app.py

import os
import asyncio
import argparse
import gradio as gr
import requests
import pandas as pd
from agno.agent import RunResponse
from agent import agent

DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"


async def _async_answer(answer_text: str) -> str:
    response: RunResponse = await agent.arun(answer_text)
    return response.content


class BasicAgent:
    def __init__(self):
        pass

    def __call__(self, question: str) -> str:
        return asyncio.run(_async_answer(question))


def run_agent(profile: gr.OAuthProfile | None, task_id: str | None = None, submit: bool = True):
    space_id = os.getenv("SPACE_ID")
    if profile:
        username = f"{profile.username}"
    else:
        return "Please log in to Hugging Face.", None

    api_url = DEFAULT_API_URL
    questions_url = f"{api_url}/questions"
    submit_url = f"{api_url}/submit"

    try:
        agent_instance = BasicAgent()
    except Exception as e:
        return f"Error initializing agent: {e}", None

    agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"

    try:
        response = requests.get(questions_url, timeout=15)
        response.raise_for_status()
        questions_data = response.json()
    except Exception as e:
        return f"Error fetching questions: {e}", None

    if task_id:
        questions_data = [q for q in questions_data if str(q.get("task_id")) == str(task_id)]
        if not questions_data:
            return f"Task {task_id} not found.", None

    results_log = []
    answers_payload = []
    for item in questions_data:
        tid = item.get("task_id")
        qtext = item.get("question")
        if not tid or qtext is None:
            continue
        try:
            submitted_answer = agent_instance(qtext)
            answers_payload.append({"task_id": tid, "submitted_answer": submitted_answer})
            results_log.append({"Task ID": tid, "Question": qtext, "Submitted Answer": submitted_answer})
        except Exception as e:
            results_log.append({"Task ID": tid, "Question": qtext, "Submitted Answer": f"AGENT ERROR: {e}"})

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

    if not submit:
        return "Test mode: nothing submitted.", pd.DataFrame(results_log)

    submission_data = {
        "username": username.strip(),
        "agent_code": agent_code,
        "answers": answers_payload,
    }

    try:
        response = requests.post(submit_url, json=submission_data, timeout=60)
        response.raise_for_status()
        result_data = response.json()
        final_status = (
            f"Submission Successful\n"
            f"User: {result_data.get('username')}\n"
            f"Score: {result_data.get('score', 'N/A')}% "
            f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')})\n"
            f"Message: {result_data.get('message', '')}"
        )
        return final_status, pd.DataFrame(results_log)
    except Exception as e:
        return f"Submission failed: {e}", pd.DataFrame(results_log)


def run_agent_single(profile: gr.OAuthProfile | None, task_id: str):
    return run_agent(profile, task_id or None, submit=False)


def run_agent_all(profile: gr.OAuthProfile | None, task_id: str):
    return run_agent(profile, task_id or None, submit=True)


with gr.Blocks() as demo:
    gr.Markdown("# Basic Agent Evaluation Runner")

    gr.LoginButton()

    task_id_input = gr.Textbox(label="Task ID (optional)", placeholder="e.g. 2023060607")
    run_test_button = gr.Button("Test Single Task (no submit)")
    run_all_button = gr.Button("Run & Submit All")

    status_output = gr.Textbox(label="Status", lines=5, interactive=False)
    results_table = gr.DataFrame(label="Results", wrap=True)

    run_test_button.click(
        fn=run_agent_single,
        inputs=[task_id_input],
        outputs=[status_output, results_table],
    )

    run_all_button.click(
        fn=run_agent_all,
        inputs=[task_id_input],
        outputs=[status_output, results_table],
    )

    gr.Markdown(
        "Running all tasks may take time. Use the single‑task button to debug quickly."
    )

if __name__ == "__main__":
    space_host = os.getenv("SPACE_HOST")
    space_id = os.getenv("SPACE_ID")
    if space_host:
        print(f"SPACE_HOST: {space_host}")
    if space_id:
        print(f"SPACE_ID: {space_id}")

    parser = argparse.ArgumentParser()
    parser.add_argument("--task-id", help="Run a single task locally without submission")
    args, _ = parser.parse_known_args()

    if args.task_id:
        status, table = run_agent(profile=None, task_id=args.task_id, submit=False)
        print(status)
        if table is not None:
            print(table)
    else:
        demo.launch(debug=True, share=False)