File size: 12,893 Bytes
30e6218 | 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 | from typing import Optional, Dict, Any, List, Tuple
from dataclasses import dataclass, asdict
from datetime import datetime, timedelta
import json
import hashlib
import numpy as np
from sklearn.metrics.pairwise import cosine_similarity
import redis
from redis.exceptions import RedisError
import logging
import pickle
from enum import Enum
from collections import OrderedDict
import asyncio
from functools import wraps
logger = logging.getLogger(__name__)
class CacheStrategy(str, Enum):
EXACT_MATCH = "exact_match"
SEMANTIC_SIMILARITY = "semantic_similarity"
HYBRID = "hybrid"
class TTLStrategy(str, Enum):
FIXED = "fixed"
SLIDING_WINDOW = "sliding_window"
ADAPTIVE = "adaptive"
@dataclass
class CacheEntry:
query: str
query_embedding: List[float]
response: str
metadata: Dict[str, Any]
timestamp: datetime
hit_count: int = 0
last_accessed: datetime = None
ttl_seconds: int = 3600
class SemanticCache:
def __init__(
self,
redis_client: Optional[redis.Redis] = None,
similarity_threshold: float = 0.95,
max_cache_size: int = 10000,
default_ttl: int = 3600,
cache_strategy: CacheStrategy = CacheStrategy.HYBRID,
ttl_strategy: TTLStrategy = TTLStrategy.SLIDING_WINDOW
):
self.redis_client = redis_client or self._create_redis_client()
self.similarity_threshold = similarity_threshold
self.max_cache_size = max_cache_size
self.default_ttl = default_ttl
self.cache_strategy = cache_strategy
self.ttl_strategy = ttl_strategy
self.local_cache: OrderedDict[str, CacheEntry] = OrderedDict()
self.embeddings_cache: Dict[str, List[float]] = {}
self.stats = {
"total_queries": 0,
"cache_hits": 0,
"cache_misses": 0,
"semantic_hits": 0,
"exact_hits": 0,
"avg_similarity_score": 0.0
}
def _create_redis_client(self) -> redis.Redis:
try:
client = redis.Redis(
host='localhost',
port=6379,
db=0,
decode_responses=False,
socket_connect_timeout=5
)
client.ping()
logger.info("Connected to Redis for semantic caching")
return client
except RedisError as e:
logger.warning(f"Redis connection failed: {e}. Using in-memory cache only.")
return None
def _get_cache_key(self, query: str) -> str:
return f"semantic_cache:{hashlib.md5(query.encode()).hexdigest()}"
def _calculate_ttl(self, entry: CacheEntry) -> int:
if self.ttl_strategy == TTLStrategy.FIXED:
return self.default_ttl
elif self.ttl_strategy == TTLStrategy.SLIDING_WINDOW:
base_ttl = self.default_ttl
extension = min(entry.hit_count * 600, 7200)
return base_ttl + extension
elif self.ttl_strategy == TTLStrategy.ADAPTIVE:
now = datetime.now()
time_since_creation = (now - entry.timestamp).seconds
if entry.hit_count > 10:
return 7200
elif entry.hit_count > 5:
return 3600
elif time_since_creation < 300:
return 1800
else:
return self.default_ttl
return self.default_ttl
def _find_semantic_match(
self,
query_embedding: List[float],
top_k: int = 5
) -> Optional[Tuple[str, float, CacheEntry]]:
if not self.embeddings_cache:
return None
query_vec = np.array(query_embedding).reshape(1, -1)
best_match = None
best_score = 0.0
for cached_query, cached_embedding in self.embeddings_cache.items():
cached_vec = np.array(cached_embedding).reshape(1, -1)
similarity = cosine_similarity(query_vec, cached_vec)[0][0]
if similarity >= self.similarity_threshold and similarity > best_score:
if cached_query in self.local_cache:
best_match = (cached_query, similarity, self.local_cache[cached_query])
best_score = similarity
return best_match
def _evict_if_needed(self):
if len(self.local_cache) >= self.max_cache_size:
num_to_evict = len(self.local_cache) // 10
sorted_entries = sorted(
self.local_cache.items(),
key=lambda x: (x[1].hit_count, x[1].last_accessed or x[1].timestamp)
)
for key, _ in sorted_entries[:num_to_evict]:
del self.local_cache[key]
if key in self.embeddings_cache:
del self.embeddings_cache[key]
if self.redis_client:
try:
self.redis_client.delete(self._get_cache_key(key))
except RedisError:
pass
async def get_async(
self,
query: str,
query_embedding: Optional[List[float]] = None
) -> Optional[str]:
return await asyncio.get_event_loop().run_in_executor(
None,
self.get,
query,
query_embedding
)
def get(
self,
query: str,
query_embedding: Optional[List[float]] = None
) -> Optional[str]:
self.stats["total_queries"] += 1
if self.cache_strategy in [CacheStrategy.EXACT_MATCH, CacheStrategy.HYBRID]:
if query in self.local_cache:
entry = self.local_cache[query]
entry.hit_count += 1
entry.last_accessed = datetime.now()
self.local_cache.move_to_end(query)
self.stats["cache_hits"] += 1
self.stats["exact_hits"] += 1
logger.debug(f"Exact cache hit for query: {query[:50]}...")
return entry.response
if self.redis_client:
try:
cache_key = self._get_cache_key(query)
cached_data = self.redis_client.get(cache_key)
if cached_data:
entry = pickle.loads(cached_data)
entry.hit_count += 1
entry.last_accessed = datetime.now()
self.local_cache[query] = entry
if len(self.local_cache) > self.max_cache_size:
self.local_cache.popitem(last=False)
ttl = self._calculate_ttl(entry)
self.redis_client.setex(cache_key, ttl, pickle.dumps(entry))
self.stats["cache_hits"] += 1
self.stats["exact_hits"] += 1
return entry.response
except RedisError as e:
logger.warning(f"Redis get error: {e}")
if self.cache_strategy in [CacheStrategy.SEMANTIC_SIMILARITY, CacheStrategy.HYBRID]:
if query_embedding:
match = self._find_semantic_match(query_embedding)
if match:
matched_query, similarity, entry = match
entry.hit_count += 1
entry.last_accessed = datetime.now()
self.stats["cache_hits"] += 1
self.stats["semantic_hits"] += 1
self.stats["avg_similarity_score"] = (
self.stats["avg_similarity_score"] * 0.9 + similarity * 0.1
)
logger.debug(
f"Semantic cache hit (similarity: {similarity:.3f}) "
f"for query: {query[:50]}..."
)
return entry.response
self.stats["cache_misses"] += 1
return None
async def set_async(
self,
query: str,
response: str,
query_embedding: Optional[List[float]] = None,
metadata: Optional[Dict[str, Any]] = None
):
return await asyncio.get_event_loop().run_in_executor(
None,
self.set,
query,
response,
query_embedding,
metadata
)
def set(
self,
query: str,
response: str,
query_embedding: Optional[List[float]] = None,
metadata: Optional[Dict[str, Any]] = None
):
self._evict_if_needed()
entry = CacheEntry(
query=query,
query_embedding=query_embedding or [],
response=response,
metadata=metadata or {},
timestamp=datetime.now(),
last_accessed=datetime.now(),
ttl_seconds=self.default_ttl
)
self.local_cache[query] = entry
if query_embedding:
self.embeddings_cache[query] = query_embedding
if self.redis_client:
try:
cache_key = self._get_cache_key(query)
ttl = self._calculate_ttl(entry)
self.redis_client.setex(cache_key, ttl, pickle.dumps(entry))
except RedisError as e:
logger.warning(f"Redis set error: {e}")
logger.debug(f"Cached response for query: {query[:50]}...")
def invalidate(self, query: str):
if query in self.local_cache:
del self.local_cache[query]
if query in self.embeddings_cache:
del self.embeddings_cache[query]
if self.redis_client:
try:
self.redis_client.delete(self._get_cache_key(query))
except RedisError:
pass
def clear(self):
self.local_cache.clear()
self.embeddings_cache.clear()
if self.redis_client:
try:
pattern = "semantic_cache:*"
for key in self.redis_client.scan_iter(match=pattern):
self.redis_client.delete(key)
except RedisError as e:
logger.warning(f"Redis clear error: {e}")
def get_statistics(self) -> Dict[str, Any]:
hit_rate = (
self.stats["cache_hits"] / self.stats["total_queries"]
if self.stats["total_queries"] > 0
else 0.0
)
return {
"total_queries": self.stats["total_queries"],
"cache_hits": self.stats["cache_hits"],
"cache_misses": self.stats["cache_misses"],
"hit_rate": hit_rate,
"semantic_hits": self.stats["semantic_hits"],
"exact_hits": self.stats["exact_hits"],
"avg_similarity_score": self.stats["avg_similarity_score"],
"cache_size": len(self.local_cache),
"embeddings_cached": len(self.embeddings_cache),
"cache_strategy": self.cache_strategy.value,
"ttl_strategy": self.ttl_strategy.value,
"similarity_threshold": self.similarity_threshold
}
def semantic_cache_decorator(
cache: SemanticCache,
get_embedding_func=None
):
def decorator(func):
@wraps(func)
async def async_wrapper(query: str, *args, **kwargs):
embedding = None
if get_embedding_func:
embedding = await get_embedding_func(query)
cached_response = await cache.get_async(query, embedding)
if cached_response:
return cached_response
response = await func(query, *args, **kwargs)
await cache.set_async(query, response, embedding)
return response
@wraps(func)
def sync_wrapper(query: str, *args, **kwargs):
embedding = None
if get_embedding_func:
embedding = get_embedding_func(query)
cached_response = cache.get(query, embedding)
if cached_response:
return cached_response
response = func(query, *args, **kwargs)
cache.set(query, response, embedding)
return response
if asyncio.iscoroutinefunction(func):
return async_wrapper
else:
return sync_wrapper
return decorator |