Update app.py

app.py

@@ -1,77 +1,156 @@
 import os
 import gradio as gr
 import requests
-import inspect
 import pandas as pd
-from smolagents import CodeAgent, DuckDuckGoSearchTool, VisitWebpageTool, LiteLLMModel
 
-# (Keep Constants as is)
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# --- Basic Agent Definition ---
-# ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
-class BasicAgent:
+# ─────────────────────────────────────────────────────────────
+# AGENT DEFINITION — smolagents + Qwen2.5-72B (HF Inference)
+# ─────────────────────────────────────────────────────────────
+from smolagents import (
+    CodeAgent,
+    HfApiModel,
+    DuckDuckGoSearchTool,
+    WikipediaSearchTool,
+    tool,
+)
+
+@tool
+def download_file_for_task(task_id: str) -> str:
+    """
+    Downloads the file associated with a GAIA task_id from the scoring API.
+    Returns the file content as text (or a description if binary).
+    Args:
+        task_id: The GAIA task identifier.
+    """
+    try:
+        url = f"{DEFAULT_API_URL}/files/{task_id}"
+        resp = requests.get(url, timeout=30)
+        if resp.status_code != 200:
+            return f"No file found for task {task_id} (HTTP {resp.status_code})"
+        content_type = resp.headers.get("Content-Type", "")
+        content = resp.content
+
+        # --- TEXT / JSON / CSV ---
+        if any(t in content_type for t in ["text", "json", "csv"]):
+            return content.decode("utf-8", errors="replace")[:4000]
+
+        # --- EXCEL ---
+        if "excel" in content_type or "spreadsheet" in content_type or task_id.endswith(".xlsx"):
+            import io, openpyxl
+            wb = openpyxl.load_workbook(io.BytesIO(content))
+            ws = wb.active
+            rows = ["\t".join(str(c) if c is not None else "" for c in row)
+                    for row in ws.iter_rows(values_only=True)]
+            return "\n".join(rows[:300])
+
+        # --- PDF ---
+        if "pdf" in content_type:
+            import io
+            try:
+                import pypdf
+                reader = pypdf.PdfReader(io.BytesIO(content))
+                text = "\n".join(p.extract_text() or "" for p in reader.pages[:10])
+                return text[:4000]
+            except Exception as e:
+                return f"[PDF parse error: {e}]"
+
+        # --- Fallback: try decoding as UTF-8 ---
+        return content.decode("utf-8", errors="replace")[:4000]
+
+    except Exception as e:
+        return f"Error downloading file: {e}"
+
+
+@tool
+def python_calculator(code: str) -> str:
+    """
+    Executes a Python code snippet and returns the printed output.
+    Use this for arithmetic, data processing, pandas operations, etc.
+    Args:
+        code: Valid Python code to execute. Use print() to output results.
+    """
+    import io, sys, traceback
+    old_stdout = sys.stdout
+    sys.stdout = buf = io.StringIO()
+    try:
+        exec(code, {"__builtins__": __builtins__, "pd": pd})
+        return buf.getvalue() or "Executed (no output). Use print() to see results."
+    except Exception:
+        return traceback.format_exc()
+    finally:
+        sys.stdout = old_stdout
+
+
+class GAIAAgent:
+    """
+    Wraps a smolagents CodeAgent powered by Qwen2.5-72B-Instruct.
+    Exposes a __call__(question) interface compatible with the template.
+    """
+
+    SYSTEM_PROMPT = """You are an expert AI assistant solving GAIA benchmark questions.
+Your answers must be SHORT and EXACT — a number, a name, a short phrase, or a comma-separated list.
+Never include explanations, preambles, or units unless explicitly asked.
+If you need to look something up, use your tools.
+At the end, output ONLY the final answer with no extra words."""
+
     def __init__(self):
-        print("=" * 50)
-        print("Initializing GAIA Agent...")
-        print("=" * 50)
-        
-        # Initialize the model
-        self.model = LiteLLMModel(
-            model_id="huggingface/Qwen/Qwen2.5-Coder-32B-Instruct",
-            max_tokens=4096,
-            temperature=0.1
+        hf_token = os.getenv("HF_TOKEN", "")
+        model = HfApiModel(
+            model_id="Qwen/Qwen2.5-72B-Instruct",
+            token=hf_token,
         )
-        
-        # Create CodeAgent with web tools
         self.agent = CodeAgent(
             tools=[
                 DuckDuckGoSearchTool(),
-                VisitWebpageTool(),
+                WikipediaSearchTool(),
+                download_file_for_task,
+                python_calculator,
             ],
-            model=self.model,
-            max_steps=15,
-            verbosity_level=0
+            model=model,
+            max_steps=6,
+            verbosity_level=1,
         )
-        print("✅ Agent initialized successfully!")
-        print("=" * 50)
-        
+        print("GAIAAgent initialized with Qwen2.5-72B + smolagents tools.")
+
     def __call__(self, question: str) -> str:
-        print(f"📝 Question: {question[:80]}...")
+        print(f"Agent received question (first 50 chars): {question[:50]}...")
         try:
-            # Prompt for exact answer
-            prompt = f"""Answer this question with ONLY the exact answer. No explanations, no extra text.
-
-Question: {question}
-
-Answer:"""
-            
+            # Inject task_id hint if present (not standard but useful for file tool)
+            prompt = (
+                f"{self.SYSTEM_PROMPT}\n\n"
+                f"Question: {question}\n\n"
+                "Give ONLY the final answer. No explanation. No sentence. Just the answer."
+            )
             result = self.agent.run(prompt)
-            answer = result.strip()
-            
-            # Clean up common prefixes
-            prefixes = ["final answer:", "answer:", "the answer is"]
-            for prefix in prefixes:
-                if answer.lower().startswith(prefix):
+            # smolagents returns the final answer as a string
+            answer = str(result).strip()
+            # Clean up common LLM verbosity
+            for prefix in ["Final answer:", "Answer:", "ANSWER:", "The answer is", "Result:"]:
+                if answer.lower().startswith(prefix.lower()):
                     answer = answer[len(prefix):].strip()
-            
-            print(f"✅ Answer: {answer[:80]}")
-            return answer if answer else "No answer found"
+            print(f"Agent answer: {answer}")
+            return answer
         except Exception as e:
-            print(f"❌ Error: {e}")
-            return ""
+            print(f"Agent error: {e}")
+            return "I don't know"
+
 
-def run_and_submit_all( profile: gr.OAuthProfile | None):
+# ─────────────────────────────────────────────────────────────
+# run_and_submit_all — structure du template conservée à 100%
+# ─────────────────────────────────────────────────────────────
+
+def run_and_submit_all(profile: gr.OAuthProfile | None):
     """
-    Fetches all questions, runs the BasicAgent on them, submits all answers,
+    Fetches all questions, runs the GAIAAgent 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
+    space_id = os.getenv("SPACE_ID")
 
     if profile:
-        username= f"{profile.username}"
+        username = f"{profile.username}"
         print(f"User logged in: {username}")
     else:
         print("User not logged in.")
@@ -81,13 +160,13 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     questions_url = f"{api_url}/questions"
     submit_url = f"{api_url}/submit"
 
-    # 1. Instantiate Agent ( modify this part to create your agent)
+    # 1. Instantiate Agent
     try:
-        agent = BasicAgent()
+        agent = GAIAAgent()
     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)
 
@@ -98,16 +177,16 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
         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("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
+        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
@@ -123,18 +202,20 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
             print(f"Skipping item with missing task_id or question: {item}")
             continue
         try:
-            submitted_answer = agent(question_text)
+            # Pass task_id via question context so file tool can use it
+            question_with_id = f"[task_id={task_id}] {question_text}"
+            submitted_answer = agent(question_with_id)
             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}"})
+            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 
+    # 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)
@@ -206,7 +287,6 @@ with gr.Blocks() as demo:
     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(
@@ -216,9 +296,8 @@ with gr.Blocks() as demo:
 
 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
+    space_id_startup = os.getenv("SPACE_ID")
 
     if space_host_startup:
         print(f"✅ SPACE_HOST found: {space_host_startup}")
@@ -226,7 +305,7 @@ if __name__ == "__main__":
     else:
         print("ℹ️  SPACE_HOST environment variable not found (running locally?).")
 
-    if space_id_startup: # Print repo URLs if SPACE_ID is found
+    if space_id_startup:
         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")
@@ -234,6 +313,5 @@ if __name__ == "__main__":
         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)