Spaces:
Sleeping
Sleeping
| import os | |
| import gradio as gr | |
| import requests | |
| import inspect | |
| import pandas as pd | |
| from langgraph.prebuilt import ToolNode, create_react_agent | |
| # from typing import Any, Dict | |
| # from typing import TypedDict, Annotated | |
| from langchain_openai import ChatOpenAI | |
| from langgraph.graph import StateGraph, START, END | |
| from langgraph.graph.message import add_messages | |
| from langchain.schema import HumanMessage, AIMessage, SystemMessage | |
| # Create a ToolNode that knows about your web_search function | |
| import json | |
| from state import AgentState | |
| # --- Constants --- | |
| DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space" | |
| from tools import ocr_image_tool, parse_excel_tool, web_search_tool, run_tools | |
| tool_node = ToolNode([ocr_image_tool, parse_excel_tool, web_search_tool]) | |
| llm = ChatOpenAI(model_name="gpt-4.1-mini", temperature=0.0) | |
| agent = create_react_agent(model=llm, tools=tool_node) | |
| # ─── Revised plan_node with NO extra arguments ─── | |
| def plan_node(state: AgentState) -> AgentState: | |
| """ | |
| Assumes that `state["messages"]` already ends with a HumanMessage of the user’s question. | |
| We look at that last HumanMessage, append it to our new history, and ask the LLM | |
| to set exactly one key in a Python dict: web_search_query, ocr_path, | |
| excel_path (+ excel_sheet_name), or final_answer. | |
| """ | |
| # 1) Grab all prior BaseMessage objects (SystemMessage/HumanMessage/AIMessage) from state | |
| prior_msgs = state.get("messages", []) | |
| # 2) Find the very last HumanMessage (the user_input). We assume the last message is one. | |
| # If there is no HumanMessage, we treat user_input as empty. | |
| user_input = "" | |
| for msg in reversed(prior_msgs): | |
| if isinstance(msg, HumanMessage): | |
| user_input = msg.content | |
| break | |
| # 3) Build our new chat history by re‐using prior_msgs. It already includes that HumanMessage. | |
| new_history = prior_msgs.copy() | |
| # 4) Add a SystemMessage that instructs the LLM how to choose exactly one key | |
| explanation = SystemMessage( | |
| content=( | |
| "You can set exactly one of the following keys in a Python dict, and nothing else:\n" | |
| " • web_search_query: <search terms> \n" | |
| " • ocr_path: <path to an image file> \n" | |
| " • excel_path: <path to a .xlsx file> \n" | |
| " • excel_sheet_name: <sheet name> \n" | |
| "Or, if no tool is needed, set final_answer: <your answer>.\n" | |
| "Example: {'web_search_query':'Mercedes Sosa discography'}\n" | |
| "Respond with only that Python dict literal—no extra text or explanation." | |
| ) | |
| ) | |
| # 5) Compose the prompt as a list of BaseMessage, then call the LLM | |
| prompt_messages = new_history + [explanation] | |
| llm_response = llm(prompt_messages) | |
| llm_out = llm_response.content.strip() | |
| # 6) Parse the LLM’s output as a dict | |
| try: | |
| parsed = eval(llm_out, {}, {}) | |
| if isinstance(parsed, dict): | |
| partial: AgentState = {"messages": new_history} | |
| allowed = { | |
| "web_search_query", | |
| "ocr_path", | |
| "excel_path", | |
| "excel_sheet_name", | |
| "final_answer" | |
| } | |
| for k, v in parsed.items(): | |
| if k in allowed: | |
| partial[k] = v | |
| return partial | |
| except Exception: | |
| pass | |
| # 7) Fallback if parsing failed | |
| return { | |
| "messages": new_history, | |
| "final_answer": "Sorry, I could not parse your intent." | |
| } | |
| # ─── Revised finalize_node with NO extra arguments ─── | |
| def finalize_node(state: AgentState) -> AgentState: | |
| """ | |
| Assumes that `state['messages']` is a list of BaseMessage, possibly ending in an AIMessage | |
| (or plan_node may have set final_answer directly). We append any tool results | |
| as SystemMessages, then prompt the LLM for one final answer. | |
| """ | |
| # 1) Copy the existing BaseMessage list | |
| history = state.get("messages", []).copy() | |
| # 2) If any tool-result fields exist, append them as SystemMessages | |
| if "web_search_result" in state and state["web_search_result"] is not None: | |
| history.append(SystemMessage(content=f"WEB_SEARCH_RESULT: {state['web_search_result']}")) | |
| if "ocr_result" in state and state["ocr_result"] is not None: | |
| history.append(SystemMessage(content=f"OCR_RESULT: {state['ocr_result']}")) | |
| if "excel_result" in state and state["excel_result"] is not None: | |
| history.append(SystemMessage(content=f"EXCEL_RESULT: {state['excel_result']}")) | |
| # 3) If plan_node already set final_answer, just return it: | |
| if state.get("final_answer") is not None: | |
| return {"final_answer": state["final_answer"]} | |
| # 4) Otherwise, ask the LLM to give the final answer now | |
| history.append(SystemMessage(content="Please provide the final answer now.")) | |
| llm_response = llm(history) | |
| return {"final_answer": llm_response.content.strip()} | |
| tool_node = ToolNode([web_search_tool, ocr_image_tool, parse_excel_tool]) | |
| # ─── 5) Build the StateGraph ─── | |
| graph = StateGraph(AgentState) | |
| # 5.a) Register nodes | |
| graph.add_node("plan", plan_node) | |
| graph.add_node("tools", tool_node) | |
| graph.add_node("run_tools", run_tools) | |
| graph.add_node("finalize", finalize_node) | |
| # 5.b) START → plan | |
| graph.add_edge(START, "plan") | |
| def route_plan(plan_out: AgentState) -> str: | |
| """ | |
| plan_out is exactly what plan_node returned (a partial AgentState). | |
| If it set any of the tool-request keys, route to 'tools'; otherwise 'finalize'. | |
| """ | |
| if plan_out.get("web_search_query") or plan_out.get("ocr_path") or plan_out.get("excel_path"): | |
| return "tools" | |
| return "finalize" | |
| graph.add_conditional_edges( | |
| "plan", | |
| route_plan, | |
| {"tools": "tools", "finalize": "finalize"} | |
| ) | |
| graph.add_edge("tools", "run_tools") | |
| # 5.e) run_tools → finalize | |
| graph.add_edge("run_tools", "finalize") | |
| # 5.f) finalize → END | |
| graph.add_edge("finalize", END) | |
| compiled_graph = graph.compile() | |
| def respond_to_input(user_input: str) -> str: | |
| """ | |
| Initialize with a SystemMessage (tools description) and the user’s question as a HumanMessage. | |
| Then run through plan → tools → run_tools → finalize. Return the "final_answer" from final_state. | |
| """ | |
| # 1) Create a SystemMessage that tells the agent about its tools | |
| system_msg = SystemMessage( | |
| content=( | |
| "You have access to exactly these tools:\n" | |
| " 1) web_search(query:str) → Returns the top search results for the query.\n" | |
| " 2) parse_excel(path:str, sheet_name:str) → Reads an Excel file and returns its contents.\n" | |
| " 3) ocr_image(path:str) → Runs OCR on an image and returns any detected text.\n\n" | |
| "If you need a tool, set exactly one of these keys in a Python‐dict:\n" | |
| " • web_search_query: <search terms>\n" | |
| " • ocr_path: <path to image>\n" | |
| " • excel_path: <path to xlsx>\n" | |
| " • excel_sheet_name: <sheet name>\n" | |
| "Otherwise, set final_answer: <your answer>.\n" | |
| "Respond with that Python dict literal—no extra text or explanation." | |
| ) | |
| ) | |
| # 2) Wrap the user_input in a HumanMessage | |
| human_msg = HumanMessage(content=user_input) | |
| # 3) Build the initial state so that "messages" contains both messages | |
| initial_state: AgentState = { | |
| "messages": [system_msg, human_msg], | |
| "user_input": user_input | |
| } | |
| # 4) Invoke the compiled graph (no second argument needed) | |
| final_state = compiled_graph.invoke(initial_state) | |
| # 5) Return the final answer (or a fallback if missing) | |
| return final_state.get("final_answer", "Error: No final answer generated.") | |
| class BasicAgent: | |
| def __init__(self): | |
| print("BasicAgent initialized.") | |
| def __call__(self, question: str) -> str: | |
| # print(f"Agent received question (first 50 chars): {question[:50]}...") | |
| # fixed_answer = "This is a default answer." | |
| # print(f"Agent returning fixed answer: {fixed_answer}") | |
| return respond_to_input(question) | |
| # return fixed_answer | |
| def run_and_submit_all( profile: gr.OAuthProfile | None): | |
| """ | |
| Fetches all questions, runs the BasicAgent on them, submits all answers, | |
| and displays the results. | |
| """ | |
| # --- Determine HF Space Runtime URL and Repo URL --- | |
| space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code | |
| 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 ( modify this part to create your agent) | |
| try: | |
| agent = BasicAgent() | |
| except Exception as e: | |
| print(f"Error instantiating agent: {e}") | |
| return f"Error initializing agent: {e}", None | |
| # 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) | |
| 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: | |
| print("Fetched questions list is empty.") | |
| return "Fetched questions list is empty or invalid format.", None | |
| print(f"Fetched {len(questions_data)} questions.") | |
| except requests.exceptions.RequestException as e: | |
| print(f"Error fetching questions: {e}") | |
| return f"Error fetching questions: {e}", None | |
| except requests.exceptions.JSONDecodeError as e: | |
| print(f"Error decoding JSON response from questions endpoint: {e}") | |
| print(f"Response text: {response.text[:500]}") | |
| return f"Error decoding server response for questions: {e}", None | |
| except Exception as e: | |
| print(f"An unexpected error occurred fetching questions: {e}") | |
| return f"An unexpected error occurred fetching questions: {e}", None | |
| # 3. Run your 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: | |
| print("Agent did not produce any answers to submit.") | |
| return "Agent did not produce any answers to submit.", pd.DataFrame(results_log) | |
| # 4. Prepare Submission | |
| submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload} | |
| status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..." | |
| print(status_update) | |
| # 5. Submit | |
| 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.") | |
| results_df = pd.DataFrame(results_log) | |
| return final_status, results_df | |
| 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 requests.exceptions.JSONDecodeError: | |
| error_detail += f" Response: {e.response.text[:500]}" | |
| status_message = f"Submission Failed: {error_detail}" | |
| print(status_message) | |
| results_df = pd.DataFrame(results_log) | |
| return status_message, results_df | |
| except requests.exceptions.Timeout: | |
| status_message = "Submission Failed: The request timed out." | |
| print(status_message) | |
| results_df = pd.DataFrame(results_log) | |
| return status_message, results_df | |
| except requests.exceptions.RequestException as e: | |
| status_message = f"Submission Failed: Network error - {e}" | |
| print(status_message) | |
| results_df = pd.DataFrame(results_log) | |
| return status_message, results_df | |
| except Exception as e: | |
| status_message = f"An unexpected error occurred during submission: {e}" | |
| print(status_message) | |
| results_df = pd.DataFrame(results_log) | |
| return status_message, results_df | |
| # --- 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("LangGraph version:", langgraph.__version__) | |
| 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 | |
| # import langgraph | |
| # print("▶︎ LangGraph version:", langgraph.__version__) | |
| 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) |