csr-01/Final_Assignment_Template
0
1import os2import gradio as gr3import requests4import inspect5import pandas as pd6import yaml7from smolagents import CodeAgent,DuckDuckGoSearchTool, LiteLLMModel,load_tool,tool, GoogleSearchTool8 9# (Keep Constants as is)10# --- Constants ---11DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"12 13# --- Basic Agent Definition ---14# ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------15 16model = LiteLLMModel(17 model_id="gemini/gemini-2.0-flash", # you can see other model names here: https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models. It is important to prefix the name with "gemini/"18 api_key=os.getenv("GOOGLE_API_KEY"),19 max_tokens=819220)21 22 23# agent = CodeAgent(tools=[DuckDuckGoSearchTool()], model=model)24agent = CodeAgent(tools=[GoogleSearchTool("serper")], model=model)25 26class BasicAgent:27 def __init__(self):28 print("BasicAgent initialized.")29 def __call__(self, question: str) -> str:30 print(f"Agent received question (first 50 chars): {question[:50]}...")31 32 fixed_answer = agent.run(f"""33 You are a helpful assistant tasked with answering questions using a set of tools. Now, I will ask you a question. Report your thoughts, and finish your answer with the following template: 34 FINAL ANSWER: [YOUR FINAL ANSWER]. 35 YOUR FINAL ANSWER should be a number OR as few words as possible OR a comma separated list of numbers and/or strings. If you are asked for a number, don't use comma to write your number neither use units such as $ or percent sign unless specified otherwise. If you are asked for a string, don't use articles, neither abbreviations (e.g. for cities), and write the digits in plain text unless specified otherwise. If you are asked for a comma separated list, apply the above rules depending of whether the element to be put in the list is a number or a string.36 Your answer should only start with "FINAL ANSWER: ", then follows with the answer.37 38 {question}39 """)40 print(f"Agent returning fixed answer: {fixed_answer}")41 return fixed_answer42 43def run_and_submit_all( profile: gr.OAuthProfile | None):44 """45 Fetches all questions, runs the BasicAgent on them, submits all answers,46 and displays the results.47 """48 # --- Determine HF Space Runtime URL and Repo URL ---49 space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code50 51 if profile:52 username= f"{profile.username}"53 print(f"User logged in: {username}")54 else:55 print("User not logged in.")56 return "Please Login to Hugging Face with the button.", None57 58 api_url = DEFAULT_API_URL59 questions_url = f"{api_url}/questions"60 submit_url = f"{api_url}/submit"61 62 # 1. Instantiate Agent ( modify this part to create your agent)63 try:64 agent = BasicAgent()65 except Exception as e:66 print(f"Error instantiating agent: {e}")67 return f"Error initializing agent: {e}", None68 # 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)69 agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"70 print(agent_code)71 72 # 2. Fetch Questions73 print(f"Fetching questions from: {questions_url}")74 try:75 response = requests.get(questions_url, timeout=15)76 response.raise_for_status()77 questions_data = response.json()78 if not questions_data:79 print("Fetched questions list is empty.")80 return "Fetched questions list is empty or invalid format.", None81 print(f"Fetched {len(questions_data)} questions.")82 except requests.exceptions.RequestException as e:83 print(f"Error fetching questions: {e}")84 return f"Error fetching questions: {e}", None85 except requests.exceptions.JSONDecodeError as e:86 print(f"Error decoding JSON response from questions endpoint: {e}")87 print(f"Response text: {response.text[:500]}")88 return f"Error decoding server response for questions: {e}", None89 except Exception as e:90 print(f"An unexpected error occurred fetching questions: {e}")91 return f"An unexpected error occurred fetching questions: {e}", None92 93 # 3. Run your Agent94 results_log = []95 answers_payload = []96 print(f"Running agent on {len(questions_data)} questions...")97 for item in questions_data:98 task_id = item.get("task_id")99 question_text = item.get("question")100 if not task_id or question_text is None:101 print(f"Skipping item with missing task_id or question: {item}")102 continue103 try:104 submitted_answer = agent(question_text)105 answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})106 results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})107 except Exception as e:108 print(f"Error running agent on task {task_id}: {e}")109 results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})110 111 if not answers_payload:112 print("Agent did not produce any answers to submit.")113 return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)114 115 # 4. Prepare Submission 116 submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}117 status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."118 print(status_update)119 120 # 5. Submit121 print(f"Submitting {len(answers_payload)} answers to: {submit_url}")122 try:123 response = requests.post(submit_url, json=submission_data, timeout=60)124 response.raise_for_status()125 result_data = response.json()126 final_status = (127 f"Submission Successful!\n"128 f"User: {result_data.get('username')}\n"129 f"Overall Score: {result_data.get('score', 'N/A')}% "130 f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"131 f"Message: {result_data.get('message', 'No message received.')}"132 )133 print("Submission successful.")134 results_df = pd.DataFrame(results_log)135 return final_status, results_df136 except requests.exceptions.HTTPError as e:137 error_detail = f"Server responded with status {e.response.status_code}."138 try:139 error_json = e.response.json()140 error_detail += f" Detail: {error_json.get('detail', e.response.text)}"141 except requests.exceptions.JSONDecodeError:142 error_detail += f" Response: {e.response.text[:500]}"143 status_message = f"Submission Failed: {error_detail}"144 print(status_message)145 results_df = pd.DataFrame(results_log)146 return status_message, results_df147 except requests.exceptions.Timeout:148 status_message = "Submission Failed: The request timed out."149 print(status_message)150 results_df = pd.DataFrame(results_log)151 return status_message, results_df152 except requests.exceptions.RequestException as e:153 status_message = f"Submission Failed: Network error - {e}"154 print(status_message)155 results_df = pd.DataFrame(results_log)156 return status_message, results_df157 except Exception as e:158 status_message = f"An unexpected error occurred during submission: {e}"159 print(status_message)160 results_df = pd.DataFrame(results_log)161 return status_message, results_df162 163 164# --- Build Gradio Interface using Blocks ---165with gr.Blocks() as demo:166 gr.Markdown("# Basic Agent Evaluation Runner")167 gr.Markdown(168 """169 **Instructions:**170 171 1. Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...172 2. Log in to your Hugging Face account using the button below. This uses your HF username for submission.173 3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.174 175 ---176 **Disclaimers:**177 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).178 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.179 """180 )181 182 gr.LoginButton()183 184 run_button = gr.Button("Run Evaluation & Submit All Answers")185 186 status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)187 # Removed max_rows=10 from DataFrame constructor188 results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)189 190 run_button.click(191 fn=run_and_submit_all,192 outputs=[status_output, results_table]193 )194 195if __name__ == "__main__":196 print("\n" + "-"*30 + " App Starting " + "-"*30)197 # Check for SPACE_HOST and SPACE_ID at startup for information198 space_host_startup = os.getenv("SPACE_HOST")199 space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup200 201 if space_host_startup:202 print(f"✅ SPACE_HOST found: {space_host_startup}")203 print(f" Runtime URL should be: https://{space_host_startup}.hf.space")204 else:205 print("ℹ️ SPACE_HOST environment variable not found (running locally?).")206 207 if space_id_startup: # Print repo URLs if SPACE_ID is found208 print(f"✅ SPACE_ID found: {space_id_startup}")209 print(f" Repo URL: https://huggingface.co/spaces/{space_id_startup}")210 print(f" Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")211 else:212 print("ℹ️ SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")213 214 print("-"*(60 + len(" App Starting ")) + "\n")215 216 print("Launching Gradio Interface for Basic Agent Evaluation...")217 demo.launch(debug=True, share=False)