mbertin/Final_Assignment_AgentCourse
0
1import os2import gradio as gr3import requests4import inspect5import pandas as pd6 7# (Keep Constants as is)8# --- Constants ---9DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"10 11# --- Basic Agent Definition ---12# ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------13class BasicAgent:14 def __init__(self):15 print("BasicAgent initialized.")16 def __call__(self, question: str) -> str:17 print(f"Agent received question (first 50 chars): {question[:50]}...")18 fixed_answer = "This is a default answer."19 print(f"Agent returning fixed answer: {fixed_answer}")20 return fixed_answer21 22def run_and_submit_all( profile: gr.OAuthProfile | None):23 """24 Fetches all questions, runs the BasicAgent on them, submits all answers,25 and displays the results.26 """27 # --- Determine HF Space Runtime URL and Repo URL ---28 space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code29 30 if profile:31 username= f"{profile.username}"32 print(f"User logged in: {username}")33 else:34 print("User not logged in.")35 return "Please Login to Hugging Face with the button.", None36 37 api_url = DEFAULT_API_URL38 questions_url = f"{api_url}/questions"39 submit_url = f"{api_url}/submit"40 41 # 1. Instantiate Agent ( modify this part to create your agent)42 try:43 agent = BasicAgent()44 except Exception as e:45 print(f"Error instantiating agent: {e}")46 return f"Error initializing agent: {e}", None47 # In the case of an app running as a hugging Face space, this link points toward your codebase ( usefull for others so please keep it public)48 agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"49 print(agent_code)50 51 # 2. Fetch Questions52 print(f"Fetching questions from: {questions_url}")53 try:54 response = requests.get(questions_url, timeout=15)55 response.raise_for_status()56 questions_data = response.json()57 if not questions_data:58 print("Fetched questions list is empty.")59 return "Fetched questions list is empty or invalid format.", None60 print(f"Fetched {len(questions_data)} questions.")61 except requests.exceptions.RequestException as e:62 print(f"Error fetching questions: {e}")63 return f"Error fetching questions: {e}", None64 except requests.exceptions.JSONDecodeError as e:65 print(f"Error decoding JSON response from questions endpoint: {e}")66 print(f"Response text: {response.text[:500]}")67 return f"Error decoding server response for questions: {e}", None68 except Exception as e:69 print(f"An unexpected error occurred fetching questions: {e}")70 return f"An unexpected error occurred fetching questions: {e}", None71 72 # 3. Run your Agent73 results_log = []74 answers_payload = []75 print(f"Running agent on {len(questions_data)} questions...")76 for item in questions_data:77 task_id = item.get("task_id")78 question_text = item.get("question")79 if not task_id or question_text is None:80 print(f"Skipping item with missing task_id or question: {item}")81 continue82 try:83 submitted_answer = agent(question_text)84 answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})85 results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})86 except Exception as e:87 print(f"Error running agent on task {task_id}: {e}")88 results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})89 90 if not answers_payload:91 print("Agent did not produce any answers to submit.")92 return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)93 94 # 4. Prepare Submission 95 submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}96 status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."97 print(status_update)98 99 # 5. Submit100 print(f"Submitting {len(answers_payload)} answers to: {submit_url}")101 try:102 response = requests.post(submit_url, json=submission_data, timeout=60)103 response.raise_for_status()104 result_data = response.json()105 final_status = (106 f"Submission Successful!\n"107 f"User: {result_data.get('username')}\n"108 f"Overall Score: {result_data.get('score', 'N/A')}% "109 f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"110 f"Message: {result_data.get('message', 'No message received.')}"111 )112 print("Submission successful.")113 results_df = pd.DataFrame(results_log)114 return final_status, results_df115 except requests.exceptions.HTTPError as e:116 error_detail = f"Server responded with status {e.response.status_code}."117 try:118 error_json = e.response.json()119 error_detail += f" Detail: {error_json.get('detail', e.response.text)}"120 except requests.exceptions.JSONDecodeError:121 error_detail += f" Response: {e.response.text[:500]}"122 status_message = f"Submission Failed: {error_detail}"123 print(status_message)124 results_df = pd.DataFrame(results_log)125 return status_message, results_df126 except requests.exceptions.Timeout:127 status_message = "Submission Failed: The request timed out."128 print(status_message)129 results_df = pd.DataFrame(results_log)130 return status_message, results_df131 except requests.exceptions.RequestException as e:132 status_message = f"Submission Failed: Network error - {e}"133 print(status_message)134 results_df = pd.DataFrame(results_log)135 return status_message, results_df136 except Exception as e:137 status_message = f"An unexpected error occurred during submission: {e}"138 print(status_message)139 results_df = pd.DataFrame(results_log)140 return status_message, results_df141 142 143# --- Build Gradio Interface using Blocks ---144with gr.Blocks() as demo:145 gr.Markdown("# Basic Agent Evaluation Runner")146 gr.Markdown(147 """148 **Instructions:**149 150 1. Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...151 2. Log in to your Hugging Face account using the button below. This uses your HF username for submission.152 3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.153 154 ---155 **Disclaimers:**156 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).157 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.158 """159 )160 161 gr.LoginButton()162 163 run_button = gr.Button("Run Evaluation & Submit All Answers")164 165 status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)166 # Removed max_rows=10 from DataFrame constructor167 results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)168 169 run_button.click(170 fn=run_and_submit_all,171 outputs=[status_output, results_table]172 )173 174if __name__ == "__main__":175 print("\n" + "-"*30 + " App Starting " + "-"*30)176 # Check for SPACE_HOST and SPACE_ID at startup for information177 space_host_startup = os.getenv("SPACE_HOST")178 space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup179 180 if space_host_startup:181 print(f"✅ SPACE_HOST found: {space_host_startup}")182 print(f" Runtime URL should be: https://{space_host_startup}.hf.space")183 else:184 print("ℹ️ SPACE_HOST environment variable not found (running locally?).")185 186 if space_id_startup: # Print repo URLs if SPACE_ID is found187 print(f"✅ SPACE_ID found: {space_id_startup}")188 print(f" Repo URL: https://huggingface.co/spaces/{space_id_startup}")189 print(f" Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")190 else:191 print("ℹ️ SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")192 193 print("-"*(60 + len(" App Starting ")) + "\n")194 195 print("Launching Gradio Interface for Basic Agent Evaluation...")196 demo.launch(debug=True, share=False)