File size: 6,908 Bytes
10e9b7d
 
4c934c3
 
1968b68
43ab812
91f5922
012ef3f
30b3077
4c934c3
30b3077
9925072
012ef3f
0dd84e4
 
 
 
 
 
 
 
 
91f5922
 
012ef3f
 
 
 
 
 
 
30b3077
012ef3f
 
91f5922
 
 
 
012ef3f
0dd84e4
 
 
 
 
 
 
 
012ef3f
30b3077
012ef3f
 
30b3077
012ef3f
0c36fa7
30b3077
012ef3f
0dd84e4
 
 
 
 
012ef3f
94d642e
30b3077
012ef3f
0dd84e4
 
 
 
 
 
 
 
 
 
012ef3f
 
 
 
 
 
 
30b3077
012ef3f
 
 
 
 
 
0dd84e4
012ef3f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
94d642e
 
012ef3f
 
30b3077
16da5cd
012ef3f
 
 
 
16da5cd
012ef3f
30b3077
012ef3f
326bc46
 
012ef3f
 
 
 
94d642e
012ef3f
 
 
94d642e
012ef3f
91f5922
 
012ef3f
94d642e
012ef3f
0c36fa7
012ef3f
 
 
 
94d642e
0c36fa7
 
94d642e
30b3077
012ef3f
 
 
 
 
 
 
 
 
 
 
 
94d642e
012ef3f
 
 
 
 
 
 
 
94d642e
012ef3f
 
 
 
 
 
d7c91b6
94d642e
012ef3f
d7c91b6
012ef3f
 
 
 
30b3077
012ef3f
 
30b3077
012ef3f
 
30b3077
012ef3f
 
 
 
30b3077
94d642e
 
012ef3f
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
import os
import gradio as gr
import requests
import pandas as pd
from smolagents import ToolCallingAgent, tool
from duckduckgo_search import DDGS
import math
from datetime import datetime
import re

# --- Enhanced Tools ---
@tool
def enhanced_search(query: str, num_results: int = 3) -> str:
    """Performs web search with result filtering and quality checks.
    
    Args:
        query: The search query string to look up.
        num_results: Number of results to return (default 3).
    
    Returns:
        A formatted string containing the search results or error message.
    """
    try:
        with DDGS() as ddgs:
            results = ddgs.text(query, max_results=num_results)
            filtered = [
                f"## {r['title']}\n{r['body']}\nURL: {r['href']}"
                for r in results
                if len(r['body']) > 30 and not any(
                    kw in r['title'].lower() 
                    for kw in ['advertisement', 'sponsored', 'ad', 'buy']
                )
            ]
            return "\n\n".join(filtered) if filtered else "No quality results found."
    except Exception as e:
        return f"Search error: {e}"

@tool
def scientific_calculator(expression: str) -> str:
    """Evaluates mathematical expressions with scientific functions.
    
    Args:
        expression: The mathematical expression to evaluate.
    
    Returns:
        The result as a string or error message.
    """
    allowed_names = {k: v for k, v in math.__dict__.items() if not k.startswith("__")}
    try:
        result = eval(expression, {"__builtins__": {}}, allowed_names)
        return str(round(result, 6)) if isinstance(result, float) else str(result)
    except Exception as e:
        return f"Calculation error: {e}"

@tool
def get_current_date() -> str:
    """Gets the current date and time.
    
    Returns:
        Current datetime in YYYY-MM-DD HH:MM:SS format.
    """
    return datetime.now().strftime("%Y-%m-%d %H:%M:%S")

@tool
def unit_converter(amount: float, from_unit: str, to_unit: str) -> str:
    """Converts between common measurement units.
    
    Args:
        amount: The numerical value to convert.
        from_unit: The source unit (e.g., 'miles').
        to_unit: The target unit (e.g., 'kilometers').
    
    Returns:
        The converted value with unit or error message.
    """
    conversions = {
        ('miles', 'kilometers'): lambda x: x * 1.60934,
        ('pounds', 'kilograms'): lambda x: x * 0.453592,
        ('fahrenheit', 'celsius'): lambda x: (x - 32) * 5/9,
    }
    key = (from_unit.lower(), to_unit.lower())
    if key in conversions:
        try:
            result = conversions[key](float(amount))
            return f"{round(result, 4)} {to_unit}"
        except:
            return "Invalid amount"
    return f"Unsupported conversion: {from_unit}{to_unit}"


# --- Agent Core ---
class GAIAAgent:
    def __init__(self):
        self.agent = ToolCallingAgent(
            name="GAIA-HF-Agent",
            description="Specialized agent for GAIA tasks",
            tools=[enhanced_search, scientific_calculator, get_current_date, unit_converter],
            model="gpt-4-turbo",  # or "gpt-3.5-turbo" if unavailable
            planning_interval=5,
            max_iterations=10
        )
        self.session_history = []

    def preprocess_question(self, question: str) -> str:
        """Clean GAIA questions"""
        question = re.sub(r'\[\d+\]', '', question)  # Remove citations
        question = question.replace("(a)", "").replace("(b)", "")  # Remove options
        return question.strip()

    def postprocess_answer(self, answer: str) -> str:
        """Extract most precise answer"""
        # Extract numbers/dates from longer answers
        numbers = re.findall(r'\d+\.?\d*', answer)
        dates = re.findall(r'\d{4}-\d{2}-\d{2}', answer)
        if dates:
            return dates[-1]
        if numbers:
            return numbers[-1]
        return answer[:500]  # Limit length

    def __call__(self, question: str) -> str:
        clean_q = self.preprocess_question(question)
        print(f"Processing: {clean_q}")
        
        try:
            answer = self.agent.run(clean_q)
            processed = self.postprocess_answer(answer)
            self.session_history.append((question, processed))
            return processed
        except Exception as e:
            return f"Agent error: {str(e)}"

# --- HF Space Integration ---
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"

def run_and_submit(profile: gr.OAuthProfile | None):
    if not profile:
        return "Please log in to submit", None
    
    space_id = os.getenv("SPACE_ID")
    agent = GAIAAgent()
    
    # Fetch questions
    try:
        response = requests.get(f"{DEFAULT_API_URL}/questions", timeout=20)
        questions = response.json()
        if not questions:
            return "No questions received", None
    except Exception as e:
        return f"Failed to get questions: {e}", None

    # Process questions
    results = []
    answers = []
    for item in questions[:20]:  # Limit to 20 for testing
        task_id = item.get("task_id")
        question = item.get("question")
        if not task_id or not question:
            continue
            
        answer = agent(question)
        results.append({
            "Task ID": task_id,
            "Question": question,
            "Answer": answer
        })
        answers.append({
            "task_id": task_id,
            "submitted_answer": answer
        })

    # Submit answers
    try:
        response = requests.post(
            f"{DEFAULT_API_URL}/submit",
            json={
                "username": profile.username,
                "agent_code": f"https://huggingface.co/spaces/{space_id}",
                "answers": answers
            },
            timeout=60
        )
        data = response.json()
        return (
            f"✅ Submitted {len(answers)} answers\n"
            f"Score: {data.get('score', 'N/A')}%\n"
            f"Correct: {data.get('correct_count', '?')}/{data.get('total_attempted', '?')}\n"
            f"Message: {data.get('message', '')}",
            pd.DataFrame(results))
    except Exception as e:
        return f"Submission failed: {e}", pd.DataFrame(results)
        
# --- Gradio UI ---
with gr.Blocks(title="GAIA Agent") as demo:
    gr.Markdown("## 🚀 GAIA Task Agent")
    gr.Markdown("Login and click submit to run evaluation")
    
    login = gr.LoginButton()
    submit_btn = gr.Button("Run & Submit Answers", variant="primary")
    
    status = gr.Textbox(label="Submission Status", interactive=False)
    results = gr.DataFrame(label="Processed Answers")
    
    submit_btn.click(
        fn=run_and_submit,
        inputs=None,
        outputs=[status, results]
    )

if __name__ == "__main__":
    demo.launch(debug=True)