Spaces:
Running
Running
File size: 8,424 Bytes
8c35759 |
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 |
"""
Query result caching for improved performance.
Provides in-memory caching with TTL for query results,
reducing latency and API costs for repeated queries.
"""
from __future__ import annotations
import hashlib
import threading
import time
from dataclasses import dataclass, field
from typing import Any, Dict, List, Optional
from langchain.schema import Document
@dataclass
class CacheEntry:
"""A single cache entry with TTL support."""
value: Any
timestamp: float
ttl: float
hits: int = 0
def is_expired(self) -> bool:
"""Check if entry has expired.
Returns:
True if entry is past its TTL.
"""
return time.time() - self.timestamp > self.ttl
def access(self) -> Any:
"""Access the cached value and increment hit counter.
Returns:
Cached value.
"""
self.hits += 1
return self.value
class QueryCache:
"""In-memory cache for query results with TTL."""
def __init__(
self,
default_ttl: float = 3600,
max_size: int = 1000,
) -> None:
self._cache: Dict[str, CacheEntry] = {}
self._lock = threading.RLock()
self.default_ttl = default_ttl
self.max_size = max_size
self._total_hits = 0
self._total_misses = 0
def _make_key(self, query: str, context_hash: str = "") -> str:
"""Create cache key from query and context.
Args:
query: Query string.
context_hash: Optional context identifier.
Returns:
MD5 hash key.
"""
combined = f"{query.lower().strip()}:{context_hash}"
return hashlib.md5(combined.encode()).hexdigest()
def _evict_if_needed(self) -> None:
"""Evict oldest entries if cache is at capacity."""
if len(self._cache) < self.max_size:
return
# Remove expired entries first
self.cleanup_expired()
# If still over capacity, remove oldest entries
if len(self._cache) >= self.max_size:
sorted_keys = sorted(
self._cache.keys(),
key=lambda k: self._cache[k].timestamp
)
# Remove oldest 10%
to_remove = max(1, len(sorted_keys) // 10)
for key in sorted_keys[:to_remove]:
del self._cache[key]
def get(
self,
query: str,
context_hash: str = ""
) -> Optional[Any]:
"""Get cached result if exists and not expired.
Args:
query: Query string.
context_hash: Optional context identifier.
Returns:
Cached value or None if not found/expired.
"""
key = self._make_key(query, context_hash)
with self._lock:
entry = self._cache.get(key)
if entry is None:
self._total_misses += 1
return None
if entry.is_expired():
del self._cache[key]
self._total_misses += 1
return None
self._total_hits += 1
return entry.access()
def set(
self,
query: str,
context_hash: str,
value: Any,
ttl: Optional[float] = None,
) -> None:
"""Cache a result.
Args:
query: Query string.
context_hash: Context identifier.
value: Value to cache.
ttl: Optional TTL override.
"""
key = self._make_key(query, context_hash)
with self._lock:
self._evict_if_needed()
self._cache[key] = CacheEntry(
value=value,
timestamp=time.time(),
ttl=ttl or self.default_ttl,
)
def invalidate(self, query: str, context_hash: str = "") -> bool:
"""Invalidate a specific cache entry.
Args:
query: Query string.
context_hash: Context identifier.
Returns:
True if entry was found and removed.
"""
key = self._make_key(query, context_hash)
with self._lock:
if key in self._cache:
del self._cache[key]
return True
return False
def invalidate_all(self) -> int:
"""Clear entire cache.
Returns:
Number of entries cleared.
"""
with self._lock:
count = len(self._cache)
self._cache.clear()
return count
def cleanup_expired(self) -> int:
"""Remove expired entries.
Returns:
Number of entries removed.
"""
with self._lock:
expired_keys = [
k for k, v in self._cache.items() if v.is_expired()
]
for key in expired_keys:
del self._cache[key]
return len(expired_keys)
def get_stats(self) -> Dict[str, Any]:
"""Get cache statistics.
Returns:
Dictionary with cache metrics.
"""
with self._lock:
total_requests = self._total_hits + self._total_misses
hit_rate = (
self._total_hits / total_requests
if total_requests > 0
else 0.0
)
return {
"size": len(self._cache),
"max_size": self.max_size,
"total_hits": self._total_hits,
"total_misses": self._total_misses,
"hit_rate": round(hit_rate, 3),
"default_ttl": self.default_ttl,
}
class AnswerCache(QueryCache):
"""Specialized cache for GraphRAG answers.
Extends QueryCache with answer-specific functionality like
caching both the answer and supporting documents.
"""
@dataclass
class AnswerEntry:
"""Cached answer with supporting documents."""
answer: str
documents: List[Document] = field(default_factory=list)
cypher_result: str = ""
metadata: Dict[str, Any] = field(default_factory=dict)
def set_answer(
self,
query: str,
answer: str,
documents: Optional[List[Document]] = None,
cypher_result: str = "",
context_hash: str = "",
ttl: Optional[float] = None,
) -> None:
"""Cache a complete answer with metadata.
Args:
query: User query.
answer: Generated answer.
documents: Supporting documents.
cypher_result: Cypher query result if any.
context_hash: Context identifier.
ttl: Optional TTL override.
"""
entry = self.AnswerEntry(
answer=answer,
documents=documents or [],
cypher_result=cypher_result,
metadata={"cached_at": time.time()},
)
self.set(query, context_hash, entry, ttl)
def get_answer(
self,
query: str,
context_hash: str = ""
) -> Optional[AnswerEntry]:
"""Get cached answer.
Args:
query: User query.
context_hash: Context identifier.
Returns:
AnswerEntry or None if not found.
"""
result = self.get(query, context_hash)
if isinstance(result, self.AnswerEntry):
return result
return None
# Global cache instances
_query_cache: Optional[QueryCache] = None
_answer_cache: Optional[AnswerCache] = None
def get_query_cache(
default_ttl: float = 3600,
max_size: int = 1000,
) -> QueryCache:
"""Get or create the global query cache.
Args:
default_ttl: Default TTL for entries.
max_size: Maximum cache size.
Returns:
QueryCache singleton instance.
"""
global _query_cache
if _query_cache is None:
_query_cache = QueryCache(default_ttl=default_ttl, max_size=max_size)
return _query_cache
def get_answer_cache(
default_ttl: float = 3600,
max_size: int = 500,
) -> AnswerCache:
"""Get or create the global answer cache.
Args:
default_ttl: Default TTL for entries.
max_size: Maximum cache size.
Returns:
AnswerCache singleton instance.
"""
global _answer_cache
if _answer_cache is None:
_answer_cache = AnswerCache(default_ttl=default_ttl, max_size=max_size)
return _answer_cache
|