Spaces:
Sleeping
Sleeping
| import os | |
| import gradio as gr | |
| import requests | |
| import pandas as pd | |
| # --- Constants --- | |
| DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space" | |
| # --- Agent Definition --- | |
| def build_agent(): | |
| """Build and return the smolagents CodeAgent with Groq backend.""" | |
| from smolagents import CodeAgent, LiteLLMModel, DuckDuckGoSearchTool, WikipediaSearchTool, VisitWebpageTool, tool | |
| def download_task_file(task_id: str) -> str: | |
| """Download a file associated with a GAIA task and return its local path. | |
| Use this when a question mentions or implies there is an attached file. | |
| Args: | |
| task_id: The task ID whose file should be downloaded. | |
| Returns: | |
| The local file path where the file was saved, or an error message. | |
| """ | |
| url = f"{DEFAULT_API_URL}/files/{task_id}" | |
| try: | |
| resp = requests.get(url, timeout=30) | |
| if resp.status_code == 404: | |
| return "No file found for this task." | |
| resp.raise_for_status() | |
| # Try to determine file extension from Content-Disposition or Content-Type | |
| content_disp = resp.headers.get("content-disposition", "") | |
| if "filename=" in content_disp: | |
| filename = content_disp.split("filename=")[-1].strip().strip('"') | |
| else: | |
| ct = resp.headers.get("content-type", "") | |
| ext_map = { | |
| "image/png": ".png", "image/jpeg": ".jpg", "image/gif": ".gif", | |
| "application/pdf": ".pdf", "text/plain": ".txt", | |
| "text/csv": ".csv", "application/json": ".json", | |
| "audio/mpeg": ".mp3", "audio/wav": ".wav", | |
| "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet": ".xlsx", | |
| } | |
| ext = next((v for k, v in ext_map.items() if k in ct), ".bin") | |
| filename = f"task_{task_id}{ext}" | |
| path = f"/tmp/{filename}" | |
| with open(path, "wb") as f: | |
| f.write(resp.content) | |
| return path | |
| except Exception as e: | |
| return f"Error downloading file: {e}" | |
| openai_api_key = os.getenv("OPENAI_API_KEY") | |
| if not openai_api_key: | |
| raise ValueError("OPENAI_API_KEY environment variable not set. Add it as a Secret in your HF Space settings.") | |
| model = LiteLLMModel( | |
| model_id="openai/gpt-4o", | |
| api_key=openai_api_key, | |
| temperature=0.0, | |
| ) | |
| agent = CodeAgent( | |
| tools=[ | |
| DuckDuckGoSearchTool(), | |
| WikipediaSearchTool(), | |
| VisitWebpageTool(), | |
| download_task_file, | |
| ], | |
| model=model, | |
| additional_authorized_imports=[ | |
| "requests", "json", "re", "math", "datetime", | |
| "csv", "io", "os", "pathlib", | |
| "PIL", "PIL.Image", | |
| "pandas", "openpyxl", | |
| ], | |
| max_steps=15, | |
| ) | |
| return agent | |
| class BasicAgent: | |
| def __init__(self): | |
| print("Initializing agent (loading smolagents + Groq)...") | |
| self._agent = build_agent() | |
| print("Agent ready.") | |
| def __call__(self, question: str) -> str: | |
| print(f"Question: {question[:100]}...") | |
| system_note = ( | |
| "You are a precise research assistant solving GAIA benchmark questions. " | |
| "Your answers are graded by EXACT STRING MATCH, so formatting is critical.\n\n" | |
| "Rules:\n" | |
| "- Reply with ONLY the answer, nothing else. No explanation, no 'FINAL ANSWER:' prefix.\n" | |
| "- Numbers: use digits (e.g. 42, 3.14). No units unless the question asks for them.\n" | |
| "- Lists: comma-separated on one line unless the question specifies otherwise.\n" | |
| "- Names/strings: exact spelling, match the question's expected format.\n" | |
| "- If a file is attached to the question, use the download_task_file tool first.\n" | |
| "- Search the web and visit pages to verify facts before answering.\n" | |
| "- Think step by step, but output ONLY the final answer." | |
| ) | |
| full_prompt = f"{system_note}\n\nQuestion: {question}" | |
| try: | |
| result = self._agent.run(full_prompt) | |
| answer = str(result).strip() | |
| print(f"Answer: {answer[:100]}") | |
| return answer | |
| except Exception as e: | |
| print(f"Agent error: {e}") | |
| return f"ERROR: {e}" | |
| def run_and_submit_all(profile: gr.OAuthProfile | None): | |
| """ | |
| Fetches all questions, runs the BasicAgent on them, submits all answers, | |
| and displays the results. | |
| """ | |
| space_id = os.getenv("SPACE_ID") | |
| if profile: | |
| username = f"{profile.username}" | |
| print(f"User logged in: {username}") | |
| else: | |
| print("User not logged in.") | |
| return "Please Login to Hugging Face with the button.", None | |
| api_url = DEFAULT_API_URL | |
| questions_url = f"{api_url}/questions" | |
| submit_url = f"{api_url}/submit" | |
| # 1. Instantiate Agent | |
| try: | |
| agent = BasicAgent() | |
| except Exception as e: | |
| print(f"Error instantiating agent: {e}") | |
| return f"Error initializing agent: {e}", None | |
| agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main" | |
| print(agent_code) | |
| # 2. Fetch Questions | |
| print(f"Fetching questions from: {questions_url}") | |
| try: | |
| response = requests.get(questions_url, timeout=15) | |
| response.raise_for_status() | |
| questions_data = response.json() | |
| if not questions_data: | |
| return "Fetched questions list is empty or invalid format.", None | |
| print(f"Fetched {len(questions_data)} questions.") | |
| except Exception as e: | |
| return f"Error fetching questions: {e}", None | |
| # 3. Run Agent | |
| results_log = [] | |
| answers_payload = [] | |
| print(f"Running agent on {len(questions_data)} questions...") | |
| for item in questions_data: | |
| task_id = item.get("task_id") | |
| question_text = item.get("question") | |
| if not task_id or question_text is None: | |
| print(f"Skipping item with missing task_id or question: {item}") | |
| continue | |
| try: | |
| submitted_answer = agent(question_text) | |
| answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer}) | |
| results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer}) | |
| except Exception as e: | |
| print(f"Error running agent on task {task_id}: {e}") | |
| results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"}) | |
| if not answers_payload: | |
| return "Agent did not produce any answers to submit.", pd.DataFrame(results_log) | |
| # 4. Submit | |
| submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload} | |
| print(f"Submitting {len(answers_payload)} answers to: {submit_url}") | |
| 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"Overall Score: {result_data.get('score', 'N/A')}% " | |
| f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n" | |
| f"Message: {result_data.get('message', 'No message received.')}" | |
| ) | |
| print("Submission successful.") | |
| return final_status, pd.DataFrame(results_log) | |
| except requests.exceptions.HTTPError as e: | |
| error_detail = f"Server responded with status {e.response.status_code}." | |
| try: | |
| error_json = e.response.json() | |
| error_detail += f" Detail: {error_json.get('detail', e.response.text)}" | |
| except Exception: | |
| error_detail += f" Response: {e.response.text[:500]}" | |
| print(f"Submission Failed: {error_detail}") | |
| return f"Submission Failed: {error_detail}", pd.DataFrame(results_log) | |
| except Exception as e: | |
| print(f"Unexpected error during submission: {e}") | |
| return f"An unexpected error occurred during submission: {e}", pd.DataFrame(results_log) | |
| # --- Gradio Interface --- | |
| with gr.Blocks() as demo: | |
| gr.Markdown("# GAIA Agent Evaluation Runner") | |
| gr.Markdown( | |
| """ | |
| **Instructions:** | |
| 1. Make sure `GROQ_API_KEY` is set as a Secret in your HF Space settings. | |
| 2. Log in with your Hugging Face account below. | |
| 3. Click **Run Evaluation & Submit All Answers** — the agent will answer all 20 GAIA questions and submit. | |
| --- | |
| *Note: This can take several minutes as the agent processes each question.* | |
| """ | |
| ) | |
| gr.LoginButton() | |
| run_button = gr.Button("Run Evaluation & Submit All Answers") | |
| status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False) | |
| 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) | |
| space_host_startup = os.getenv("SPACE_HOST") | |
| space_id_startup = os.getenv("SPACE_ID") | |
| if space_host_startup: | |
| print(f"✅ SPACE_HOST found: {space_host_startup}") | |
| else: | |
| print("ℹ️ SPACE_HOST not found (running locally?).") | |
| if space_id_startup: | |
| print(f"✅ SPACE_ID found: {space_id_startup}") | |
| else: | |
| print("ℹ️ SPACE_ID not found (running locally?).") | |
| print("-" * (60 + len(" App Starting ")) + "\n") | |
| print("Launching Gradio Interface...") | |
| demo.launch(debug=True, share=False) | |