Spaces:
Running
Running
File size: 14,791 Bytes
690700c |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 |
# """
# Multi-LLM Manager for Google Gemini, Groq, and HuggingFace
# All three APIs co-exist for different purposes (no fallback logic)
# Architecture:
# - Google Gemini (Primary): User-facing chat responses (best quality)
# - Groq (Secondary): Fast inference for evaluation and specific tasks
# - HuggingFace: Model downloads and embeddings (always required)
# Each API has its designated purpose based on config settings.
# """
# import time
# import google.generativeai as genai
# from typing import List, Dict, Optional, Literal
# from langchain_groq import ChatGroq
# from langchain_core.messages import HumanMessage, SystemMessage, AIMessage
# from app.config import settings
# # ============================================================================
# # GOOGLE GEMINI MANAGER
# # ============================================================================
# class GeminiManager:
# """
# Google Gemini API Manager (Primary LLM)
# Handles Google Pro account with gemini-2.0-flash-lite model
# """
# def __init__(self):
# """Initialize Gemini API with your Google API key"""
# self.api_key = settings.GOOGLE_API_KEY
# self.model_name = settings.GEMINI_MODEL
# # Configure Gemini
# genai.configure(api_key=self.api_key)
# # Create model instance with safety settings
# self.model = genai.GenerativeModel(
# model_name=self.model_name,
# generation_config={
# "temperature": settings.LLM_TEMPERATURE,
# "max_output_tokens": settings.LLM_MAX_TOKENS,
# }
# )
# # Rate limiting tracking
# self.requests_this_minute = 0
# self.tokens_this_minute = 0
# self.last_reset = time.time()
# print(f"✅ Gemini Manager initialized: {self.model_name}")
# def _check_rate_limits(self):
# """
# Check and reset rate limit counters.
# Gemini Pro: 60 requests/min, 60,000 tokens/min
# """
# current_time = time.time()
# # Reset counters every minute
# if current_time - self.last_reset > 60:
# self.requests_this_minute = 0
# self.tokens_this_minute = 0
# self.last_reset = current_time
# # Check if limits exceeded
# if self.requests_this_minute >= settings.GEMINI_REQUESTS_PER_MINUTE:
# wait_time = 60 - (current_time - self.last_reset)
# print(f"⚠️ Gemini rate limit hit. Waiting {wait_time:.1f}s...")
# time.sleep(wait_time)
# self._check_rate_limits() # Recursive check after waiting
# async def generate(
# self,
# messages: List[Dict[str, str]],
# system_prompt: Optional[str] = None
# ) -> str:
# """
# Generate response using Gemini.
# Args:
# messages: List of conversation messages
# Format: [{'role': 'user'/'assistant', 'content': '...'}]
# system_prompt: Optional system prompt (prepended to first message)
# Returns:
# str: Generated response text
# """
# self._check_rate_limits()
# try:
# # Format messages for Gemini
# # Gemini uses 'user' and 'model' roles
# formatted_messages = []
# # Add system prompt as first user message if provided
# if system_prompt:
# formatted_messages.append({
# 'role': 'user',
# 'parts': [system_prompt]
# })
# # Convert messages
# for msg in messages:
# role = 'model' if msg['role'] == 'assistant' else 'user'
# formatted_messages.append({
# 'role': role,
# 'parts': [msg['content']]
# })
# # Generate response
# chat = self.model.start_chat(history=formatted_messages[:-1])
# response = chat.send_message(formatted_messages[-1]['parts'][0])
# # Track rate limits
# self.requests_this_minute += 1
# # Note: Token counting would require additional API call
# # For now, estimate ~4 chars per token
# estimated_tokens = len(response.text) // 4
# self.tokens_this_minute += estimated_tokens
# return response.text
# except Exception as e:
# print(f"❌ Gemini API error: {e}")
# raise
# # ============================================================================
# # GROQ MANAGER
# # ============================================================================
# class GroqManager:
# """
# Groq API Manager (Secondary LLM)
# Handles fast inference with Llama-3-70B
# """
# def __init__(self):
# """Initialize Groq API with single API key"""
# self.api_key = settings.GROQ_API_KEY
# self.model_name = settings.GROQ_MODEL
# # Create ChatGroq instance
# self.llm = ChatGroq(
# api_key=self.api_key,
# model_name=self.model_name,
# temperature=settings.LLM_TEMPERATURE,
# max_tokens=settings.LLM_MAX_TOKENS
# )
# # Rate limiting tracking
# self.requests_this_minute = 0
# self.tokens_this_minute = 0
# self.last_reset = time.time()
# print(f"✅ Groq Manager initialized: {self.model_name}")
# def _check_rate_limits(self):
# """
# Check and reset rate limit counters.
# Groq Free: 30 requests/min, 30,000 tokens/min
# """
# current_time = time.time()
# # Reset counters every minute
# if current_time - self.last_reset > 60:
# self.requests_this_minute = 0
# self.tokens_this_minute = 0
# self.last_reset = current_time
# # Check if limits exceeded
# if self.requests_this_minute >= settings.GROQ_REQUESTS_PER_MINUTE:
# wait_time = 60 - (current_time - self.last_reset)
# print(f"⚠️ Groq rate limit hit. Waiting {wait_time:.1f}s...")
# time.sleep(wait_time)
# self._check_rate_limits()
# async def generate(
# self,
# messages: List[Dict[str, str]],
# system_prompt: Optional[str] = None
# ) -> str:
# """
# Generate response using Groq.
# Args:
# messages: List of conversation messages
# Format: [{'role': 'user'/'assistant', 'content': '...'}]
# system_prompt: Optional system prompt
# Returns:
# str: Generated response text
# """
# self._check_rate_limits()
# try:
# # Format messages for LangChain
# formatted_messages = []
# # Add system message if provided
# if system_prompt:
# formatted_messages.append(SystemMessage(content=system_prompt))
# # Convert conversation messages
# for msg in messages:
# if msg['role'] == 'user':
# formatted_messages.append(HumanMessage(content=msg['content']))
# elif msg['role'] == 'assistant':
# formatted_messages.append(AIMessage(content=msg['content']))
# # Generate response
# response = await self.llm.ainvoke(formatted_messages)
# # Track rate limits
# self.requests_this_minute += 1
# # Estimate tokens (rough approximation)
# estimated_tokens = len(response.content) // 4
# self.tokens_this_minute += estimated_tokens
# return response.content
# except Exception as e:
# print(f"❌ Groq API error: {e}")
# raise
# # ============================================================================
# # UNIFIED LLM MANAGER (Routes to appropriate LLM)
# # ============================================================================
# class LLMManager:
# """
# Unified LLM Manager that routes requests to appropriate LLM.
# Routing strategy (from config):
# - Chat responses → Gemini (best quality for users)
# - Evaluation → Groq (fast, good enough for RL)
# - Policy → Local BERT (no API call)
# """
# def __init__(self):
# """Initialize all LLM managers"""
# self.gemini = None
# self.groq = None
# # Initialize Gemini if configured
# if settings.is_gemini_enabled():
# try:
# self.gemini = GeminiManager()
# except Exception as e:
# print(f"⚠️ Failed to initialize Gemini: {e}")
# # Initialize Groq if configured
# if settings.is_groq_enabled():
# try:
# self.groq = GroqManager()
# except Exception as e:
# print(f"⚠️ Failed to initialize Groq: {e}")
# print("✅ LLM Manager initialized")
# async def generate(
# self,
# messages: List[Dict[str, str]],
# system_prompt: Optional[str] = None,
# task: Literal["chat", "evaluation"] = "chat"
# ) -> str:
# """
# Generate response using appropriate LLM based on task.
# Args:
# messages: Conversation messages
# system_prompt: Optional system prompt
# task: Task type - "chat" (user-facing) or "evaluation" (RL training)
# Returns:
# str: Generated response
# Raises:
# ValueError: If appropriate LLM is not configured
# """
# # Determine which LLM to use based on task
# llm_choice = settings.get_llm_for_task(task)
# if llm_choice == "gemini":
# if self.gemini is None:
# raise ValueError("Gemini API not configured. Set GOOGLE_API_KEY in .env")
# return await self.gemini.generate(messages, system_prompt)
# elif llm_choice == "groq":
# if self.groq is None:
# raise ValueError("Groq API not configured. Set GROQ_API_KEY in .env")
# return await self.groq.generate(messages, system_prompt)
# else:
# raise ValueError(f"Unknown LLM choice: {llm_choice}")
# # async def generate_chat_response(
# # self,
# # query: str,
# # context: str,
# # history: List[Dict[str, str]]
# # ) -> str:
# # """
# # Generate chat response (uses Gemini by default).
# # Args:
# # query: User query
# # context: Retrieved context (from FAISS)
# # history: Conversation history
# # Returns:
# # str: Chat response
# # """
# # # Build system prompt
# # system_prompt = settings.SYSTEM_PROMPT
# # if context:
# # system_prompt += f"\n\nRelevant Information:\n{context}"
# # # Build messages
# # messages = history + [{'role': 'user', 'content': query}]
# # # Generate using chat LLM (Gemini)
# # return await self.generate(messages, system_prompt, task="chat")
# async def generate_chat_response(
# self,
# query: str,
# context: str,
# history: List[Dict[str, str]]
# ) -> str:
# """Generate chat response (uses Gemini by default)."""
# # Import the detailed prompt
# from app.services.chat_service import BANKING_SYSTEM_PROMPT
# # Build enhanced system prompt with context
# system_prompt = BANKING_SYSTEM_PROMPT
# if context:
# system_prompt += f"\n\nRelevant Knowledge Base Context:\n{context}"
# else:
# system_prompt += "\n\nNo specific banking documents were retrieved for this query. Provide a helpful general response while acknowledging your banking specialization."
# # Build messages
# messages = history + [{'role': 'user', 'content': query}]
# # Generate using chat LLM (Gemini)
# return await self.generate(messages, system_prompt, task="chat")
# async def evaluate_response(
# self,
# query: str,
# response: str,
# context: str = ""
# ) -> Dict:
# """
# Evaluate response quality (uses Groq for speed).
# Used during RL training.
# Args:
# query: User query
# response: Generated response
# context: Retrieved context (if any)
# Returns:
# dict: Evaluation results
# {'quality': 'Good'/'Bad', 'explanation': '...'}
# """
# eval_prompt = f"""Evaluate this response:
# Query: {query}
# Response: {response}
# Context used: {context if context else 'None'}
# Is this response Good or Bad? Respond with just "Good" or "Bad" and brief explanation."""
# messages = [{'role': 'user', 'content': eval_prompt}]
# # Generate using evaluation LLM (Groq)
# result = await self.generate(messages, task="evaluation")
# # Parse result
# quality = "Good" if "Good" in result else "Bad"
# return {
# 'quality': quality,
# 'explanation': result
# }
# # ============================================================================
# # GLOBAL LLM MANAGER INSTANCE
# # ============================================================================
# llm_manager = LLMManager()
# # ============================================================================
# # USAGE EXAMPLE (for reference)
# # ============================================================================
# """
# # In your service file:
# from app.core.llm_manager import llm_manager
# # Generate chat response (uses Gemini)
# response = await llm_manager.generate_chat_response(
# query="What is my account balance?",
# context="Your balance is $1000",
# history=[]
# )
# # Evaluate response (uses Groq)
# evaluation = await llm_manager.evaluate_response(
# query="What is my balance?",
# response="Your balance is $1000",
# context="Balance: $1000"
# )
# """
|