push agent

app.py

@@ -1,196 +1,336 @@
 import os
+import re
+from dataclasses import dataclass
+from typing import Any
+
 import gradio as gr
-import requests
-import inspect
 import pandas as pd
+import requests
+from smolagents import CodeAgent, InferenceClientModel, tool
+from smolagents.default_tools import DuckDuckGoSearchTool, VisitWebpageTool
 
-# (Keep Constants as is)
-# --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
+DEFAULT_HF_MODEL = os.getenv("HF_MODEL", "Qwen/Qwen2.5-72B-Instruct")
+
+
+@dataclass
+class AgentConfig:
+    api_base_url: str = DEFAULT_API_URL
+    hf_model: str = DEFAULT_HF_MODEL
+    hf_token_env: str = "HF_TOKEN"
+    max_steps: int = 8
+    web_timeout_sec: int = 15
+    max_file_chars: int = 12000
+
+
+def normalize_answer(text: str) -> str:
+    value = (text or "").strip()
+    value = re.sub(r"^FINAL\s*:\s*", "", value, flags=re.IGNORECASE).strip()
+    value = value.strip().strip('"').strip()
+    value = value.replace("FINAL ANSWER:", "").replace("Final answer:", "").strip()
+    return value or "unknown"
+
+
+def fetch_questions(api_base_url: str) -> list[dict[str, Any]]:
+    response = requests.get(f"{api_base_url}/questions", timeout=20)
+    response.raise_for_status()
+    data = response.json()
+    if not isinstance(data, list):
+        raise ValueError("Invalid /questions response format.")
+    return data
+
+
+def submit_answers(api_base_url: str, payload: dict[str, Any]) -> dict[str, Any]:
+    response = requests.post(f"{api_base_url}/submit", json=payload, timeout=90)
+    response.raise_for_status()
+    return response.json()
+
+
+class GAIASmolAgent:
+    def __init__(self, config: AgentConfig):
+        self.config = config
+        token = os.getenv(config.hf_token_env)
+        if not token:
+            raise ValueError(f"Missing required secret: {config.hf_token_env}")
+
+        self.model = InferenceClientModel(
+            model_id=config.hf_model,
+            token=token,
+            temperature=0,
+            max_tokens=1200,
+        )
+        self.http = requests.Session()
+        self.http.headers.update({"User-Agent": "gaia-smolagent/1.0"})
+
+        @tool
+        def fetch_gaia_file(task_id: str) -> str:
+            """
+            Fetch and read the file attached to a GAIA task.
+
+            Args:
+                task_id: The GAIA task id.
+            """
+            url = f"{self.config.api_base_url}/files/{task_id}"
+            try:
+                response = self.http.get(url, timeout=self.config.web_timeout_sec)
+                if response.status_code >= 400:
+                    return f"TOOL_ERROR: could not fetch file for task {task_id}. HTTP {response.status_code}"
+                content_type = (response.headers.get("content-type") or "").lower()
+                if "text" in content_type or "json" in content_type or "csv" in content_type:
+                    text = response.text
+                    text = re.sub(r"\s+", " ", text).strip()
+                    if len(text) > self.config.max_file_chars:
+                        text = text[: self.config.max_file_chars] + " ...[truncated]"
+                    return text
+                size = len(response.content or b"")
+                return f"Binary file fetched. Content-Type: {content_type or 'unknown'}, bytes: {size}"
+            except requests.RequestException as e:
+                return f"TOOL_ERROR: request failed: {e}"
+
+        @tool
+        def add_numbers(a: float, b: float) -> float:
+            """
+            Add two numbers.
+
+            Args:
+                a: First number.
+                b: Second number.
+            """
+            return a + b
+
+        @tool
+        def subtract_numbers(a: float, b: float) -> float:
+            """
+            Subtract two numbers.
+
+            Args:
+                a: First number.
+                b: Second number.
+            """
+            return a - b
+
+        @tool
+        def multiply_numbers(a: float, b: float) -> float:
+            """
+            Multiply two numbers.
+
+            Args:
+                a: First number.
+                b: Second number.
+            """
+            return a * b
+
+        @tool
+        def divide_numbers(a: float, b: float) -> float:
+            """
+            Divide two numbers.
+
+            Args:
+                a: Numerator.
+                b: Denominator.
+            """
+            if b == 0:
+                return float("inf")
+            return a / b
+
+        @tool
+        def power_number(base: float, exponent: float) -> float:
+            """
+            Raise a number to a power.
+
+            Args:
+                base: Base value.
+                exponent: Exponent value.
+            """
+            return base**exponent
+
+        self.agent = CodeAgent(
+            model=self.model,
+            tools=[
+                fetch_gaia_file,
+                add_numbers,
+                subtract_numbers,
+                multiply_numbers,
+                divide_numbers,
+                power_number,
+                DuckDuckGoSearchTool(),
+                VisitWebpageTool(),
+            ],
+            max_steps=self.config.max_steps,
+            add_base_tools=False,
+        )
+
+    def solve_task(self, task_id: str, question: str) -> tuple[str, dict[str, Any]]:
+        prompt = (
+            "You are solving one GAIA benchmark question.\n"
+            "You must use tools when needed (duckduckgo search, webpage visit, arithmetic, fetch_gaia_file).\n"
+            "Critical scoring rule: exact match. Return only the final answer text, nothing else.\n"
+            "Never include labels like 'FINAL ANSWER'.\n\n"
+            f"Task ID: {task_id}\n"
+            f"Question: {question}\n\n"
+            "If the question depends on an attached file, call fetch_gaia_file(task_id) with the exact task id."
+        )
+        result = self.agent.run(prompt, reset=True)
+        answer = normalize_answer(str(result))
+        meta = {
+            "status": "ok",
+            "steps": len(getattr(self.agent, "logs", []) or []),
+            "tools": "smolagents",
+        }
+        return answer, meta
+
+
+def _agent_code_url() -> str:
+    space_id = os.getenv("SPACE_ID")
+    if space_id:
+        return f"https://huggingface.co/spaces/{space_id}/tree/main"
+    return "https://huggingface.co/spaces/unknown/tree/main"
+
+
+def generate_answers(profile: gr.OAuthProfile | None):
+    if not profile:
+        return "Please login to Hugging Face first.", None, [], ""
+
+    username = profile.username.strip()
+    config = AgentConfig()
 
-# --- 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) -> 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 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()
+        questions = fetch_questions(config.api_base_url)
     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}")
+        return f"Failed to fetch questions: {e}", None, [], username
+
     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
+        agent = GAIASmolAgent(config=config)
     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:
+        return f"Failed to initialize smolagents agent: {e}", None, [], username
+
+    answers_payload: list[dict[str, str]] = []
+    rows: list[dict[str, Any]] = []
+
+    for item in questions:
         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})
+            answer, meta = agent.solve_task(task_id=str(task_id), question=str(question_text))
+            answers_payload.append({"task_id": str(task_id), "submitted_answer": answer})
+            rows.append(
+                {
+                    "Task ID": str(task_id),
+                    "Question": str(question_text),
+                    "Submitted Answer": answer,
+                    "Status": meta["status"],
+                    "Steps": meta["steps"],
+                    "Tools": meta["tools"],
+                }
+            )
         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}"})
+            rows.append(
+                {
+                    "Task ID": str(task_id),
+                    "Question": str(question_text),
+                    "Submitted Answer": "unknown",
+                    "Status": f"agent_error: {e}",
+                    "Steps": 0,
+                    "Tools": "smolagents",
+                }
+            )
+
+    if not answers_payload:
+        return "No answers were generated.", pd.DataFrame(rows), [], username
+
+    status = (
+        f"Generated {len(answers_payload)} answers for user '{username}'. "
+        "Review the table, then click submit."
+    )
+    return status, pd.DataFrame(rows), answers_payload, username
+
 
+def submit_generated_answers(answers_payload: list[dict[str, str]], username: str):
+    if not username:
+        return "Missing username in session. Click 'Generate Answers' after logging in."
     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)
+        return "No generated answers found. Click 'Generate Answers' first."
 
-    # 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)
+    clean_answers: list[dict[str, str]] = []
+    for item in answers_payload:
+        task_id = str(item.get("task_id", "")).strip()
+        submitted = normalize_answer(str(item.get("submitted_answer", "")))
+        if not task_id:
+            continue
+        clean_answers.append({"task_id": task_id, "submitted_answer": submitted})
+
+    if not clean_answers:
+        return "Generated answers are invalid or empty."
+
+    payload = {
+        "username": username,
+        "agent_code": _agent_code_url(),
+        "answers": clean_answers,
+    }
 
-    # 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 = (
+        result = submit_answers(DEFAULT_API_URL, payload)
+        return (
             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.')}"
+            f"User: {result.get('username', 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.')}"
         )
-        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}."
+        detail = f"HTTP {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
+            body = e.response.json()
+            detail = f"{detail} - {body.get('detail', body)}"
+        except Exception:
+            detail = f"{detail} - {e.response.text[:500]}"
+        return f"Submission failed: {detail}"
     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
+        return f"Submission failed: {e}"
 
 
-# --- Build Gradio Interface using Blocks ---
 with gr.Blocks() as demo:
-    gr.Markdown("# Basic Agent Evaluation Runner")
+    gr.Markdown("# GAIA smolagents Runner")
     gr.Markdown(
         """
-        **Instructions:**
+        Two-step flow:
+        1. Generate answers for all tasks.
+        2. Submit generated answers to leaderboard scoring.
 
-        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.
+        Required Space secrets:
+        - `HF_TOKEN`
+        Optional:
+        - `HF_MODEL` (default: `Qwen/Qwen2.5-72B-Instruct`)
         """
     )
 
     gr.LoginButton()
+    generated_answers_state = gr.State([])
+    username_state = gr.State("")
 
-    run_button = gr.Button("Run Evaluation & Submit All Answers")
+    with gr.Row():
+        generate_button = gr.Button("1) Generate Answers", variant="primary")
+        submit_button = gr.Button("2) Submit Generated 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)
+    status_output = gr.Textbox(label="Status", lines=6, interactive=False)
+    results_table = gr.DataFrame(label="Generated Answers", wrap=True)
 
-    run_button.click(
-        fn=run_and_submit_all,
-        outputs=[status_output, results_table]
+    generate_button.click(
+        fn=generate_answers,
+        outputs=[status_output, results_table, generated_answers_state, username_state],
+    )
+    submit_button.click(
+        fn=submit_generated_answers,
+        inputs=[generated_answers_state, username_state],
+        outputs=[status_output],
     )
 
+
 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)
+    print("\n" + "-" * 30 + " App Starting " + "-" * 30)
+    print(f"SPACE_HOST: {os.getenv('SPACE_HOST', 'not set')}")
+    print(f"SPACE_ID: {os.getenv('SPACE_ID', 'not set')}")
+    print("-" * (60 + len(" App Starting ")) + "\n")
+    demo.launch(debug=True, share=False)

requirements.txt

@@ -1,2 +1,5 @@
 gradio
-requests
+requests
+pandas
+huggingface_hub
+smolagents[toolkit]