent: {content}\n\n"
    
    return formatted_results

# Combined function for search and query
def search_and_query(question):
    # First, perform the vector search
    search_results = vector_search(question)
    
    # Then, use these results to query OpenAI
    prompt = f"""Given the following context and question, provide a comprehensive and accurate answer. Use ONLY the information provided in the context to answer. If the context doesn't contain relevant information to answer the question, state that clearly.

Context:
{search_results}

Question: {question}

Answer:"""
    
    openai_response = query_openai(prompt)
    
    # Return both the search results and the OpenAI response
    return search_results, openai_response

# Gradio interface
with gr.Blocks() as demo:
    question_input = gr.Textbox(label="Enter your question")
    search_output = gr.Textbox(label="Vector Search Results")
    answer_output = gr.Textbox(label="OpenAI Answer")
    query_button = gr.Button("Get Answer")
    
    query_button.click(search_and_query, inputs=question_input, outputs=[search_output, answer_output])

demo.launch()