Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -25,16 +25,14 @@ def get_current_date_time() -> str:
|
|
| 25 |
from datetime import datetime
|
| 26 |
return datetime.now().isoformat()
|
| 27 |
|
| 28 |
-
|
| 29 |
class BasicAgent:
|
| 30 |
def __init__(self):
|
| 31 |
print("BasicAgent initialized.")
|
| 32 |
hf_token = os.getenv("HF_TOKEN")
|
| 33 |
-
# Use InferenceClientModel without specifying provider
|
| 34 |
-
# so HF router picks the best available provider automatically
|
| 35 |
model = InferenceClientModel(
|
| 36 |
model_id="meta-llama/Llama-3.3-70B-Instruct",
|
| 37 |
token=hf_token,
|
|
|
|
| 38 |
)
|
| 39 |
tools = [
|
| 40 |
DuckDuckGoSearchTool(),
|
|
@@ -50,27 +48,22 @@ class BasicAgent:
|
|
| 50 |
)
|
| 51 |
|
| 52 |
def __call__(self, question: str) -> str:
|
| 53 |
-
print(f"Agent received question (first 50 chars): {question[:50]}
|
| 54 |
try:
|
| 55 |
answer = self.agent.run(question)
|
| 56 |
-
|
|
|
|
| 57 |
except Exception as e:
|
| 58 |
print(f"Agent error: {e}")
|
| 59 |
-
|
| 60 |
-
print(f"Agent returning answer: {final_answer}")
|
| 61 |
-
return final_answer
|
| 62 |
|
| 63 |
|
| 64 |
-
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
and displays the results.
|
| 68 |
-
"""
|
| 69 |
-
# --- Determine HF Space Runtime URL and Repo URL ---
|
| 70 |
-
space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code
|
| 71 |
|
| 72 |
if profile:
|
| 73 |
-
username= f"{profile.username}"
|
| 74 |
print(f"User logged in: {username}")
|
| 75 |
else:
|
| 76 |
print("User not logged in.")
|
|
@@ -80,17 +73,15 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
|
|
| 80 |
questions_url = f"{api_url}/questions"
|
| 81 |
submit_url = f"{api_url}/submit"
|
| 82 |
|
| 83 |
-
# 1. Instantiate Agent ( modify this part to create your agent)
|
| 84 |
try:
|
| 85 |
agent = BasicAgent()
|
| 86 |
except Exception as e:
|
| 87 |
print(f"Error instantiating agent: {e}")
|
| 88 |
return f"Error initializing agent: {e}", None
|
| 89 |
-
|
| 90 |
agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
|
| 91 |
print(agent_code)
|
| 92 |
|
| 93 |
-
# 2. Fetch Questions
|
| 94 |
print(f"Fetching questions from: {questions_url}")
|
| 95 |
try:
|
| 96 |
response = requests.get(questions_url, timeout=15)
|
|
@@ -100,18 +91,10 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
|
|
| 100 |
print("Fetched questions list is empty.")
|
| 101 |
return "Fetched questions list is empty or invalid format.", None
|
| 102 |
print(f"Fetched {len(questions_data)} questions.")
|
| 103 |
-
except
|
| 104 |
print(f"Error fetching questions: {e}")
|
| 105 |
return f"Error fetching questions: {e}", None
|
| 106 |
-
except requests.exceptions.JSONDecodeError as e:
|
| 107 |
-
print(f"Error decoding JSON response from questions endpoint: {e}")
|
| 108 |
-
print(f"Response text: {response.text[:500]}")
|
| 109 |
-
return f"Error decoding server response for questions: {e}", None
|
| 110 |
-
except Exception as e:
|
| 111 |
-
print(f"An unexpected error occurred fetching questions: {e}")
|
| 112 |
-
return f"An unexpected error occurred fetching questions: {e}", None
|
| 113 |
|
| 114 |
-
# 3. Run your Agent
|
| 115 |
results_log = []
|
| 116 |
answers_payload = []
|
| 117 |
print(f"Running agent on {len(questions_data)} questions...")
|
|
@@ -133,12 +116,10 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
|
|
| 133 |
print("Agent did not produce any answers to submit.")
|
| 134 |
return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
|
| 135 |
|
| 136 |
-
# 4. Prepare Submission
|
| 137 |
submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
|
| 138 |
status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."
|
| 139 |
print(status_update)
|
| 140 |
|
| 141 |
-
# 5. Submit
|
| 142 |
print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
|
| 143 |
try:
|
| 144 |
response = requests.post(submit_url, json=submission_data, timeout=60)
|
|
@@ -147,36 +128,15 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
|
|
| 147 |
final_status = (
|
| 148 |
f"Submission Successful!\n"
|
| 149 |
f"User: {result_data.get('username')}\n"
|
| 150 |
-
f"Overall Score: {result_data.get('score', 'N/A')}
|
| 151 |
f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
|
| 152 |
f"Message: {result_data.get('message', 'No message received.')}"
|
| 153 |
)
|
| 154 |
print("Submission successful.")
|
| 155 |
results_df = pd.DataFrame(results_log)
|
| 156 |
return final_status, results_df
|
| 157 |
-
except requests.exceptions.HTTPError as e:
|
| 158 |
-
error_detail = f"Server responded with status {e.response.status_code}."
|
| 159 |
-
try:
|
| 160 |
-
error_json = e.response.json()
|
| 161 |
-
error_detail += f" Detail: {error_json.get('detail', e.response.text)}"
|
| 162 |
-
except requests.exceptions.JSONDecodeError:
|
| 163 |
-
error_detail += f" Response: {e.response.text[:500]}"
|
| 164 |
-
status_message = f"Submission Failed: {error_detail}"
|
| 165 |
-
print(status_message)
|
| 166 |
-
results_df = pd.DataFrame(results_log)
|
| 167 |
-
return status_message, results_df
|
| 168 |
-
except requests.exceptions.Timeout:
|
| 169 |
-
status_message = "Submission Failed: The request timed out."
|
| 170 |
-
print(status_message)
|
| 171 |
-
results_df = pd.DataFrame(results_log)
|
| 172 |
-
return status_message, results_df
|
| 173 |
-
except requests.exceptions.RequestException as e:
|
| 174 |
-
status_message = f"Submission Failed: Network error - {e}"
|
| 175 |
-
print(status_message)
|
| 176 |
-
results_df = pd.DataFrame(results_log)
|
| 177 |
-
return status_message, results_df
|
| 178 |
except Exception as e:
|
| 179 |
-
status_message = f"
|
| 180 |
print(status_message)
|
| 181 |
results_df = pd.DataFrame(results_log)
|
| 182 |
return status_message, results_df
|
|
@@ -188,51 +148,40 @@ with gr.Blocks() as demo:
|
|
| 188 |
gr.Markdown(
|
| 189 |
"""
|
| 190 |
**Instructions:**
|
| 191 |
-
|
| 192 |
1. Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...
|
| 193 |
2. Log in to your Hugging Face account using the button below. This uses your HF username for submission.
|
| 194 |
3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
|
| 195 |
-
|
| 196 |
---
|
| 197 |
**Disclaimers:**
|
| 198 |
Once clicking on the "submit button, it can take quite some time ( this is the time for the agent to go through all the questions).
|
| 199 |
This space provides a basic setup and is intentionally sub-optimal to encourage you to develop your own, more robust solution. For instance for the delay process of the submit button, a solution could be to cache the answers and submit in a seperate action or even to answer the questions in async.
|
| 200 |
"""
|
| 201 |
)
|
| 202 |
-
|
| 203 |
gr.LoginButton()
|
| 204 |
-
|
| 205 |
run_button = gr.Button("Run Evaluation & Submit All Answers")
|
| 206 |
-
|
| 207 |
status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
|
| 208 |
# Removed max_rows=10 from DataFrame constructor
|
| 209 |
results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
|
| 210 |
-
|
| 211 |
run_button.click(
|
| 212 |
fn=run_and_submit_all,
|
| 213 |
outputs=[status_output, results_table]
|
| 214 |
)
|
| 215 |
-
|
| 216 |
if __name__ == "__main__":
|
| 217 |
print("\n" + "-" * 30 + " App Starting " + "-" * 30)
|
| 218 |
# Check for SPACE_HOST and SPACE_ID at startup for information
|
| 219 |
space_host_startup = os.getenv("SPACE_HOST")
|
| 220 |
space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup
|
| 221 |
-
|
| 222 |
if space_host_startup:
|
| 223 |
-
print(f"
|
| 224 |
print(f" Runtime URL should be: https://{space_host_startup}.hf.space")
|
| 225 |
else:
|
| 226 |
-
print("
|
| 227 |
-
|
| 228 |
if space_id_startup: # Print repo URLs if SPACE_ID is found
|
| 229 |
-
print(f"
|
| 230 |
print(f" Repo URL: https://huggingface.co/spaces/{space_id_startup}")
|
| 231 |
print(f" Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")
|
| 232 |
else:
|
| 233 |
-
print("
|
| 234 |
-
|
| 235 |
print("-" * (60 + len(" App Starting ")) + "\n")
|
| 236 |
-
|
| 237 |
print("Launching Gradio Interface for Basic Agent Evaluation...")
|
| 238 |
-
demo.launch(debug=True, share=False)
|
|
|
|
| 25 |
from datetime import datetime
|
| 26 |
return datetime.now().isoformat()
|
| 27 |
|
|
|
|
| 28 |
class BasicAgent:
|
| 29 |
def __init__(self):
|
| 30 |
print("BasicAgent initialized.")
|
| 31 |
hf_token = os.getenv("HF_TOKEN")
|
|
|
|
|
|
|
| 32 |
model = InferenceClientModel(
|
| 33 |
model_id="meta-llama/Llama-3.3-70B-Instruct",
|
| 34 |
token=hf_token,
|
| 35 |
+
provider="cerebras",
|
| 36 |
)
|
| 37 |
tools = [
|
| 38 |
DuckDuckGoSearchTool(),
|
|
|
|
| 48 |
)
|
| 49 |
|
| 50 |
def __call__(self, question: str) -> str:
|
| 51 |
+
print(f"Agent received question (first 50 chars): {question[:50]}")
|
| 52 |
try:
|
| 53 |
answer = self.agent.run(question)
|
| 54 |
+
print(f"Agent answer: {str(answer)[:100]}")
|
| 55 |
+
return str(answer)
|
| 56 |
except Exception as e:
|
| 57 |
print(f"Agent error: {e}")
|
| 58 |
+
return f"Error: {e}"
|
|
|
|
|
|
|
| 59 |
|
| 60 |
|
| 61 |
+
# --- run_and_submit_all function (Keep as is) ---
|
| 62 |
+
def run_and_submit_all(profile: gr.OAuthProfile | None):
|
| 63 |
+
space_id = os.getenv("SPACE_ID")
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
|
| 65 |
if profile:
|
| 66 |
+
username = f"{profile.username}"
|
| 67 |
print(f"User logged in: {username}")
|
| 68 |
else:
|
| 69 |
print("User not logged in.")
|
|
|
|
| 73 |
questions_url = f"{api_url}/questions"
|
| 74 |
submit_url = f"{api_url}/submit"
|
| 75 |
|
|
|
|
| 76 |
try:
|
| 77 |
agent = BasicAgent()
|
| 78 |
except Exception as e:
|
| 79 |
print(f"Error instantiating agent: {e}")
|
| 80 |
return f"Error initializing agent: {e}", None
|
| 81 |
+
|
| 82 |
agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
|
| 83 |
print(agent_code)
|
| 84 |
|
|
|
|
| 85 |
print(f"Fetching questions from: {questions_url}")
|
| 86 |
try:
|
| 87 |
response = requests.get(questions_url, timeout=15)
|
|
|
|
| 91 |
print("Fetched questions list is empty.")
|
| 92 |
return "Fetched questions list is empty or invalid format.", None
|
| 93 |
print(f"Fetched {len(questions_data)} questions.")
|
| 94 |
+
except Exception as e:
|
| 95 |
print(f"Error fetching questions: {e}")
|
| 96 |
return f"Error fetching questions: {e}", None
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 97 |
|
|
|
|
| 98 |
results_log = []
|
| 99 |
answers_payload = []
|
| 100 |
print(f"Running agent on {len(questions_data)} questions...")
|
|
|
|
| 116 |
print("Agent did not produce any answers to submit.")
|
| 117 |
return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
|
| 118 |
|
|
|
|
| 119 |
submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
|
| 120 |
status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."
|
| 121 |
print(status_update)
|
| 122 |
|
|
|
|
| 123 |
print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
|
| 124 |
try:
|
| 125 |
response = requests.post(submit_url, json=submission_data, timeout=60)
|
|
|
|
| 128 |
final_status = (
|
| 129 |
f"Submission Successful!\n"
|
| 130 |
f"User: {result_data.get('username')}\n"
|
| 131 |
+
f"Overall Score: {result_data.get('score', 'N/A')} "
|
| 132 |
f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
|
| 133 |
f"Message: {result_data.get('message', 'No message received.')}"
|
| 134 |
)
|
| 135 |
print("Submission successful.")
|
| 136 |
results_df = pd.DataFrame(results_log)
|
| 137 |
return final_status, results_df
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 138 |
except Exception as e:
|
| 139 |
+
status_message = f"Submission Failed: {e}"
|
| 140 |
print(status_message)
|
| 141 |
results_df = pd.DataFrame(results_log)
|
| 142 |
return status_message, results_df
|
|
|
|
| 148 |
gr.Markdown(
|
| 149 |
"""
|
| 150 |
**Instructions:**
|
|
|
|
| 151 |
1. Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...
|
| 152 |
2. Log in to your Hugging Face account using the button below. This uses your HF username for submission.
|
| 153 |
3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
|
|
|
|
| 154 |
---
|
| 155 |
**Disclaimers:**
|
| 156 |
Once clicking on the "submit button, it can take quite some time ( this is the time for the agent to go through all the questions).
|
| 157 |
This space provides a basic setup and is intentionally sub-optimal to encourage you to develop your own, more robust solution. For instance for the delay process of the submit button, a solution could be to cache the answers and submit in a seperate action or even to answer the questions in async.
|
| 158 |
"""
|
| 159 |
)
|
|
|
|
| 160 |
gr.LoginButton()
|
|
|
|
| 161 |
run_button = gr.Button("Run Evaluation & Submit All Answers")
|
|
|
|
| 162 |
status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
|
| 163 |
# Removed max_rows=10 from DataFrame constructor
|
| 164 |
results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
|
|
|
|
| 165 |
run_button.click(
|
| 166 |
fn=run_and_submit_all,
|
| 167 |
outputs=[status_output, results_table]
|
| 168 |
)
|
|
|
|
| 169 |
if __name__ == "__main__":
|
| 170 |
print("\n" + "-" * 30 + " App Starting " + "-" * 30)
|
| 171 |
# Check for SPACE_HOST and SPACE_ID at startup for information
|
| 172 |
space_host_startup = os.getenv("SPACE_HOST")
|
| 173 |
space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup
|
|
|
|
| 174 |
if space_host_startup:
|
| 175 |
+
print(f"\u2705 SPACE_HOST found: {space_host_startup}")
|
| 176 |
print(f" Runtime URL should be: https://{space_host_startup}.hf.space")
|
| 177 |
else:
|
| 178 |
+
print("\u2139\ufe0f SPACE_HOST environment variable not found (running locally?).")
|
|
|
|
| 179 |
if space_id_startup: # Print repo URLs if SPACE_ID is found
|
| 180 |
+
print(f"\u2705 SPACE_ID found: {space_id_startup}")
|
| 181 |
print(f" Repo URL: https://huggingface.co/spaces/{space_id_startup}")
|
| 182 |
print(f" Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")
|
| 183 |
else:
|
| 184 |
+
print("\u2139\ufe0f SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")
|
|
|
|
| 185 |
print("-" * (60 + len(" App Starting ")) + "\n")
|
|
|
|
| 186 |
print("Launching Gradio Interface for Basic Agent Evaluation...")
|
| 187 |
+
demo.launch(debug=True, share=False)
|