import os import re import base64 from enum import Enum from pydantic import BaseModel from io import BytesIO from tempfile import SpooledTemporaryFile from typing import Optional import logging import gradio as gr import requests import pandas as pd from langchain_core.messages import HumanMessage from concurrent.futures import ThreadPoolExecutor, as_completed from agent import gaia_agent logging.basicConfig(level=logging.INFO) logger = logging.getLogger(__name__) class ContentType(Enum): IMAGE = "image" PDF = "pdf" AUDIO = "audio" TEXT = "text" # (Keep Constants as is) # --- Constants --- DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space" MAX_WORKERS = 8 class LLMFile(BaseModel): filename: str file: bytes mime: str content_type: ContentType # --- Basic Agent Definition --- # ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------ class BasicAgent: def __init__(self): print("BasicAgent initialized.") def __call__(self, question: str, content: Optional[LLMFile]) -> str: print(f"Agent received question (first 50 chars): {question[:50]}...") message_content = [{"type": "text", "text": question}] if content: if content.content_type == ContentType.AUDIO: media = { "type": "input_audio", "input_audio": {"data": base64.b64encode(content.file).encode("ascii"), "format": "wav"} } elif content.content_type == ContentType.IMAGE: media = { "type": "image", "image_url": {"url": f"data:image/jpeg;base64,{base64.b64encode(content.file).encode("ascii")}"} } elif content.content_type == ContentType.PDF: media = { "type": "file", "file": { "filename": content.filename, "file_data": f"data:application/pdf;base64,{base64.b64encode(content.file).encode("ascii")}", } } message_content.append(media) messages = gaia_agent.invoke({"messages": [ HumanMessage(content=message_content) ]}) message = messages['messages'][-1].content match = re.search(r'FINAL ANSWER:\s*(.*)', message) if match: answer = match.group(1) else: answer = "ERROR" print(f"Agent returning answer: {answer}") return answer def run_and_submit_all(profile: Optional[gr.OAuthProfile]): if not profile: logger.warning("User not logged in.") return "Please Login to Hugging Face with the button.", None username = profile.username.strip() logger.info(f"User logged in: {username}") session = requests.Session() space_id = os.getenv("SPACE_ID") agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main" # --- Fetch questions --- questions_url = f"{DEFAULT_API_URL}/questions" try: response = session.get(questions_url, timeout=15) response.raise_for_status() questions_data = response.json() if not questions_data: raise ValueError("Fetched questions list is empty or invalid.") logger.info(f"Fetched {len(questions_data)} questions.") except Exception as e: logger.exception("Error fetching questions.") return f"Error fetching questions: {e}", None # --- Instantiate agent --- try: agent = BasicAgent() except Exception as e: logger.exception("Error initializing agent.") return f"Error initializing agent: {e}", None # --- Run agent in parallel --- def process_question(item): task_id = item.get("task_id") question = item.get("question") if not task_id or question is None: return None, {"Task ID": task_id, "Question": question, "Submitted Answer": "INVALID QUESTION FORMAT"} if item.get("filename", None): # --- Fetch file --- file_url = f"{DEFAULT_API_URL}/files/{task_id}" try: response = session.get(file_url, timeout=15) response.raise_for_status() content_disposition = response.headers.get("content-disposition", "") filename = task_id + ".bin" if "filename=" in content_disposition: filename = content_disposition.split("filename=")[-1].strip('"') mime = response.headers.get("content-type", "") if mime.startswith("audio/"): media = LLMFile(filename=filename, mime=mime, content_type=ContentType.AUDIO, file=response.content) elif mime.startswith("image/"): media = LLMFile(filename=filename, mime=mime, content_type=ContentType.IMAGE, file=response.content) elif mime.startswith("image/"): media = LLMFile(filename=filename, mime=mime, content_type=ContentType.IMAGE, file=response.content) elif mime.startswith("text/"): media = LLMFile(filename=filename, mime=mime, content_type=ContentType.TEXT, file=response.content) except Exception as e: logger.exception("Error fetching file for task id %s.", str(task_id)) return f"Error fetching file for task id ({task_id}): {e}", None try: answer = agent(question, media if item.get("filename", None) else None) return {"task_id": task_id, "submitted_answer": answer}, { "Task ID": task_id, "Question": question, "Submitted Answer": answer } except Exception as e: logger.warning(f"Agent error on task {task_id}: {e}") return None, { "Task ID": task_id, "Question": question, "Submitted Answer": f"AGENT ERROR: {e}" } answers_payload = [] results_log = [] with ThreadPoolExecutor(max_workers=MAX_WORKERS) as executor: futures = [executor.submit(process_question, item) for item in questions_data] for future in as_completed(futures): answer, log = future.result() if answer: answers_payload.append(answer) results_log.append(log) if not answers_payload: return "Agent did not produce any answers to submit.", pd.DataFrame(results_log) # --- Submit answers --- submit_url = f"{DEFAULT_API_URL}/submit" submission_data = { "username": username, "agent_code": agent_code, "answers": answers_payload, } logger.info(f"Submitting {len(answers_payload)} answers to: {submit_url}") try: response = session.post(submit_url, json=submission_data, timeout=60) response.raise_for_status() result = response.json() final_status = ( f"Submission Successful!\n" f"User: {result.get('username')}\n" f"Overall Score: {result.get('score', 'N/A')}% " f"({result.get('correct_count', '?')}/{result.get('total_attempted', '?')} correct)\n" f"Message: {result.get('message', 'No message received.')}" ) return final_status, pd.DataFrame(results_log) except requests.exceptions.HTTPError as e: try: error_detail = e.response.json().get("detail", e.response.text) except Exception: error_detail = e.response.text[:500] status_message = f"Submission Failed: {error_detail}" except requests.exceptions.Timeout: status_message = "Submission Failed: The request timed out." except requests.exceptions.RequestException as e: status_message = f"Submission Failed: Network error - {e}" except Exception as e: status_message = f"Unexpected error during submission: {e}" logger.error(status_message) return status_message, pd.DataFrame(results_log) # --- Build Gradio Interface using Blocks --- with gr.Blocks() as demo: gr.Markdown("# Basic Agent Evaluation Runner") gr.Markdown( """ **Instructions:** 1. Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ... 2. Log in to your Hugging Face account using the button below. This uses your HF username for submission. 3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score. --- **Disclaimers:** Once clicking on the "submit button, it can take quite some time ( this is the time for the agent to go through all the questions). This space provides a basic setup and is intentionally sub-optimal to encourage you to develop your own, more robust solution. For instance for the delay process of the submit button, a solution could be to cache the answers and submit in a seperate action or even to answer the questions in async. """ ) gr.LoginButton() run_button = gr.Button("Run Evaluation & Submit All Answers") status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False) # Removed max_rows=10 from DataFrame constructor results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True) run_button.click( fn=run_and_submit_all, outputs=[status_output, results_table] ) if __name__ == "__main__": print("\n" + "-"*30 + " App Starting " + "-"*30) # Check for SPACE_HOST and SPACE_ID at startup for information space_host_startup = os.getenv("SPACE_HOST") space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup if space_host_startup: print(f"✅ SPACE_HOST found: {space_host_startup}") print(f" Runtime URL should be: https://{space_host_startup}.hf.space") else: print("ℹ️ SPACE_HOST environment variable not found (running locally?).") if space_id_startup: # Print repo URLs if SPACE_ID is found print(f"✅ SPACE_ID found: {space_id_startup}") print(f" Repo URL: https://huggingface.co/spaces/{space_id_startup}") print(f" Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main") else: print("ℹ️ SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.") print("-"*(60 + len(" App Starting ")) + "\n") print("Launching Gradio Interface for Basic Agent Evaluation...") demo.launch(debug=True, share=False)