Spaces:
Sleeping
Sleeping
File size: 13,561 Bytes
7637d2f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 |
import os
from dotenv import load_dotenv
import streamlit as st
from streamlit.runtime.scriptrunner import RerunException, StopException
from openai import OpenAI
from pymongo import MongoClient
from pinecone import Pinecone
import uuid
from datetime import datetime
import time
from streamlit.runtime.caching import cache_data
# Load environment variables
load_dotenv()
# Configuration
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
MONGODB_URI = os.getenv("MONGODB_URI")
PINECONE_API_KEY = os.getenv("PINECONE_API_KEY")
PINECONE_ENVIRONMENT = os.getenv("PINECONE_ENVIRONMENT")
PINECONE_INDEX_NAME = os.getenv("PINECONE_INDEX_NAME")
# Initialize clients
openai_client = OpenAI(api_key=OPENAI_API_KEY)
mongo_client = MongoClient(MONGODB_URI)
db = mongo_client["Wall_Street"]
conversation_history = db["conversation_history"]
global_common_memory = db["global_common_memory"] # New global common memory collection
chat_history = db["chat_history"] # New collection for storing all chats
# Initialize Pinecone
pc = Pinecone(api_key=PINECONE_API_KEY)
pinecone_index = pc.Index(PINECONE_INDEX_NAME)
# Set up Streamlit page configuration
st.set_page_config(page_title="GPT-Driven Chat System - User", page_icon="💬", layout="wide")
# Custom CSS to improve the UI
st.markdown("""
<style>
/* Your custom CSS styles */
</style>
""", unsafe_allow_html=True)
# Initialize Streamlit session state
if 'chat_history' not in st.session_state:
st.session_state['chat_history'] = []
if 'user_type' not in st.session_state:
st.session_state['user_type'] = None
if 'session_id' not in st.session_state:
st.session_state['session_id'] = str(uuid.uuid4())
# --- Common Memory Functions ---
@cache_data(ttl=300) # Cache for 5 minutes
def get_global_common_memory():
"""Retrieve the global common memory."""
memory_doc = global_common_memory.find_one({"memory_id": "global_common_memory_id"})
return memory_doc.get('memory', []) if memory_doc else []
# --- Relevant Context Retrieval ---
@cache_data(ttl=60) # Cache for 1 minute
def get_relevant_context(query, top_k=3):
"""
Retrieve relevant context from Pinecone based on the user query.
"""
try:
query_embedding = openai_client.embeddings.create(
model="text-embedding-3-large", # Updated to use the larger model
input=query
).data[0].embedding
results = pinecone_index.query(vector=query_embedding, top_k=top_k, include_metadata=True)
contexts = [item['metadata']['text'] for item in results['matches']]
return " ".join(contexts)
except Exception as e:
print(f"Error retrieving context: {str(e)}")
return ""
# --- GPT Response Function ---
def get_gpt_response(prompt, context=""):
"""
Generates a response from the GPT model based on the user prompt, retrieved context, and global chat memory.
Assesses confidence and marks the response as 'uncertain' if confidence is low.
"""
try:
print(prompt)
#print(context)
common_memory = get_global_common_memory()
system_message = (
"You are a helpful assistant. Use the following context and global chat memory to inform your responses, "
"but don't mention them explicitly unless directly relevant to the user's question. "
"If you are uncertain about the answer, respond with 'I am not sure about that.'"
)
if common_memory:
# Join the memory items into a single string
memory_str = "\n".join(common_memory)
system_message += f"\n\nGlobal Chat Memory:\n{memory_str}"
print(system_message)
messages = [
{"role": "system", "content": system_message},
{"role": "user", "content": f"Context: {context}\n\nUser query: {prompt}"}
]
completion = openai_client.chat.completions.create(
model="gpt-4o-mini",
messages=messages,
temperature=0.7 # Adjust temperature for confidence control
)
response = completion.choices[0].message.content.strip()
print(response)
# Determine if the response indicates uncertainty
is_uncertain = "i am not sure about that" in response.lower()
print(is_uncertain)
return response, is_uncertain
except Exception as e:
return f"Error generating response: {str(e)}", False
# --- Send User Message ---
def send_message(message):
"""
Sends a user message. If the chatbot is uncertain, messages are sent to the operator for approval.
"""
context = get_relevant_context(message)
user_message = {
"role": "user",
"content": message,
"timestamp": datetime.utcnow(),
"status": "approved" # User messages are always approved
}
# Upsert the user message immediately
result = conversation_history.update_one(
{"session_id": st.session_state['session_id']},
{
"$push": {"messages": user_message},
"$set": {"last_updated": datetime.utcnow()},
"$setOnInsert": {"created_at": datetime.utcnow()}
},
upsert=True
)
# Update or create the chat history document
chat_history.update_one(
{"session_id": st.session_state['session_id']},
{
"$push": {"messages": user_message},
"$set": {"last_updated": datetime.utcnow()},
"$setOnInsert": {"created_at": datetime.utcnow()}
},
upsert=True
)
# Update the session state with the user message
st.session_state['chat_history'].append(user_message)
if not st.session_state.get('admin_takeover_active'):
# Generate GPT response if takeover is not active
gpt_response, is_uncertain = get_gpt_response(message, context)
if is_uncertain:
status = "pending" # Mark as pending for operator approval
else:
status = "approved"
assistant_message = {
"role": "assistant",
"content": gpt_response,
"timestamp": datetime.utcnow(),
"status": status
}
# Upsert the assistant message
result = conversation_history.update_one(
{"session_id": st.session_state['session_id']},
{
"$push": {"messages": assistant_message},
"$set": {"last_updated": datetime.utcnow()}
}
)
# Update the chat history document
chat_history.update_one(
{"session_id": st.session_state['session_id']},
{
"$push": {"messages": assistant_message},
"$set": {"last_updated": datetime.utcnow()}
}
)
# Update the session state with the assistant message
st.session_state['chat_history'].append(assistant_message)
# --- Send Admin Message ---
def send_admin_message(message):
"""
Sends an admin message directly to the user during a takeover.
"""
admin_message = {
"role": "admin",
"content": message,
"timestamp": datetime.utcnow(),
"status": "approved"
}
# Upsert the admin message
result = conversation_history.update_one(
{"session_id": st.session_state['session_id']},
{
"$push": {"messages": admin_message},
"$set": {"last_updated": datetime.utcnow()}
}
)
# Update the chat history document
chat_history.update_one(
{"session_id": st.session_state['session_id']},
{
"$push": {"messages": admin_message},
"$set": {"last_updated": datetime.utcnow()}
}
)
# Update the session state with the admin message
st.session_state['chat_history'].append(admin_message)
# --- User Page ---
def user_page():
if 'session_id' not in st.session_state:
st.session_state['session_id'] = str(uuid.uuid4())
st.title("Chat Interface")
chat_col, info_col = st.columns([3, 1])
with chat_col:
# Create a placeholder for the chat interface
chat_placeholder = st.empty()
# Add a manual refresh button
if st.button("Refresh Chat"):
fetch_and_update_chat()
# Handle new user input outside the loop
user_input = st.chat_input("Type your message here...", key="user_chat_input")
if user_input:
send_message(user_input)
# If admin takeover is active, allow admin to send messages
if st.session_state.get('admin_takeover_active'):
admin_input = st.chat_input("Admin is currently taking over the chat...", key="admin_chat_input")
if admin_input:
send_admin_message(admin_input)
# Main loop for continuous updates
while True:
with chat_placeholder.container():
# Display all messages in the chat history
for message in st.session_state['chat_history']:
if message["role"] == "user":
with st.chat_message("user"):
st.markdown(message["content"])
elif message["role"] == "assistant":
with st.chat_message("assistant"):
if message.get("status") == "approved":
st.markdown(message["content"])
elif message.get("status") == "pending":
st.info("This response is pending operator approval...")
elif message["role"] == "admin":
with st.chat_message("admin"):
st.markdown(f"**Admin:** {message['content']}")
# Fetch updates every 5 seconds
fetch_and_update_chat()
time.sleep(5)
with info_col:
st.subheader("Session Information")
stat_cols = st.columns(2)
with stat_cols[0]:
st.write(f"**Session ID:** {st.session_state['session_id'][:8]}...")
with stat_cols[1]:
st.write(f"**User Type:** {st.session_state.get('user_type', 'Regular User')}")
stat_cols = st.columns(2)
with stat_cols[0]:
st.write(f"**Chat History Count:** {len(st.session_state['chat_history'])}")
with stat_cols[1]:
st.write(f"**Last Active:** {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
# Add more user-specific information or settings as needed
# --- Fetch and Update Chat ---
def fetch_and_update_chat():
"""
Fetches the latest chat history from the database and updates the session state.
"""
chat = conversation_history.find_one({"session_id": st.session_state['session_id']})
if chat:
st.session_state['chat_history'] = chat.get('messages', [])
else:
st.session_state['chat_history'] = []
# Check if admin takeover is active
takeover_status = db.takeover_status.find_one({"session_id": st.session_state['session_id']})
is_active = takeover_status and takeover_status.get("active", False)
st.session_state['admin_takeover_active'] = is_active
# --- View Full Chat (User Perspective) ---
def view_full_chat(session_id):
st.title("Full Chat View")
chat = conversation_history.find_one({"session_id": session_id})
if chat:
st.subheader(f"Session ID: {session_id[:8]}...")
last_updated = chat.get('last_updated', datetime.utcnow())
st.write(f"Last Updated: {last_updated}")
# Display global common memory
st.subheader("Global Chat Memory")
common_memory = get_global_common_memory()
if common_memory:
for idx, item in enumerate(common_memory, 1):
st.text(f"{idx}. {item}")
else:
st.info("No global chat memory found.")
# Add button to clear global common memory
if st.button("Clear Global Chat Memory"):
clear_global_common_memory()
st.success("Global chat memory cleared successfully!")
st.rerun()
# Add admin takeover button
if st.button("Admin Takeover"):
st.session_state['admin_takeover'] = session_id
st.rerun()
for message in chat.get('messages', []):
if message['role'] == 'user':
with st.chat_message("user"):
st.markdown(message["content"])
elif message['role'] == 'assistant':
with st.chat_message("assistant"):
if message.get("status") == "approved":
st.markdown(message["content"])
else:
st.info("Waiting for admin approval...")
elif message['role'] == 'admin': # Display admin messages
with st.chat_message("admin"):
st.markdown(f"**Admin:** {message['content']}")
st.caption(f"Timestamp: {message.get('timestamp', 'N/A')}")
else:
st.error("Chat not found.")
if st.button("Back to Admin Dashboard"):
st.session_state.pop('selected_chat', None)
st.rerun()
def main():
try:
user_page()
except (RerunException, StopException):
# These exceptions are used by Streamlit for page navigation and stopping the script
raise
except Exception as e:
st.error(f"An unexpected error occurred: {str(e)}")
if __name__ == "__main__":
main() |