LamiaYT's picture
fixing
5696ad8
raw
history blame
17.8 kB
import os
from transformers import pipeline
import gradio as gr
import requests
import inspect
import pandas as pd
from smolagents import CodeAgent
from smolagents.tools import PythonInterpreterTool
import json
import tempfile
import urllib.parse
from pathlib import Path
from duckduckgo_search import DDGS
# --- Constants ---
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
class HfApiModel:
"""
Simple wrapper for Hugging Face pipeline as a replacement for smolagents.HfApiModel
"""
def __init__(self, model_id: str, token: str = None):
self.model_id = model_id
self.token = token or os.getenv("HUGGINGFACE_INFERENCE_TOKEN")
self.pipe = pipeline("text-generation", model=model_id, token=self.token)
def __call__(self, prompt: str) -> str:
outputs = self.pipe(prompt, max_new_tokens=512, do_sample=True)
return outputs[0]["generated_text"]
class DuckDuckGoSearchTool:
name = "duckduckgo_search"
description = "Use DuckDuckGo to search the web."
def __call__(self, query: str) -> str:
try:
results = []
with DDGS() as ddgs:
for r in ddgs.text(query, max_results=3):
results.append(f"Title: {r['title']}\nURL: {r['href']}\nSnippet: {r['body']}\n---")
return "\n".join(results) if results else "No results found."
except Exception as e:
return f"Error using DuckDuckGoSearchTool: {e}"
# --- Custom Tools ---
class SerperSearchTool:
"""Enhanced search tool using Serper API for more reliable results"""
name = "serper_search"
description = "Search the web using Serper API. Use this for finding current information, facts, and data."
def __init__(self):
self.api_key = os.getenv("SERPER_API_KEY")
if not self.api_key:
print("Warning: SERPER_API_KEY not found, falling back to DuckDuckGo")
def __call__(self, query: str) -> str:
"""Search the web and return formatted results"""
if not self.api_key:
return f"Search query: {query} - API key not available"
try:
url = "https://google.serper.dev/search"
payload = json.dumps({
"q": query,
"num": 5
})
headers = {
'X-API-KEY': self.api_key,
'Content-Type': 'application/json'
}
response = requests.post(url, headers=headers, data=payload, timeout=10)
response.raise_for_status()
data = response.json()
results = []
if 'organic' in data:
for item in data['organic'][:3]:
results.append(f"Title: {item.get('title', 'N/A')}")
results.append(f"Content: {item.get('snippet', 'N/A')}")
results.append(f"URL: {item.get('link', 'N/A')}")
results.append("---")
if 'answerBox' in data:
answer = data['answerBox']
results.insert(0, f"Answer: {answer.get('answer', answer.get('snippet', 'N/A'))}")
results.insert(1, "---")
return "\n".join(results) if results else f"No results found for: {query}"
except Exception as e:
print(f"Serper search error: {e}")
return f"Search error for '{query}': {str(e)}"
class MathCalculatorTool:
"""Tool for mathematical calculations and computations"""
name = "math_calculator"
description = "Perform mathematical calculations, solve equations, and handle numerical computations."
def __call__(self, expression: str) -> str:
"""Safely evaluate mathematical expressions"""
try:
# Import math functions for calculations
import math
import operator
# Safe evaluation context
safe_dict = {
"abs": abs, "round": round, "min": min, "max": max,
"sum": sum, "pow": pow, "sqrt": math.sqrt,
"sin": math.sin, "cos": math.cos, "tan": math.tan,
"log": math.log, "log10": math.log10, "exp": math.exp,
"pi": math.pi, "e": math.e
}
# Clean the expression
expression = expression.replace("^", "**") # Handle exponents
result = eval(expression, {"__builtins__": {}}, safe_dict)
return f"Result: {result}"
except Exception as e:
return f"Math calculation error: {str(e)}"
class FileProcessorTool:
"""Tool for processing various file formats"""
name = "file_processor"
description = "Process and extract information from files (text, CSV, JSON, etc.)"
def __call__(self, file_path: str, action: str = "read") -> str:
"""Process files based on action type"""
try:
if not os.path.exists(file_path):
return f"File not found: {file_path}"
file_ext = Path(file_path).suffix.lower()
if file_ext in ['.txt', '.md']:
with open(file_path, 'r', encoding='utf-8') as f:
content = f.read()
return f"File content ({len(content)} chars):\n{content[:1000]}..."
elif file_ext == '.csv':
import pandas as pd
df = pd.read_csv(file_path)
return f"CSV file with {len(df)} rows and {len(df.columns)} columns:\n{df.head().to_string()}"
elif file_ext == '.json':
with open(file_path, 'r', encoding='utf-8') as f:
data = json.load(f)
return f"JSON data:\n{json.dumps(data, indent=2)[:1000]}..."
else:
return f"Unsupported file type: {file_ext}"
except Exception as e:
return f"File processing error: {str(e)}"
# --- Enhanced Agent Definition ---
class GAIAAgent:
def __init__(self):
"""Initialize the GAIA agent with tools and model"""
print("Initializing GAIA Agent...")
# Initialize model
try:
hf_token = os.getenv("HUGGINGFACE_INFERENCE_TOKEN")
if not hf_token:
print("Warning: HUGGINGFACE_INFERENCE_TOKEN not found")
# Use a good model for reasoning
model = HfApiModel(
model_id="meta-llama/Llama-3.1-70B-Instruct",
token=hf_token
)
# Initialize tools
self.tools = [
SerperSearchTool(),
PythonInterpreterTool(),
MathCalculatorTool(),
FileProcessorTool(),
DuckDuckGoSearchTool() # Backup search
]
# Initialize the agent
self.agent = CodeAgent(
tools=self.tools,
model=model,
max_steps=10,
verbosity_level=1
)
print("GAIA Agent initialized successfully with tools:", [tool.name for tool in self.tools])
except Exception as e:
print(f"Error initializing GAIA Agent: {e}")
# Fallback to basic setup
try:
model = HfApiModel(model_id="microsoft/DialoGPT-medium")
self.agent = CodeAgent(tools=[PythonInterpreterTool()], model=model)
print("Fallback agent initialized")
except Exception as fallback_error:
print(f"Fallback initialization failed: {fallback_error}")
self.agent = None
def __call__(self, question: str) -> str:
"""Process a question using the GAIA agent"""
print(f"Processing question: {question[:100]}...")
if not self.agent:
return "Agent initialization failed. Please check your configuration."
try:
# Enhanced prompt for better reasoning
enhanced_prompt = f"""
You are an AI assistant designed to answer questions accurately and thoroughly.
You have access to web search, Python interpreter, math calculator, and file processing tools.
Question: {question}
Please think step by step:
1. Analyze what type of question this is
2. Determine what tools or information you need
3. Use appropriate tools to gather information
4. Reason through the problem
5. Provide a clear, accurate answer
If the question requires:
- Current information or facts: Use search tools
- Calculations: Use the math calculator or Python interpreter
- File analysis: Use the file processor tool
- Multi-step reasoning: Break it down systematically
Answer:"""
# Run the agent
result = self.agent.run(enhanced_prompt)
# Extract the final answer if it's structured
if isinstance(result, dict) and 'output' in result:
answer = result['output']
else:
answer = str(result)
# Clean up the answer
if "Answer:" in answer:
answer = answer.split("Answer:")[-1].strip()
print(f"Agent response: {answer[:100]}...")
return answer
except Exception as e:
error_msg = f"Error processing question: {str(e)}"
print(error_msg)
# Fallback to basic response
try:
basic_response = f"I encountered an error while processing this question: {question}. Error: {str(e)}"
return basic_response
except:
return "Unable to process this question due to technical difficulties."
def run_and_submit_all(profile: gr.OAuthProfile | None):
"""
Fetches all questions, runs the GAIA Agent on them, submits all answers,
and displays the results.
"""
# --- Determine HF Space Runtime URL and Repo URL ---
space_id = os.getenv("SPACE_ID")
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
try:
agent = GAIAAgent()
if not agent.agent:
return "Failed to initialize GAIA Agent. Please check your tokens and try again.", None
except Exception as e:
print(f"Error instantiating agent: {e}")
return f"Error initializing agent: {e}", None
# Agent code URL
agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main" if space_id else "local"
print(f"Agent code: {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}")
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 GAIA Agent
results_log = []
answers_payload = []
print(f"Running GAIA agent on {len(questions_data)} questions...")
for i, item in enumerate(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:
print(f"Processing question {i+1}/{len(questions_data)}: {task_id}")
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[:100] + "..." if len(question_text) > 100 else question_text,
"Submitted Answer": submitted_answer[:200] + "..." if len(submitted_answer) > 200 else submitted_answer
})
except Exception as e:
print(f"Error running agent on task {task_id}: {e}")
error_answer = f"AGENT ERROR: {e}"
answers_payload.append({"task_id": task_id, "submitted_answer": error_answer})
results_log.append({
"Task ID": task_id,
"Question": question_text[:100] + "..." if len(question_text) > 100 else question_text,
"Submitted Answer": error_answer
})
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=120) # Increased timeout
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 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 ---
with gr.Blocks(title="GAIA Agent Evaluation") as demo:
gr.Markdown("# GAIA Benchmark Agent Evaluation")
gr.Markdown(
"""
**Enhanced GAIA Agent with Multiple Tools:**
- ๐Ÿ” Web Search (Serper API + DuckDuckGo fallback)
- ๐Ÿ Python Interpreter for calculations
- ๐Ÿงฎ Mathematical calculator
- ๐Ÿ“ File processor for various formats
- ๐Ÿง  Advanced reasoning with Llama-3.1-70B
**Instructions:**
1. Make sure you have SERPER_API_KEY and HUGGINGFACE_INFERENCE_TOKEN set
2. Log in to your Hugging Face account
3. Click 'Run GAIA Evaluation' to start the benchmark
**Target:** >40% accuracy on GAIA benchmark questions
"""
)
gr.LoginButton()
run_button = gr.Button("๐Ÿš€ Run GAIA Evaluation & Submit", variant="primary")
status_output = gr.Textbox(
label="Evaluation Status & Results",
lines=6,
interactive=False,
placeholder="Click the button above to start evaluation..."
)
results_table = gr.DataFrame(
label="Questions and Agent Responses",
wrap=True,
interactive=False
)
run_button.click(
fn=run_and_submit_all,
outputs=[status_output, results_table]
)
if __name__ == "__main__":
print("\n" + "="*50)
print("๐Ÿค– GAIA Agent Evaluation System Starting")
print("="*50)
# Check environment variables
serper_key = os.getenv("SERPER_API_KEY")
hf_token = os.getenv("HUGGINGFACE_INFERENCE_TOKEN")
space_id = os.getenv("SPACE_ID")
print(f"โœ… SERPER_API_KEY: {'Found' if serper_key else 'Missing (will use fallback search)'}")
print(f"โœ… HF_TOKEN: {'Found' if hf_token else 'Missing (required for model access)'}")
print(f"โœ… SPACE_ID: {space_id if space_id else 'Not found (running locally)'}")
if space_id:
print(f"๐Ÿ”— Space URL: https://huggingface.co/spaces/{space_id}")
print("="*50)
print("๐ŸŽฏ Target: >40% accuracy on GAIA benchmark")
print("๐Ÿ› ๏ธ Tools: Search, Python, Math, File Processing")
print("๐Ÿง  Model: Llama-3.1-70B-Instruct")
print("="*50 + "\n")
demo.launch(debug=True, share=False)