Spaces:
Sleeping
Sleeping
Commit ·
efa9970
1
Parent(s): e21cbb5
IP limit testing (set 1 per day)
Browse files
app.py
CHANGED
|
@@ -4,12 +4,22 @@ import time
|
|
| 4 |
import html
|
| 5 |
import re
|
| 6 |
import traceback
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
|
| 8 |
"""
|
| 9 |
For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
|
| 10 |
"""
|
| 11 |
client = InferenceClient("Trinoid/Data_Management")
|
| 12 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
# Comprehensive patterns to filter out thinking and meta-commentary
|
| 14 |
THINKING_PATTERNS = [
|
| 15 |
r"Okay, so I('m| am) (trying to|going to|attempting to)",
|
|
@@ -32,6 +42,39 @@ THINKING_PATTERNS = [
|
|
| 32 |
r"I'll approach this by",
|
| 33 |
]
|
| 34 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
def process_final_response(response_text):
|
| 36 |
"""Comprehensive processing of the final response to ensure quality"""
|
| 37 |
|
|
@@ -107,6 +150,23 @@ def respond(
|
|
| 107 |
temperature,
|
| 108 |
top_p,
|
| 109 |
):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 110 |
# Create a more effective system prompt
|
| 111 |
enhanced_system_message = f"""You are an expert in Microsoft 365 services including SharePoint, OneDrive, Teams, and the Microsoft 365 compliance ecosystem.
|
| 112 |
|
|
@@ -202,6 +262,30 @@ If comparing services:
|
|
| 202 |
error_msg = f"I apologize, but I encountered an error while generating a response. Error details: {str(e)}"
|
| 203 |
yield error_msg
|
| 204 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 205 |
|
| 206 |
# Custom CSS for Plant Wisdom.AI styling
|
| 207 |
custom_css = """
|
|
@@ -309,12 +393,25 @@ summary:hover {
|
|
| 309 |
.thinking-step:last-child {
|
| 310 |
border-bottom: none;
|
| 311 |
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 312 |
"""
|
| 313 |
|
| 314 |
"""
|
| 315 |
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
|
| 316 |
"""
|
| 317 |
-
|
|
|
|
| 318 |
respond,
|
| 319 |
title="AI Data Management Expert",
|
| 320 |
description="Hello! I am your Data Management Expert, specialized in Microsoft 365. I'm here to help you with guidance on Data Management procedures. How can I assist you today?",
|
|
@@ -412,6 +509,39 @@ IMPORTANT: If a question has been asked before in the conversation, acknowledge
|
|
| 412 |
],
|
| 413 |
)
|
| 414 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 415 |
|
| 416 |
if __name__ == "__main__":
|
| 417 |
demo.launch(server_name="0.0.0.0")
|
|
|
|
| 4 |
import html
|
| 5 |
import re
|
| 6 |
import traceback
|
| 7 |
+
import datetime
|
| 8 |
+
import threading
|
| 9 |
+
from collections import defaultdict
|
| 10 |
+
from flask import request
|
| 11 |
|
| 12 |
"""
|
| 13 |
For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
|
| 14 |
"""
|
| 15 |
client = InferenceClient("Trinoid/Data_Management")
|
| 16 |
|
| 17 |
+
# Rate limiting settings
|
| 18 |
+
MAX_REQUESTS_PER_DAY = 1 # Maximum number of requests per IP per day (set to 1 for testing)
|
| 19 |
+
ip_request_counters = defaultdict(int) # Tracks request count per IP
|
| 20 |
+
ip_last_reset = {} # Tracks when counters were last reset for each IP
|
| 21 |
+
rate_limit_lock = threading.Lock() # Lock for thread-safe counter access
|
| 22 |
+
|
| 23 |
# Comprehensive patterns to filter out thinking and meta-commentary
|
| 24 |
THINKING_PATTERNS = [
|
| 25 |
r"Okay, so I('m| am) (trying to|going to|attempting to)",
|
|
|
|
| 42 |
r"I'll approach this by",
|
| 43 |
]
|
| 44 |
|
| 45 |
+
def get_client_ip():
|
| 46 |
+
"""Get the client's IP address from the request"""
|
| 47 |
+
try:
|
| 48 |
+
# Try to get IP from request context in a Flask environment
|
| 49 |
+
if request:
|
| 50 |
+
return request.remote_addr
|
| 51 |
+
except:
|
| 52 |
+
pass
|
| 53 |
+
|
| 54 |
+
# Fallback if we can't get a real IP
|
| 55 |
+
return "127.0.0.1"
|
| 56 |
+
|
| 57 |
+
def check_rate_limit():
|
| 58 |
+
"""Check if the current IP has exceeded its daily limit"""
|
| 59 |
+
current_ip = get_client_ip()
|
| 60 |
+
current_date = datetime.datetime.now().date()
|
| 61 |
+
|
| 62 |
+
with rate_limit_lock:
|
| 63 |
+
# Reset counter if it's a new day
|
| 64 |
+
if current_ip in ip_last_reset and ip_last_reset[current_ip] != current_date:
|
| 65 |
+
ip_request_counters[current_ip] = 0
|
| 66 |
+
|
| 67 |
+
# Update last reset date
|
| 68 |
+
ip_last_reset[current_ip] = current_date
|
| 69 |
+
|
| 70 |
+
# Check if limit is exceeded
|
| 71 |
+
if ip_request_counters[current_ip] >= MAX_REQUESTS_PER_DAY:
|
| 72 |
+
return False
|
| 73 |
+
|
| 74 |
+
# Increment counter
|
| 75 |
+
ip_request_counters[current_ip] += 1
|
| 76 |
+
return True
|
| 77 |
+
|
| 78 |
def process_final_response(response_text):
|
| 79 |
"""Comprehensive processing of the final response to ensure quality"""
|
| 80 |
|
|
|
|
| 150 |
temperature,
|
| 151 |
top_p,
|
| 152 |
):
|
| 153 |
+
# Check rate limit before processing the request
|
| 154 |
+
if not check_rate_limit():
|
| 155 |
+
current_ip = get_client_ip()
|
| 156 |
+
next_reset = (datetime.datetime.now() + datetime.timedelta(days=1)).replace(hour=0, minute=0, second=0)
|
| 157 |
+
hours_until_reset = int((next_reset - datetime.datetime.now()).total_seconds() / 3600)
|
| 158 |
+
|
| 159 |
+
limit_message = f"""
|
| 160 |
+
<div class="rate-limit-warning">
|
| 161 |
+
<h3>Daily Request Limit Reached</h3>
|
| 162 |
+
<p>You've reached the maximum of {MAX_REQUESTS_PER_DAY} requests allowed per day.</p>
|
| 163 |
+
<p>Your limit will reset in approximately {hours_until_reset} hours (at midnight your local time).</p>
|
| 164 |
+
<p>Please try again tomorrow. Thank you for your understanding!</p>
|
| 165 |
+
</div>
|
| 166 |
+
"""
|
| 167 |
+
yield limit_message
|
| 168 |
+
return
|
| 169 |
+
|
| 170 |
# Create a more effective system prompt
|
| 171 |
enhanced_system_message = f"""You are an expert in Microsoft 365 services including SharePoint, OneDrive, Teams, and the Microsoft 365 compliance ecosystem.
|
| 172 |
|
|
|
|
| 262 |
error_msg = f"I apologize, but I encountered an error while generating a response. Error details: {str(e)}"
|
| 263 |
yield error_msg
|
| 264 |
|
| 265 |
+
# Function to get admin stats about rate limiting
|
| 266 |
+
def get_rate_limit_stats():
|
| 267 |
+
stats = {
|
| 268 |
+
"max_requests_per_day": MAX_REQUESTS_PER_DAY,
|
| 269 |
+
"active_ips": len(ip_request_counters),
|
| 270 |
+
"ip_usage": dict(ip_request_counters),
|
| 271 |
+
"last_reset_dates": {ip: date.strftime("%Y-%m-%d") for ip, date in ip_last_reset.items()}
|
| 272 |
+
}
|
| 273 |
+
return stats
|
| 274 |
+
|
| 275 |
+
# Admin interface to adjust rate limit settings
|
| 276 |
+
def update_rate_limit(new_limit):
|
| 277 |
+
global MAX_REQUESTS_PER_DAY
|
| 278 |
+
|
| 279 |
+
if new_limit < 1:
|
| 280 |
+
return {"success": False, "message": "Limit must be at least 1"}
|
| 281 |
+
|
| 282 |
+
old_limit = MAX_REQUESTS_PER_DAY
|
| 283 |
+
MAX_REQUESTS_PER_DAY = new_limit
|
| 284 |
+
|
| 285 |
+
return {
|
| 286 |
+
"success": True,
|
| 287 |
+
"message": f"Rate limit updated from {old_limit} to {new_limit} requests per IP per day"
|
| 288 |
+
}
|
| 289 |
|
| 290 |
# Custom CSS for Plant Wisdom.AI styling
|
| 291 |
custom_css = """
|
|
|
|
| 393 |
.thinking-step:last-child {
|
| 394 |
border-bottom: none;
|
| 395 |
}
|
| 396 |
+
|
| 397 |
+
/* Rate limit warning styling */
|
| 398 |
+
.rate-limit-warning {
|
| 399 |
+
background-color: #fff3cd;
|
| 400 |
+
color: #856404;
|
| 401 |
+
border: 1px solid #ffeeba;
|
| 402 |
+
border-radius: 8px;
|
| 403 |
+
padding: 16px;
|
| 404 |
+
margin: 16px 0;
|
| 405 |
+
text-align: center;
|
| 406 |
+
font-weight: 500;
|
| 407 |
+
}
|
| 408 |
"""
|
| 409 |
|
| 410 |
"""
|
| 411 |
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
|
| 412 |
"""
|
| 413 |
+
# Main chat interface
|
| 414 |
+
chat_interface = gr.ChatInterface(
|
| 415 |
respond,
|
| 416 |
title="AI Data Management Expert",
|
| 417 |
description="Hello! I am your Data Management Expert, specialized in Microsoft 365. I'm here to help you with guidance on Data Management procedures. How can I assist you today?",
|
|
|
|
| 509 |
],
|
| 510 |
)
|
| 511 |
|
| 512 |
+
# Create Gradio Blocks app with both interfaces
|
| 513 |
+
with gr.Blocks(theme=gr.themes.Base()) as demo:
|
| 514 |
+
# Main chat interface
|
| 515 |
+
chat_interface.render()
|
| 516 |
+
|
| 517 |
+
# Hidden admin panel (accessible at /admin)
|
| 518 |
+
with gr.Tab("Admin", visible=False):
|
| 519 |
+
gr.Markdown("## Rate Limiting Settings")
|
| 520 |
+
|
| 521 |
+
with gr.Row():
|
| 522 |
+
current_limit = gr.Number(label="Current Request Limit Per IP Per Day", value=MAX_REQUESTS_PER_DAY, interactive=False)
|
| 523 |
+
new_limit = gr.Number(label="New Request Limit", value=MAX_REQUESTS_PER_DAY, minimum=1, step=1)
|
| 524 |
+
update_btn = gr.Button("Update Limit")
|
| 525 |
+
|
| 526 |
+
result_text = gr.Textbox(label="Result", interactive=False)
|
| 527 |
+
|
| 528 |
+
# Stats display
|
| 529 |
+
with gr.Accordion("Usage Statistics", open=False):
|
| 530 |
+
refresh_btn = gr.Button("Refresh Stats")
|
| 531 |
+
stats_json = gr.JSON(label="Current Usage Stats")
|
| 532 |
+
|
| 533 |
+
# Connect buttons to functions
|
| 534 |
+
update_btn.click(
|
| 535 |
+
fn=update_rate_limit,
|
| 536 |
+
inputs=[new_limit],
|
| 537 |
+
outputs=[result_text, current_limit]
|
| 538 |
+
)
|
| 539 |
+
|
| 540 |
+
refresh_btn.click(
|
| 541 |
+
fn=get_rate_limit_stats,
|
| 542 |
+
inputs=[],
|
| 543 |
+
outputs=[stats_json]
|
| 544 |
+
)
|
| 545 |
|
| 546 |
if __name__ == "__main__":
|
| 547 |
demo.launch(server_name="0.0.0.0")
|