import os
import gradio as gr
import requests
import pandas as pd
from transformers import pipeline

# --- Constants ---
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
HF_MODEL_NAME = "facebook/bart-large-mnli"  # Smaller, free model that works well in Spaces

# --- Enhanced Agent Definition ---
class BasicAgent:
    def __init__(self, hf_token=None):
        print("Initializing LLM Agent...")
        self.hf_token = hf_token
        self.llm = None
        
        try:
            # Using a smaller model that works better in Spaces
            self.llm = pipeline(
                "text-generation",
                model=HF_MODEL_NAME,
                token=hf_token,
                device_map="auto"
            )
            print("LLM initialized successfully")
        except Exception as e:
            print(f"Error initializing LLM: {e}")
            # Fallback to simple responses if LLM fails
            self.llm = None

    def __call__(self, question: str) -> str:
        if not self.llm:
            return "This is a default answer (LLM not available)"
            
        try:
            print(f"Generating answer for: {question[:50]}...")
            response = self.llm(
                question,
                max_length=100,
                do_sample=True,
                temperature=0.7
            )
            return response[0]['generated_text']
        except Exception as e:
            print(f"Error generating answer: {e}")
            return f"Error generating answer: {e}"

def run_and_submit_all(request: gr.Request):
    """
    Modified to work with Gradio's auth system
    """
    # Get username from auth
    if not request.username:
        return "Please login with Hugging Face account", None
    
    username = request.username
    space_id = os.getenv("SPACE_ID")
    api_url = DEFAULT_API_URL
    questions_url = f"{api_url}/questions"
    submit_url = f"{api_url}/submit"

    # 1. Instantiate Agent
    try:
        agent = BasicAgent(hf_token=os.getenv("HF_TOKEN"))
    except Exception as e:
        return f"Error initializing agent: {e}", None
    
    agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"

    # 2. Fetch Questions
    try:
        response = requests.get(questions_url, timeout=15)
        response.raise_for_status()
        questions_data = response.json()
        if not questions_data:
            return "No questions received from server", None
    except Exception as e:
        return f"Error fetching questions: {e}", None

    # 3. Process Questions
    results_log = []
    answers_payload = []
    for item in questions_data:
        task_id = item.get("task_id")
        question_text = item.get("question")
        if not task_id or not question_text:
            continue
            
        try:
            answer = agent(question_text)
            answers_payload.append({
                "task_id": task_id,
                "submitted_answer": answer
            })
            results_log.append({
                "Task ID": task_id,
                "Question": question_text,
                "Submitted Answer": answer
            })
        except Exception as e:
            results_log.append({
                "Task ID": task_id,
                "Question": question_text,
                "Submitted Answer": f"ERROR: {str(e)}"
            })

    if not answers_payload:
        return "No valid answers generated", pd.DataFrame(results_log)

    # 4. Submit Answers
    submission_data = {
        "username": username,
        "agent_code": agent_code,
        "answers": answers_payload
    }
    
    try:
        response = requests.post(submit_url, json=submission_data, timeout=60)
        response.raise_for_status()
        result = response.json()
        
        status = (
            f"Submission Successful!\n"
            f"User: {result.get('username')}\n"
            f"Score: {result.get('score', 'N/A')}% "
            f"({result.get('correct_count', '?')}/{result.get('total_attempted', '?')})\n"
            f"Message: {result.get('message', '')}"
        )
        return status, pd.DataFrame(results_log)
    except Exception as e:
        return f"Submission failed: {str(e)}", pd.DataFrame(results_log)

# --- Gradio Interface ---
with gr.Blocks() as demo:
    gr.Markdown("# LLM Agent Evaluation Runner")
    gr.Markdown("""
        **Instructions:**
        1. Log in with your Hugging Face account
        2. Click 'Run Evaluation'
        3. View your results
    """)

    gr.LoginButton()
    
    with gr.Row():
        run_btn = gr.Button("Run Evaluation & Submit Answers", variant="primary")
    
    status_output = gr.Textbox(label="Status", interactive=False)
    results_table = gr.DataFrame(label="Results", wrap=True)

    run_btn.click(
        fn=run_and_submit_all,
        inputs=[],
        outputs=[status_output, results_table]
    )

if __name__ == "__main__":
    demo.launch()