Spaces:
Sleeping
Sleeping
File size: 16,464 Bytes
e0aa230 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 |
"""
Embedding Generator Module
This module is responsible for generating vector embeddings for text chunks
using Gemini Embedding v3 with complete API integration.
Technology: Gemini Embedding v3 (gemini-embedding-exp-03-07)
"""
import logging
import os
import time
import hashlib
from datetime import datetime, timedelta
from typing import Dict, List, Any, Optional, Union
import json
# Import Gemini API and caching libraries
try:
import google.generativeai as genai
from cachetools import TTLCache
except ImportError as e:
logging.warning(f"Some embedding libraries are not installed: {e}")
from utils.error_handler import EmbeddingError, error_handler, ErrorType
class EmbeddingGenerator:
"""
Generates vector embeddings for text chunks using Gemini Embedding v3 with full functionality.
Features:
- Gemini Embedding v3 API integration
- Batch processing with rate limiting
- Intelligent retry logic with exponential backoff
- Embedding caching mechanism
- Cost optimization
"""
def __init__(self, config: Optional[Dict[str, Any]] = None):
"""
Initialize the EmbeddingGenerator with configuration.
Args:
config: Configuration dictionary with API parameters
"""
self.config = config or {}
self.logger = logging.getLogger(__name__)
# API Configuration
self.api_key = self.config.get("api_key", os.environ.get("GEMINI_API_KEY"))
self.model = self.config.get("model", "gemini-embedding-exp-03-07")
self.batch_size = self.config.get("batch_size", 5)
self.max_retries = self.config.get("max_retries", 3)
self.retry_delay = self.config.get("retry_delay", 1)
# Performance settings
self.rate_limit_delay = self.config.get("rate_limit_delay", 0.1)
self.max_text_length = self.config.get(
"max_text_length", 8192
) # ✨ 8K token limit for latest model
self.enable_caching = self.config.get("enable_caching", True)
self.cache_ttl = self.config.get("cache_ttl", 3600) # 1 hour
# Statistics tracking
self.stats = {
"total_requests": 0,
"successful_requests": 0,
"failed_requests": 0,
"cache_hits": 0,
"total_tokens_processed": 0,
"start_time": datetime.now(),
}
# Initialize cache
if self.enable_caching:
self.cache = TTLCache(maxsize=1000, ttl=self.cache_ttl)
else:
self.cache = None
# Validate and initialize API client
self._initialize_client()
def _initialize_client(self):
"""Initialize Gemini API client with validation."""
if not self.api_key:
self.logger.warning(
"No Gemini API key provided. Embeddings will not be generated."
)
self.client = None
return
try:
# Configure Gemini API
genai.configure(api_key=self.api_key)
# Test API connection
self._test_api_connection()
self.client = genai
self.logger.info("Gemini API client initialized successfully")
except Exception as e:
self.logger.error(f"Failed to initialize Gemini API client: {str(e)}")
self.client = None
def _test_api_connection(self):
"""Test API connection with a simple request."""
try:
# Test with a simple embedding request
test_result = genai.embed_content(
model=self.model,
content="test connection",
task_type="retrieval_document",
)
if not test_result.get("embedding"):
raise Exception("No embedding returned from test request")
self.logger.info("API connection test successful")
except Exception as e:
raise EmbeddingError(f"API connection test failed: {str(e)}")
@error_handler(ErrorType.EMBEDDING_GENERATION)
def generate_embeddings(self, texts: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
"""
Generate embeddings for a list of text chunks with full functionality.
Args:
texts: List of dictionaries containing text chunks and metadata
Each dict should have 'content' and 'metadata' keys
Returns:
List of dictionaries with original content, metadata, and embeddings
"""
if not self.client or not texts:
self.logger.warning("No API client or empty text list")
return texts
self.logger.info(f"Generating embeddings for {len(texts)} text chunks")
start_time = time.time()
# Filter and validate texts
valid_texts = self._validate_texts(texts)
if not valid_texts:
self.logger.warning("No valid texts to process")
return texts
# Process in batches to respect API limits
results = []
total_batches = (len(valid_texts) + self.batch_size - 1) // self.batch_size
for i in range(0, len(valid_texts), self.batch_size):
batch_num = (i // self.batch_size) + 1
batch = valid_texts[i : i + self.batch_size]
self.logger.info(
f"Processing batch {batch_num}/{total_batches} ({len(batch)} items)"
)
try:
batch_results = self._process_batch(batch)
results.extend(batch_results)
# Rate limiting between batches
if i + self.batch_size < len(valid_texts):
time.sleep(self.rate_limit_delay)
except Exception as e:
self.logger.error(f"Batch {batch_num} failed: {str(e)}")
# Add original items without embeddings
for item in batch:
item_copy = item.copy()
item_copy["embedding"] = []
item_copy["embedding_error"] = str(e)
results.append(item_copy)
# Update statistics
processing_time = time.time() - start_time
self.logger.info(f"Embedding generation completed in {processing_time:.2f}s")
return results
def _validate_texts(self, texts: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
"""
Validate and filter text inputs.
Args:
texts: List of text dictionaries
Returns:
List of valid text dictionaries
"""
valid_texts = []
for i, item in enumerate(texts):
if not isinstance(item, dict) or "content" not in item:
self.logger.warning(f"Invalid item at index {i}: missing 'content' key")
continue
content = item["content"]
if not content or not isinstance(content, str):
self.logger.warning(
f"Invalid content at index {i}: empty or non-string"
)
continue
# Truncate if too long
if len(content) > self.max_text_length:
self.logger.warning(
f"Truncating text at index {i}: {len(content)} -> {self.max_text_length} chars"
)
item = item.copy()
item["content"] = content[: self.max_text_length]
item["metadata"] = item.get("metadata", {})
item["metadata"]["truncated"] = True
item["metadata"]["original_length"] = len(content)
valid_texts.append(item)
return valid_texts
def _process_batch(self, batch: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
"""
Process a batch of text chunks to generate embeddings.
Args:
batch: List of dictionaries containing text chunks and metadata
Returns:
List of dictionaries with original content, metadata, and embeddings
"""
# Extract content and check cache
contents = []
cache_results = {}
for i, item in enumerate(batch):
content = item["content"]
# Check cache first
if self.cache is not None:
cache_key = self._get_cache_key(content)
if cache_key in self.cache:
cache_results[i] = self.cache[cache_key]
self.stats["cache_hits"] += 1
continue
contents.append((i, content))
# Generate embeddings for non-cached content
embeddings_map = {}
if contents:
content_texts = [content for _, content in contents]
embeddings = self._generate_with_retry(content_texts)
# Map embeddings back to indices
for j, (original_index, content) in enumerate(contents):
if j < len(embeddings):
embedding = embeddings[j]
embeddings_map[original_index] = embedding
# Cache the result
if self.cache is not None:
cache_key = self._get_cache_key(content)
self.cache[cache_key] = embedding
# 🔗 Combine results
results = []
for i, item in enumerate(batch):
result = item.copy()
# Add embedding from cache or new generation
if i in cache_results:
result["embedding"] = cache_results[i]
result["embedding_source"] = "cache"
elif i in embeddings_map:
result["embedding"] = embeddings_map[i]
result["embedding_source"] = "api"
else:
result["embedding"] = []
result["embedding_source"] = "failed"
self.logger.warning(f"Missing embedding for batch item {i}")
# Add embedding metadata
if result["embedding"]:
result["metadata"] = result.get("metadata", {})
result["metadata"].update(
{
"embedding_model": self.model,
"embedding_dimension": len(result["embedding"]),
"embedding_generated_at": datetime.now().isoformat(),
}
)
results.append(result)
return results
def _generate_with_retry(self, texts: List[str]) -> List[List[float]]:
"""
Generate embeddings with intelligent retry logic.
Args:
texts: List of text strings to embed
Returns:
List of embedding vectors (each is a list of floats)
"""
for attempt in range(self.max_retries):
try:
self.stats["total_requests"] += 1
# Generate embeddings using Gemini API
embeddings = []
for text in texts:
try:
# Track tokens
self.stats["total_tokens_processed"] += len(text.split())
# Call Gemini API
result = self.client.embed_content(
model=self.model,
content=text,
task_type="retrieval_document",
title="Document chunk for RAG system",
)
if result and "embedding" in result:
embeddings.append(result["embedding"])
else:
self.logger.warning(
f"No embedding in API response for text: {text[:50]}..."
)
embeddings.append([])
except Exception as e:
self.logger.warning(
f"Failed to embed individual text: {str(e)}"
)
embeddings.append([])
self.stats["successful_requests"] += 1
return embeddings
except Exception as e:
self.stats["failed_requests"] += 1
self.logger.warning(
f"Embedding generation failed (attempt {attempt+1}/{self.max_retries}): {str(e)}"
)
if attempt < self.max_retries - 1:
# Exponential backoff with jitter
delay = self.retry_delay * (2**attempt) + (time.time() % 1)
self.logger.info(f"Retrying in {delay:.1f} seconds...")
time.sleep(delay)
# All retries failed
self.logger.error("All embedding generation attempts failed")
return [[] for _ in texts]
@error_handler(ErrorType.EMBEDDING_GENERATION)
def generate_query_embedding(self, query: str) -> List[float]:
"""
Generate embedding for a single query string.
Args:
query: Query text to embed
Returns:
Embedding vector as a list of floats
"""
if not self.client or not query:
return []
self.logger.info(f"Generating embedding for query: {query[:50]}...")
# Check cache first
if self.cache is not None:
cache_key = self._get_cache_key(query, "query")
if cache_key in self.cache:
self.stats["cache_hits"] += 1
return self.cache[cache_key]
# Generate embedding
embeddings = self._generate_with_retry([query])
embedding = embeddings[0] if embeddings else []
# Cache the result
if embedding and self.cache is not None:
cache_key = self._get_cache_key(query, "query")
self.cache[cache_key] = embedding
return embedding
def _get_cache_key(self, text: str, prefix: str = "doc") -> str:
"""
Generate cache key for text.
Args:
text: Text content
prefix: Key prefix
Returns:
Cache key string
"""
# 🔐 Create hash of text + model for unique key
content_hash = hashlib.md5(f"{self.model}:{text}".encode()).hexdigest()
return f"{prefix}:{content_hash}"
def get_statistics(self) -> Dict[str, Any]:
"""
Get embedding generation statistics.
Returns:
Dictionary with statistics
"""
runtime = datetime.now() - self.stats["start_time"]
return {
**self.stats,
"runtime_seconds": runtime.total_seconds(),
"cache_hit_rate": (
self.stats["cache_hits"] / max(1, self.stats["total_requests"]) * 100
),
"success_rate": (
self.stats["successful_requests"]
/ max(1, self.stats["total_requests"])
* 100
),
"avg_tokens_per_request": (
self.stats["total_tokens_processed"]
/ max(1, self.stats["total_requests"])
),
"cache_size": len(self.cache) if self.cache else 0,
"model": self.model,
"batch_size": self.batch_size,
}
def clear_cache(self):
"""Clear the embedding cache."""
if self.cache:
self.cache.clear()
self.logger.info("Embedding cache cleared")
def warm_up_cache(self, sample_texts: List[str]):
"""
🔥 Warm up the cache with sample texts.
Args:
sample_texts: List of sample texts to pre-generate embeddings
"""
if not sample_texts:
return
self.logger.info(f"🔥 Warming up cache with {len(sample_texts)} sample texts")
sample_items = [{"content": text, "metadata": {}} for text in sample_texts]
self.generate_embeddings(sample_items)
self.logger.info("Cache warm-up completed")
|