Agent added

agent.py

@@ -0,0 +1,198 @@
+from smolagents import CodeAgent, LiteLLMModel, DuckDuckGoSearchTool, WikipediaSearchTool, Tool, VisitWebpageTool
+import os
+import json
+import requests
+import pandas as pd
+from huggingface_hub import InferenceClient
+
+class getFile(Tool):
+    name = "getFileTool"
+    description = "This tool allows to download the file attached to the question"
+    output_type = "string"
+    inputs = {
+        "task_id": {
+            "type": "string",
+            "description": "The task id of the question",
+        },
+        "file_name": {
+            "type": "string",
+            "description": "The name of the file to download"
+        }
+    }
+
+    def forward(self, task_id: str, file_name: str) -> str:
+        # Download the file from the task id
+        file = f"https://agents-course-unit4-scoring.hf.space/files/{task_id}"
+        # Save the file with the file name
+        with open(file_name, "wb") as f:
+            f.write(requests.get(file).content)
+
+        # Return the file name  
+        return os.path.abspath(file_name)
+    
+class LoadXlsxFileTool(Tool):
+    name = "load_xlsx_file_tool"
+    description = """This tool loads xlsx file into pandas and returns it"""
+    inputs = {
+        "file_path": {"type": "string", "description": "File path"}
+    }
+    output_type = "object"
+
+    def forward(self, file_path: str) -> object:
+        return pd.read_excel(file_path)
+
+class LoadTextFileTool(Tool):
+    name = "load_text_file_tool"
+    description = """This tool loads any text file"""
+    inputs = {
+        "file_path": {"type": "string", "description": "File path"}
+    }
+    output_type = "string"
+
+    def forward(self, file_path: str) -> object:
+        with open(file_path, 'r', encoding='utf-8') as file:
+            return file.read() 
+
+class AudioToTextTool(Tool):
+    name = "audio_to_text_tool"
+    description = """This tool transcribes audio files into text"""
+    inputs = {
+        "file_path": {"type": "string", "description": "File path"}
+    }
+    output_type = "string"
+
+    def forward(self, file_path: str) -> str:
+        try:
+            # Check if file exists
+            if not os.path.exists(file_path):
+                return f"Error: File {file_path} does not exist"
+
+            # Read the audio file as raw bytes
+            with open(file_path, "rb") as f:
+                audio_data = f.read()
+
+            # Set up the API URL and headers
+            api_url = "https://router.huggingface.co/hf-inference/models/openai/whisper-large-v3-turbo"
+            headers = {
+                "Authorization": f"Bearer {os.getenv('HF_API_KEY')}",
+                "Content-Type": "audio/mpeg"  # Assuming MP3 format
+            }
+
+            # Make the API request
+            response = requests.post(api_url, headers=headers, data=audio_data)
+            response.raise_for_status()  # Raise an exception for bad status codes
+            
+            # Parse and return the response
+            output = response.json()
+            return output["text"]
+
+        except Exception as e:
+            return f"Error transcribing audio: {str(e)}"
+    
+class ImageAnalysisTool(Tool):
+    name = "image_analysis_tool"
+    description = """This tool analyzes images and returns the text and the information in the image"""
+    inputs = {
+        "task_id": {"type": "string", "description": "The task id of the question"},
+    }
+    output_type = "string"
+
+    def forward(self, task_id: str) -> str:
+        client = InferenceClient(
+            provider="nebius",
+            api_key=os.getenv("HF_API_KEY"),
+        )
+            
+        completion = client.chat.completions.create(
+            model="mistralai/Mistral-Small-3.1-24B-Instruct-2503",
+            messages=[
+                {
+                    "role": "user",
+                    "content": [
+                    {
+                        "type": "text",
+                        "text": "Describe this image in markdown format"
+                    },
+                    {
+                        "type": "image_url",
+                        "image_url": {
+                            "url": f"https://agents-course-unit4-scoring.hf.space/files/{task_id}"
+                        }
+                    }
+                ]
+            }
+        ],
+        )
+
+        return completion.choices[0].message.content
+
+
+def final_answer_formatting(answer, question):
+    model = LiteLLMModel(
+        model_id="gemini/gemini-2.0-flash",
+        api_key=os.getenv("GOOGLE_API_KEY"),
+    )
+
+    prompt = f"""
+    You are an AI assistant specialized in the GAIA benchmark. For the question provided, generate the answer in the exact format requested by the question. Do not include any other text or creative additions.
+
+    Question: {question}
+    Answer: {answer}
+    """
+
+    messages = [
+    {"role": "user", "content": [{"type": "text", "text": prompt}]}
+    ]
+
+    output = model(messages).content
+    return output
+
+
+web_agent = CodeAgent(
+    model=LiteLLMModel(
+        model_id="gemini/gemini-2.0-flash",
+        api_key=os.getenv("GOOGLE_API_KEY"),
+    ),
+    tools=[
+        WikipediaSearchTool(),
+        DuckDuckGoSearchTool(), 
+        VisitWebpageTool()
+    ],
+    add_base_tools=False,
+    additional_authorized_imports=[
+        "os", "requests", "inspect", "pandas", 
+        "datetime", "re", "bs4", "markdownify"
+    ],
+    max_steps=10,
+    name="web_agent",
+    description="This agent is used to search the web for information"
+)
+
+audio_agent = CodeAgent(
+    model=LiteLLMModel(
+        model_id="gemini/gemini-2.0-flash",
+        api_key=os.getenv("GOOGLE_API_KEY"),
+    ),
+    tools=[AudioToTextTool()],
+    add_base_tools=False,
+    max_steps=10,
+    name="audio_agent",
+    description="This agent is used to analyze and transcribe audio files"
+)
+
+
+manager_agent = CodeAgent(
+    name="manager_agent",
+    model=LiteLLMModel(
+        model_id="gemini/gemini-2.5-flash-preview-04-17",
+        api_key=os.getenv("GOOGLE_API_KEY"),
+    ),
+    tools=[getFile(), LoadXlsxFileTool(), LoadTextFileTool(), ImageAnalysisTool()],
+    managed_agents=[web_agent, audio_agent],
+    additional_authorized_imports=[
+        "pandas"
+    ],
+    planning_interval=5,
+    verbosity_level=1,
+    max_steps=10,
+)

app.py

@@ -3,25 +3,15 @@ import gradio as gr
 import requests
 import inspect
 import pandas as pd
+from agent import manager_agent, final_answer_formatting
 
 # (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:
-    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):
+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 manager_agent on them, submits all answers,
     and displays the results.
     """
     # --- Determine HF Space Runtime URL and Repo URL ---
@@ -38,12 +28,13 @@ 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. Use the manager_agent from agent.py
     try:
-        agent = BasicAgent()
+        agent = manager_agent
     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)
@@ -76,13 +67,22 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     for item in questions_data:
         task_id = item.get("task_id")
         question_text = item.get("question")
+        file_name = item.get("file_name")
         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})
+            # Format the input for the manager agent
+            agent_input = f"""
+                task_id: {task_id}
+                file_name: {file_name if file_name else None}
+                Question to answer: {question_text}
+            """
+            submitted_answer = agent(agent_input)
+            # Format the answer using the final_answer_formatting function
+            formatted_answer = final_answer_formatting(submitted_answer, question_text)
+            answers_payload.append({"task_id": task_id, "submitted_answer": formatted_answer})
+            results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": formatted_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}"})
@@ -142,14 +142,13 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
 
 # --- Build Gradio Interface using Blocks ---
 with gr.Blocks() as demo:
-    gr.Markdown("# Basic Agent Evaluation Runner")
+    gr.Markdown("# GAIA 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.
+        1. Log in to your Hugging Face account using the button below. This uses your HF username for submission.
+        2. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
 
         ---
         **Disclaimers:**
@@ -163,7 +162,6 @@ 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(
@@ -192,5 +190,5 @@ if __name__ == "__main__":
 
     print("-"*(60 + len(" App Starting ")) + "\n")
 
-    print("Launching Gradio Interface for Basic Agent Evaluation...")
+    print("Launching Gradio Interface for GAIA Agent Evaluation...")
     demo.launch(debug=True, share=False)

requirements.txt

@@ -1,2 +1,7 @@
 gradio
-requests
+requests
+pandas
+openpyxl
+huggingface-hub
+smolagents
+python-dotenv