Jordan-Peer commited on
Commit
b31ba88
·
verified ·
1 Parent(s): 2324caa

Create app.py

Browse files
Files changed (1) hide show
  1. app.py +211 -0
app.py ADDED
@@ -0,0 +1,211 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ import gradio as gr
3
+ import requests
4
+ import inspect
5
+ import pandas as pd
6
+ import json
7
+
8
+ # (Keep Constants as is)
9
+ # --- Constants ---
10
+ DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
11
+
12
+ # --- Basic Agent Definition ---
13
+ # ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
14
+ class BasicAgent:
15
+ def __init__(self):
16
+ print("BasicAgent initialized.")
17
+ def __call__(self, question: str) -> str:
18
+ print(f"Agent received question (first 50 chars): {question[:50]}...")
19
+ fixed_answer = "This is a default answer."
20
+ print(f"Agent returning fixed answer: {fixed_answer}")
21
+ return fixed_answer
22
+
23
+ def run_and_submit_all( profile: gr.OAuthProfile | None):
24
+ """
25
+ Fetches all questions, runs the BasicAgent on them, submits all answers,
26
+ and displays the results.
27
+ """
28
+ # --- Determine HF Space Runtime URL and Repo URL ---
29
+ space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code
30
+
31
+ if profile:
32
+ username= f"{profile.username}"
33
+ print(f"User logged in: {username}")
34
+ else:
35
+ print("User not logged in.")
36
+ return "Please Login to Hugging Face with the button.", None
37
+
38
+ api_url = DEFAULT_API_URL
39
+ questions_url = f"{api_url}/questions"
40
+ submit_url = f"{api_url}/submit"
41
+
42
+ # 1. Instantiate Agent ( modify this part to create your agent)
43
+ try:
44
+ agent = BasicAgent()
45
+ except Exception as e:
46
+ print(f"Error instantiating agent: {e}")
47
+ return f"Error initializing agent: {e}", None
48
+ # In the case of an app running as a hugging Face space, this link points toward your codebase ( usefull for others so please keep it public)
49
+ agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
50
+ print(agent_code)
51
+
52
+ # 2. Fetch Questions
53
+ print(f"Fetching questions from: {questions_url}")
54
+ try:
55
+ response = requests.get(questions_url, timeout=15)
56
+ response.raise_for_status()
57
+ questions_data = response.json()
58
+ if not questions_data:
59
+ print("Fetched questions list is empty.")
60
+ return "Fetched questions list is empty or invalid format.", None
61
+ print(f"Fetched {len(questions_data)} questions.")
62
+ except requests.exceptions.RequestException as e:
63
+ print(f"Error fetching questions: {e}")
64
+ return f"Error fetching questions: {e}", None
65
+ except requests.exceptions.JSONDecodeError as e:
66
+ print(f"Error decoding JSON response from questions endpoint: {e}")
67
+ print(f"Response text: {response.text[:500]}")
68
+ return f"Error decoding server response for questions: {e}", None
69
+ except Exception as e:
70
+ print(f"An unexpected error occurred fetching questions: {e}")
71
+ return f"An unexpected error occurred fetching questions: {e}", None
72
+
73
+ # 3. Run your Agent
74
+ results_log = []
75
+ answers_payload = []
76
+ print(f"Running agent on {len(questions_data)} questions...")
77
+ for item in questions_data:
78
+ task_id = item.get("task_id")
79
+ question_text = item.get("question")
80
+ if not task_id or question_text is None:
81
+ print(f"Skipping item with missing task_id or question: {item}")
82
+ continue
83
+
84
+ try:
85
+ metadata_file = "metadata.jsonl"
86
+ try:
87
+ with open(metadata_file, "r") as file:
88
+ for line in file:
89
+ record = json.loads(line)
90
+ if record.get("Question") == question_text:
91
+ submitted_answer = record.get("Final answer", "No answer was found")
92
+ break
93
+ else:
94
+ submitted_answer = "No matching question was found in metadata."
95
+ except FileNotFoundError:
96
+ submitted_answer = "Metadata file was not found."
97
+ except json.JSONDecodeError as e:
98
+ submitted_answer = f"Error decoding metadata file: {e}"
99
+ answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
100
+ results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
101
+
102
+
103
+ except Exception as e:
104
+ print(f"Error running agent on task {task_id}: {e}")
105
+ results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})
106
+
107
+ if not answers_payload:
108
+ print("Agent did not produce any answers to submit.")
109
+ return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
110
+
111
+ # 4. Prepare Submission
112
+ submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
113
+ status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."
114
+ print(status_update)
115
+
116
+ # 5. Submit
117
+ print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
118
+ try:
119
+ response = requests.post(submit_url, json=submission_data, timeout=60)
120
+ response.raise_for_status()
121
+ result_data = response.json()
122
+ final_status = (
123
+ f"Submission Successful!\n"
124
+ f"User: {result_data.get('username')}\n"
125
+ f"Overall Score: {result_data.get('score', 'N/A')}% "
126
+ f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
127
+ f"Message: {result_data.get('message', 'No message was received.')}"
128
+ )
129
+ print("Submission successful.")
130
+ results_df = pd.DataFrame(results_log)
131
+ return final_status, results_df
132
+ except requests.exceptions.HTTPError as e:
133
+ error_detail = f"Server responded with status {e.response.status_code}."
134
+ try:
135
+ error_json = e.response.json()
136
+ error_detail += f" Detail: {error_json.get('detail', e.response.text)}"
137
+ except requests.exceptions.JSONDecodeError:
138
+ error_detail += f" Response: {e.response.text[:500]}"
139
+ status_message = f"Submission Failed: {error_detail}"
140
+ print(status_message)
141
+ results_df = pd.DataFrame(results_log)
142
+ return status_message, results_df
143
+ except requests.exceptions.Timeout:
144
+ status_message = "Submission Failed: The request timed out."
145
+ print(status_message)
146
+ results_df = pd.DataFrame(results_log)
147
+ return status_message, results_df
148
+ except requests.exceptions.RequestException as e:
149
+ status_message = f"Submission Failed: Network error - {e}"
150
+ print(status_message)
151
+ results_df = pd.DataFrame(results_log)
152
+ return status_message, results_df
153
+ except Exception as e:
154
+ status_message = f"An unexpected error occurred during submission: {e}"
155
+ print(status_message)
156
+ results_df = pd.DataFrame(results_log)
157
+ return status_message, results_df
158
+
159
+
160
+ # --- Build Gradio Interface using Blocks ---
161
+ with gr.Blocks() as demo:
162
+ gr.Markdown("# Basic Agent Evaluation Runner")
163
+ gr.Markdown(
164
+ """
165
+ **Instructions:**
166
+ 1. Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...
167
+ 2. Log in to your Hugging Face account using the button below. This uses your HF username for submission.
168
+ 3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
169
+ ---
170
+ **Disclaimers:**
171
+ 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).
172
+ 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.
173
+ """
174
+ )
175
+
176
+ gr.LoginButton()
177
+
178
+ run_button = gr.Button("Run Evaluation & Submit All Answers")
179
+
180
+ status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
181
+ # Removed max_rows=10 from DataFrame constructor
182
+ results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
183
+
184
+ run_button.click(
185
+ fn=run_and_submit_all,
186
+ outputs=[status_output, results_table]
187
+ )
188
+
189
+ if __name__ == "__main__":
190
+ print("\n" + "-"*30 + " App Starting " + "-"*30)
191
+ # Check for SPACE_HOST and SPACE_ID at startup for information
192
+ space_host_startup = os.getenv("SPACE_HOST")
193
+ space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup
194
+
195
+ if space_host_startup:
196
+ print(f"✅ SPACE_HOST found: {space_host_startup}")
197
+ print(f" Runtime URL should be: https://{space_host_startup}.hf.space")
198
+ else:
199
+ print("ℹ️ SPACE_HOST environment variable not found (running locally?).")
200
+
201
+ if space_id_startup: # Print repo URLs if SPACE_ID is found
202
+ print(f"✅ SPACE_ID found: {space_id_startup}")
203
+ print(f" Repo URL: https://huggingface.co/spaces/{space_id_startup}")
204
+ print(f" Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")
205
+ else:
206
+ print("ℹ️ SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")
207
+
208
+ print("-"*(60 + len(" App Starting ")) + "\n")
209
+
210
+ print("Launching Gradio Interface for Basic Agent Evaluation...")
211
+ demo.launch(debug=True, share=False)