Updated the README file

.env.example

@@ -12,5 +12,5 @@ CHESS_ENGINE_PATH=stock-fish-engine-location-here
 
 # Configure this if you want to enable observability with LangSmith
 LANGSMITH_API_KEY=your-langsmith-key
-LANGSMITH_TRACING=true
+LANGSMITH_TRACING=false
 LANGSMITH_PROJECT=gaia_agent

README.md

@@ -39,7 +39,8 @@ GAIA introductory paper [”GAIA: A Benchmark for General AI Assistants”](http
 
 **Nodes**
 
-- **Pre-processor**: Pre-processing of the question. 
+- **Pre-processor**: Initialization and preparation of the state. Input handling: attached pictures are directly sent to the assistant node and to the model.
+Other type of attachments are loaded only in the tools.
 - **Assistant**: The brain of the agent. Decides which tool to call. Uses `gpt-4.1`.
 - **Tools**: The invocation of a tool.
 - **Optimize memory**: This step summarizes and the removes all the messages except the last 2. If the last message is the response of a web
@@ -106,12 +107,24 @@ TODO: analyze YouTube videos and answer questions about objects in the video.
 ## Future work and improvements 🔜
 - **Evaluation**:  Evaluate the agent against other questions from the GAIA validation set.
 - **Large Web Extracts**: Try other chunking strategies.
-- **Audio Analysis**:Use a lesser expensive model to get the transcripts (like whisper) and if this is not enough to answer the question and more sophiticated processing is needed
+- **Audio Analysis**: Use a lesser expensive model to get the transcripts (like `whisper`) and if this is not enough to answer the question and more sophiticated processing is needed
 for other sounds like music, barks or other type of sounds then indeed use a better model.
 - **Python File execution**:Improve safety when executing python code or python files.
 - **Video Analysis**: Answer questions about objects in the video.
 - **Chessboard Images Analysis**: Detect correctly all pieces on a chess board image. 
 
+
+## How to use
+Please check the `.env.example` file for the environment variables that need to be configured. If running in a Huggingface Space, you can
+set them as secrets, otherwise rename the `.env.example` to `.env`. 
+
+The `CHESS_ENGINE_PATH` needs to be configured only if running on a Windows machine and needs to point
+to the `stockfish` executable, otherwise the `stockfish`installation is automatically detected. 
+
+The `LANGSMITH_*` properties need to be configured only if you want to enable observability with LangSmith.
+
+
+
 ## References 📚
 The math tool implementation was inspired from this repo https://github.com/langchain-ai/open_deep_research
 

app.py

@@ -1,7 +1,6 @@
 import os
 import time
 from pathlib import Path
-from typing import Optional
 
 import gradio as gr
 import pandas as pd
@@ -16,23 +15,6 @@ from utils.dependencies_checker import check_dependencies
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
 
-# --- Basic Agent Definition ---
-class BasicAgent:
-    agent: GaiaAgent
-
-    def __init__(self):
-        self.agent = GaiaAgent()
-
-    def __call__(self, question: str, attachment: Optional[Attachment] = None) -> str:
-        print(f"Agent received question (first 50 chars): {question[:50]}...")
-        if attachment:
-            print(f"Agent received an attachment : {attachment.file_path}...")
-
-        answer = self.agent.__call__(question, attachment)
-        print(f"Agent returning fixed answer: {answer}")
-        return answer
-
-
 def get_question_attached_file(task_id, file_name) -> Attachment:
     api_url = DEFAULT_API_URL
     attachment_url = f"{api_url}/files/{task_id}"
@@ -82,7 +64,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
 
     # 1. Instantiate Agent ( modify this part to create your agent)
     try:
-        agent = BasicAgent()
+        agent = GaiaAgent()
     except Exception as e:
         print(f"Error instantiating agent: {e}")
         return f"Error initializing agent: {e}", None
@@ -135,8 +117,13 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
             # check if the answer is cached, if not invoke the agent
             submitted_answer = cache.get(task_id)
             if submitted_answer is None:
+                print(f"Agent received question (first 50 chars): {question_text[:50]}...")
+                if attachment:
+                    print(f"Agent received an attachment : {attachment.file_path}...")
                 submitted_answer = agent(question_text, attachment)
-                time.sleep(20)
+                print(f"Agent returning fixed answer: {submitted_answer}")
+                #sleep in
+                time.sleep(30)
                 cache.set(task_id, submitted_answer)
             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})

nodes/nodes.py

@@ -8,7 +8,7 @@ from langchain_openai import ChatOpenAI
 from config.settings import config
 from core.messages import attachmentHandler
 from core.state import State
-from nodes.chunking_node import OversizedContentHandler
+from nodes.chunking_handler import OversizedContentHandler
 from tools.audio_tool import query_audio
 from tools.chess_tool import chess_analysis_tool
 from tools.excel_tool import query_excel_file
@@ -17,14 +17,16 @@ from tools.python_executor import execute_python_code
 from tools.tavily_tools import web_search_tools
 from utils.prompt_manager import prompt_mgmt
 
-web_search_tools.append(query_audio)
-web_search_tools.append(query_excel_file)
-web_search_tools.append(execute_python_code)
-web_search_tools.append(math_tool)
-web_search_tools.append(chess_analysis_tool)
+agent_tools = []
+agent_tools.extend(web_search_tools)
+agent_tools.append(query_audio)
+agent_tools.append(query_excel_file)
+agent_tools.append(execute_python_code)
+agent_tools.append(math_tool)
+agent_tools.append(chess_analysis_tool)
 
 model = ChatOpenAI(model=config.model_name)
-model = model.bind_tools(web_search_tools, parallel_tool_calls=False)
+model = model.bind_tools(agent_tools, parallel_tool_calls=False)
 
 response_processing_model = ChatOpenAI(model=config.response_processing_model_name)
 
@@ -115,7 +117,7 @@ def optimize_memory(state: State):
     # Delete all but the 2 most recent messages and the first one
     remaining_messages = [RemoveMessage(id=m.id) for m in state["messages"][:-2]]
 
-    # If the last message returned from a tool is oversized, chunk it and retrieve only the relevant chunks
+    # If the last message returned from a tool is oversized, chunk it and replace it with the relevant chunks
     content_handler = OversizedContentHandler()
     chunked = content_handler.process_oversized_message(state["messages"][-1], state.get("question"))
 

tools/math_agent.py

@@ -1,9 +1,3 @@
-"""Research Agent Implementation.
-
-This module implements a research agent that can perform iterative web searches
-and synthesis to answer complex research questions.
-"""
-
 from typing import Literal
 
 from langchain.chat_models import init_chat_model
@@ -30,8 +24,6 @@ compress_model = init_chat_model(model="openai:gpt-4.1",
                                  max_tokens=32000)
 
 
-# ===== AGENT NODES =====
-
 def llm_call(state: MathAgentState):
     """Analyze current state and decide on next actions.
 
@@ -100,12 +92,10 @@ def compress_research(state: MathAgentState) -> dict:
     }
 
 
-# ===== ROUTING LOGIC =====
-
 def should_continue(state: MathAgentState) -> Literal["tool_node", "compress_research"]:
-    """Determine whether to continue research or provide final answer.
+    """Determine whether to continue the solving of the problem or provide final answer.
 
-    Determines whether the agent should continue the research loop or provide
+    Determines whether the agent should continue the solving loop or provide
     a final answer based on whether the LLM made tool calls.
 
     Returns: