Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -2,12 +2,11 @@ import os
|
|
| 2 |
import gradio as gr
|
| 3 |
import requests
|
| 4 |
import pandas as pd
|
|
|
|
| 5 |
|
| 6 |
# --- Constants ---
|
| 7 |
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
|
| 8 |
|
| 9 |
-
import re
|
| 10 |
-
|
| 11 |
# ----- Robust Hardcoded Agent Definition -----
|
| 12 |
class RobustHardcodedAgent:
|
| 13 |
def __init__(self):
|
|
@@ -28,7 +27,7 @@ class RobustHardcodedAgent:
|
|
| 28 |
# Supprime retours à la ligne, espaces multiples, ponctuation et met en minuscules
|
| 29 |
text = text.lower()
|
| 30 |
text = re.sub(r'\s+', ' ', text)
|
| 31 |
-
text = re.sub(r'[^\w\s]', '', text)
|
| 32 |
return text.strip()
|
| 33 |
|
| 34 |
def __call__(self, question: str) -> str:
|
|
@@ -38,11 +37,10 @@ class RobustHardcodedAgent:
|
|
| 38 |
print(f"Agent returning answer: {answer}")
|
| 39 |
return answer
|
| 40 |
|
| 41 |
-
|
| 42 |
-
# ----- Run & Submit Function -----
|
| 43 |
def run_and_submit_all(profile: gr.OAuthProfile | None):
|
| 44 |
-
""" Fetches all questions, runs the
|
| 45 |
-
|
| 46 |
if profile:
|
| 47 |
username = profile.username
|
| 48 |
print(f"User logged in: {username}")
|
|
@@ -54,25 +52,26 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
|
|
| 54 |
questions_url = f"{api_url}/questions"
|
| 55 |
submit_url = f"{api_url}/submit"
|
| 56 |
|
| 57 |
-
# Instantiate
|
| 58 |
try:
|
| 59 |
-
agent =
|
| 60 |
except Exception as e:
|
|
|
|
| 61 |
return f"Error initializing agent: {e}", None
|
| 62 |
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
# Fetch questions
|
| 66 |
try:
|
| 67 |
response = requests.get(questions_url, timeout=15)
|
| 68 |
response.raise_for_status()
|
| 69 |
questions_data = response.json()
|
| 70 |
if not questions_data:
|
| 71 |
return "Fetched questions list is empty or invalid format.", None
|
|
|
|
| 72 |
except Exception as e:
|
|
|
|
| 73 |
return f"Error fetching questions: {e}", None
|
| 74 |
|
| 75 |
-
# Run
|
| 76 |
results_log = []
|
| 77 |
answers_payload = []
|
| 78 |
for item in questions_data:
|
|
@@ -90,8 +89,12 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
|
|
| 90 |
if not answers_payload:
|
| 91 |
return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
|
| 92 |
|
| 93 |
-
#
|
|
|
|
|
|
|
| 94 |
submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
|
|
|
|
|
|
|
| 95 |
try:
|
| 96 |
response = requests.post(submit_url, json=submission_data, timeout=60)
|
| 97 |
response.raise_for_status()
|
|
@@ -103,37 +106,29 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
|
|
| 103 |
f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
|
| 104 |
f"Message: {result_data.get('message', 'No message received.')}"
|
| 105 |
)
|
| 106 |
-
|
|
|
|
| 107 |
except Exception as e:
|
| 108 |
-
|
| 109 |
-
return
|
| 110 |
|
| 111 |
-
# -----
|
| 112 |
with gr.Blocks() as demo:
|
| 113 |
-
gr.Markdown("# Hardcoded Agent Evaluation Runner")
|
| 114 |
-
gr.Markdown(
|
| 115 |
-
|
| 116 |
-
|
| 117 |
-
|
| 118 |
-
|
| 119 |
-
"""
|
|
|
|
| 120 |
gr.LoginButton()
|
| 121 |
run_button = gr.Button("Run Evaluation & Submit All Answers")
|
| 122 |
status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
|
| 123 |
results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
|
| 124 |
-
|
| 125 |
run_button.click(fn=run_and_submit_all, outputs=[status_output, results_table])
|
| 126 |
|
|
|
|
| 127 |
if __name__ == "__main__":
|
| 128 |
-
print("\
|
| 129 |
-
space_host_startup = os.getenv("SPACE_HOST")
|
| 130 |
-
space_id_startup = os.getenv("SPACE_ID")
|
| 131 |
-
if space_host_startup:
|
| 132 |
-
print(f"✅ SPACE_HOST found: {space_host_startup}")
|
| 133 |
-
print(f" Runtime URL should be: https://{space_host_startup}.hf.space")
|
| 134 |
-
if space_id_startup:
|
| 135 |
-
print(f"✅ SPACE_ID found: {space_id_startup}")
|
| 136 |
-
print(f" Repo URL: https://huggingface.co/spaces/{space_id_startup}")
|
| 137 |
-
print("-"*(60 + len(" App Starting ")) + "\n")
|
| 138 |
-
print("Launching Gradio Interface for Hardcoded Agent Evaluation...")
|
| 139 |
demo.launch(debug=True, share=False)
|
|
|
|
| 2 |
import gradio as gr
|
| 3 |
import requests
|
| 4 |
import pandas as pd
|
| 5 |
+
import re
|
| 6 |
|
| 7 |
# --- Constants ---
|
| 8 |
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
|
| 9 |
|
|
|
|
|
|
|
| 10 |
# ----- Robust Hardcoded Agent Definition -----
|
| 11 |
class RobustHardcodedAgent:
|
| 12 |
def __init__(self):
|
|
|
|
| 27 |
# Supprime retours à la ligne, espaces multiples, ponctuation et met en minuscules
|
| 28 |
text = text.lower()
|
| 29 |
text = re.sub(r'\s+', ' ', text)
|
| 30 |
+
text = re.sub(r'[^\w\s]', '', text)
|
| 31 |
return text.strip()
|
| 32 |
|
| 33 |
def __call__(self, question: str) -> str:
|
|
|
|
| 37 |
print(f"Agent returning answer: {answer}")
|
| 38 |
return answer
|
| 39 |
|
| 40 |
+
# ----- Run and Submit All -----
|
|
|
|
| 41 |
def run_and_submit_all(profile: gr.OAuthProfile | None):
|
| 42 |
+
""" Fetches all questions, runs the RobustHardcodedAgent on them, submits answers, and returns results. """
|
| 43 |
+
|
| 44 |
if profile:
|
| 45 |
username = profile.username
|
| 46 |
print(f"User logged in: {username}")
|
|
|
|
| 52 |
questions_url = f"{api_url}/questions"
|
| 53 |
submit_url = f"{api_url}/submit"
|
| 54 |
|
| 55 |
+
# 1. Instantiate Agent
|
| 56 |
try:
|
| 57 |
+
agent = RobustHardcodedAgent()
|
| 58 |
except Exception as e:
|
| 59 |
+
print(f"Error instantiating agent: {e}")
|
| 60 |
return f"Error initializing agent: {e}", None
|
| 61 |
|
| 62 |
+
# 2. Fetch Questions
|
|
|
|
|
|
|
| 63 |
try:
|
| 64 |
response = requests.get(questions_url, timeout=15)
|
| 65 |
response.raise_for_status()
|
| 66 |
questions_data = response.json()
|
| 67 |
if not questions_data:
|
| 68 |
return "Fetched questions list is empty or invalid format.", None
|
| 69 |
+
print(f"Fetched {len(questions_data)} questions.")
|
| 70 |
except Exception as e:
|
| 71 |
+
print(f"Error fetching questions: {e}")
|
| 72 |
return f"Error fetching questions: {e}", None
|
| 73 |
|
| 74 |
+
# 3. Run Agent
|
| 75 |
results_log = []
|
| 76 |
answers_payload = []
|
| 77 |
for item in questions_data:
|
|
|
|
| 89 |
if not answers_payload:
|
| 90 |
return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
|
| 91 |
|
| 92 |
+
# 4. Prepare Submission
|
| 93 |
+
space_id = os.getenv("SPACE_ID")
|
| 94 |
+
agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
|
| 95 |
submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
|
| 96 |
+
|
| 97 |
+
# 5. Submit
|
| 98 |
try:
|
| 99 |
response = requests.post(submit_url, json=submission_data, timeout=60)
|
| 100 |
response.raise_for_status()
|
|
|
|
| 106 |
f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
|
| 107 |
f"Message: {result_data.get('message', 'No message received.')}"
|
| 108 |
)
|
| 109 |
+
results_df = pd.DataFrame(results_log)
|
| 110 |
+
return final_status, results_df
|
| 111 |
except Exception as e:
|
| 112 |
+
results_df = pd.DataFrame(results_log)
|
| 113 |
+
return f"Submission Failed: {e}", results_df
|
| 114 |
|
| 115 |
+
# ----- Gradio Interface -----
|
| 116 |
with gr.Blocks() as demo:
|
| 117 |
+
gr.Markdown("# Robust Hardcoded Agent Evaluation Runner")
|
| 118 |
+
gr.Markdown(
|
| 119 |
+
"""
|
| 120 |
+
**Instructions:**
|
| 121 |
+
1. Log in to your Hugging Face account.
|
| 122 |
+
2. Click 'Run Evaluation & Submit All Answers' to fetch questions, run the agent, submit answers, and see the score.
|
| 123 |
+
"""
|
| 124 |
+
)
|
| 125 |
gr.LoginButton()
|
| 126 |
run_button = gr.Button("Run Evaluation & Submit All Answers")
|
| 127 |
status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
|
| 128 |
results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
|
|
|
|
| 129 |
run_button.click(fn=run_and_submit_all, outputs=[status_output, results_table])
|
| 130 |
|
| 131 |
+
# ----- Main -----
|
| 132 |
if __name__ == "__main__":
|
| 133 |
+
print("\nLaunching Gradio Interface for Robust Hardcoded Agent...")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 134 |
demo.launch(debug=True, share=False)
|