Spaces:
Build error
Build error
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,200 +1,6 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
import re
|
| 8 |
-
|
| 9 |
-
# Constants
|
| 10 |
-
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
|
| 11 |
-
|
| 12 |
-
# --- Generalized Tools ---
|
| 13 |
-
class KnowledgeExtractionTool:
|
| 14 |
-
def __init__(self):
|
| 15 |
-
self.api_url = DEFAULT_API_URL
|
| 16 |
-
|
| 17 |
-
def fetch_data(self, url):
|
| 18 |
-
"""Fetches data from a URL (Wikipedia or other sources)"""
|
| 19 |
-
response = requests.get(url)
|
| 20 |
-
if response.status_code == 200:
|
| 21 |
-
return response.text
|
| 22 |
-
return None
|
| 23 |
-
|
| 24 |
-
def extract_winners_from_wikipedia(self, url):
|
| 25 |
-
"""Extracts competition winners from a Wikipedia page"""
|
| 26 |
-
page_content = self.fetch_data(url)
|
| 27 |
-
if not page_content:
|
| 28 |
-
return []
|
| 29 |
-
|
| 30 |
-
soup = BeautifulSoup(page_content, 'html.parser')
|
| 31 |
-
winners_data = []
|
| 32 |
-
|
| 33 |
-
# Look for the table of winners
|
| 34 |
-
winners_table = soup.find('table', {'class': 'wikitable'})
|
| 35 |
-
if winners_table:
|
| 36 |
-
for row in winners_table.find_all('tr')[1:]: # Skipping the header row
|
| 37 |
-
cells = row.find_all('td')
|
| 38 |
-
if len(cells) > 3:
|
| 39 |
-
name = cells[0].text.strip() # Winner's name
|
| 40 |
-
year = int(cells[1].text.strip()) # Year
|
| 41 |
-
nationality = cells[3].text.strip() # Nationality
|
| 42 |
-
winners_data.append({"name": name, "year": year, "nationality": nationality})
|
| 43 |
-
return winners_data
|
| 44 |
-
|
| 45 |
-
def filter_winners(self, winners_data, year_threshold=1977, countries_to_check=None):
|
| 46 |
-
"""Filters winners by year and nationality (if country no longer exists)"""
|
| 47 |
-
if not countries_to_check:
|
| 48 |
-
countries_to_check = ["Yugoslavia", "Soviet Union", "East Germany"] # Example list
|
| 49 |
-
|
| 50 |
-
filtered_winners = []
|
| 51 |
-
for winner in winners_data:
|
| 52 |
-
if winner["year"] > year_threshold and any(country in winner["nationality"] for country in countries_to_check):
|
| 53 |
-
filtered_winners.append(winner)
|
| 54 |
-
return filtered_winners
|
| 55 |
-
|
| 56 |
-
def extract_first_name(self, full_name):
|
| 57 |
-
"""Extract the first name from a full name string"""
|
| 58 |
-
return full_name.split()[0] if full_name else None
|
| 59 |
-
|
| 60 |
-
def process_question(self, question):
|
| 61 |
-
"""Generalized function to process a wide variety of questions"""
|
| 62 |
-
if "Malko Competition" in question and "first name" in question:
|
| 63 |
-
print("Processing Malko Competition query...")
|
| 64 |
-
|
| 65 |
-
# Fetch data from Wikipedia (or other sources)
|
| 66 |
-
winners_data = self.extract_winners_from_wikipedia("https://en.wikipedia.org/wiki/Malko_Competition")
|
| 67 |
-
|
| 68 |
-
# Filter by year and nationality of winners
|
| 69 |
-
filtered_winners = self.filter_winners(winners_data)
|
| 70 |
-
|
| 71 |
-
# Extract and return the first name
|
| 72 |
-
if filtered_winners:
|
| 73 |
-
first_name = self.extract_first_name(filtered_winners[0]["name"])
|
| 74 |
-
return f"The first name of the winner is: {first_name}"
|
| 75 |
-
else:
|
| 76 |
-
return "No winners found with the specified conditions."
|
| 77 |
-
|
| 78 |
-
# Add other types of questions and their handling here as needed
|
| 79 |
-
return "Question could not be processed."
|
| 80 |
-
|
| 81 |
-
# --- Basic Agent Definition ---
|
| 82 |
-
class BasicAgent:
|
| 83 |
-
def __init__(self):
|
| 84 |
-
self.knowledge_tool = KnowledgeExtractionTool()
|
| 85 |
-
|
| 86 |
-
def __call__(self, question: str) -> str:
|
| 87 |
-
print(f"Processing question: {question}")
|
| 88 |
-
return self.knowledge_tool.process_question(question)
|
| 89 |
-
|
| 90 |
-
# --- Run and Submit All Function ---
|
| 91 |
-
def run_and_submit_all(profile: gr.OAuthProfile | None):
|
| 92 |
-
"""
|
| 93 |
-
Fetches all questions, runs the BasicAgent on them, submits all answers,
|
| 94 |
-
and displays the results.
|
| 95 |
-
"""
|
| 96 |
-
space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code
|
| 97 |
-
|
| 98 |
-
if profile:
|
| 99 |
-
username = f"{profile.username}"
|
| 100 |
-
print(f"User logged in: {username}")
|
| 101 |
-
else:
|
| 102 |
-
print("User not logged in.")
|
| 103 |
-
return "Please Login to Hugging Face with the button.", None
|
| 104 |
-
|
| 105 |
-
api_url = DEFAULT_API_URL
|
| 106 |
-
questions_url = f"{api_url}/questions"
|
| 107 |
-
submit_url = f"{api_url}/submit"
|
| 108 |
-
|
| 109 |
-
# Instantiate Agent
|
| 110 |
-
try:
|
| 111 |
-
agent = BasicAgent()
|
| 112 |
-
except Exception as e:
|
| 113 |
-
print(f"Error instantiating agent: {e}")
|
| 114 |
-
return f"Error initializing agent: {e}", None
|
| 115 |
-
|
| 116 |
-
# Fetch Questions
|
| 117 |
-
print(f"Fetching questions from: {questions_url}")
|
| 118 |
-
try:
|
| 119 |
-
response = requests.get(questions_url, timeout=15)
|
| 120 |
-
response.raise_for_status()
|
| 121 |
-
questions_data = response.json()
|
| 122 |
-
if not questions_data:
|
| 123 |
-
print("Fetched questions list is empty.")
|
| 124 |
-
return "Fetched questions list is empty or invalid format.", None
|
| 125 |
-
print(f"Fetched {len(questions_data)} questions.")
|
| 126 |
-
except requests.exceptions.RequestException as e:
|
| 127 |
-
print(f"Error fetching questions: {e}")
|
| 128 |
-
return f"Error fetching questions: {e}", None
|
| 129 |
-
|
| 130 |
-
# Run agent on each question
|
| 131 |
-
results_log = []
|
| 132 |
-
answers_payload = []
|
| 133 |
-
print(f"Running agent on {len(questions_data)} questions...")
|
| 134 |
-
for item in questions_data:
|
| 135 |
-
task_id = item.get("task_id")
|
| 136 |
-
question_text = item.get("question")
|
| 137 |
-
if not task_id or question_text is None:
|
| 138 |
-
print(f"Skipping item with missing task_id or question: {item}")
|
| 139 |
-
continue
|
| 140 |
-
try:
|
| 141 |
-
submitted_answer = agent(question_text)
|
| 142 |
-
answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
|
| 143 |
-
results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
|
| 144 |
-
except Exception as e:
|
| 145 |
-
print(f"Error running agent on task {task_id}: {e}")
|
| 146 |
-
results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})
|
| 147 |
-
|
| 148 |
-
if not answers_payload:
|
| 149 |
-
print("Agent did not produce any answers to submit.")
|
| 150 |
-
return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
|
| 151 |
-
|
| 152 |
-
# Prepare and submit answers
|
| 153 |
-
submission_data = {"username": username.strip(), "agent_code": f"https://huggingface.co/spaces/{os.getenv('SPACE_ID')}/tree/main", "answers": answers_payload}
|
| 154 |
-
print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
|
| 155 |
-
try:
|
| 156 |
-
response = requests.post(submit_url, json=submission_data, timeout=60)
|
| 157 |
-
response.raise_for_status()
|
| 158 |
-
result_data = response.json()
|
| 159 |
-
final_status = (
|
| 160 |
-
f"Submission Successful!\n"
|
| 161 |
-
f"User: {result_data.get('username')}\n"
|
| 162 |
-
f"Overall Score: {result_data.get('score', 'N/A')}% "
|
| 163 |
-
f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
|
| 164 |
-
f"Message: {result_data.get('message', 'No message received.')}"
|
| 165 |
-
)
|
| 166 |
-
print("Submission successful.")
|
| 167 |
-
results_df = pd.DataFrame(results_log)
|
| 168 |
-
return final_status, results_df
|
| 169 |
-
except requests.exceptions.RequestException as e:
|
| 170 |
-
status_message = f"Submission Failed: Network error - {e}"
|
| 171 |
-
print(status_message)
|
| 172 |
-
results_df = pd.DataFrame(results_log)
|
| 173 |
-
return status_message, results_df
|
| 174 |
-
|
| 175 |
-
# --- Build Gradio Interface using Blocks ---
|
| 176 |
-
with gr.Blocks() as demo:
|
| 177 |
-
gr.Markdown("# Generalized Agent Evaluation Runner")
|
| 178 |
-
gr.Markdown(
|
| 179 |
-
"""
|
| 180 |
-
**Instructions:**
|
| 181 |
-
|
| 182 |
-
1. Clone this space and modify the code to define your agent's logic, tools, and packages.
|
| 183 |
-
2. Log in to your Hugging Face account using the button below.
|
| 184 |
-
3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
|
| 185 |
-
"""
|
| 186 |
-
)
|
| 187 |
-
|
| 188 |
-
gr.LoginButton()
|
| 189 |
-
|
| 190 |
-
run_button = gr.Button("Run Evaluation & Submit All Answers")
|
| 191 |
-
status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
|
| 192 |
-
results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
|
| 193 |
-
|
| 194 |
-
run_button.click(
|
| 195 |
-
fn=run_and_submit_all,
|
| 196 |
-
outputs=[status_output, results_table]
|
| 197 |
-
)
|
| 198 |
-
|
| 199 |
-
if __name__ == "__main__":
|
| 200 |
-
demo.launch(debug=True, share=False)
|
|
|
|
| 1 |
+
login_btn = gr.LoginButton()
|
| 2 |
+
run_btn.click(
|
| 3 |
+
fn=run_and_submit_all,
|
| 4 |
+
inputs=[login_btn], # pass the profile from LoginButton
|
| 5 |
+
outputs=[status_out, results_tbl]
|
| 6 |
+
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|