Spaces:
Running
Running
File size: 14,201 Bytes
5e1a30c |
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 |
"""
Knowledge Database Cache System
==============================
Persistent storage for processed documents, chunks, and embeddings to avoid
reprocessing on system restart.
"""
import logging
import pickle
import json
import hashlib
import time
from pathlib import Path
from typing import Dict, Any, List, Optional, Tuple
import numpy as np
from dataclasses import asdict
logger = logging.getLogger(__name__)
class KnowledgeCache:
"""Persistent cache for processed documents and embeddings"""
def __init__(self, cache_dir: Path = Path("cache")):
"""
Initialize knowledge cache
Args:
cache_dir: Directory to store cache files
"""
self.cache_dir = cache_dir
self.cache_dir.mkdir(exist_ok=True)
# Cache file paths
self.metadata_file = self.cache_dir / "metadata.json"
self.documents_file = self.cache_dir / "documents.pkl"
self.embeddings_file = self.cache_dir / "embeddings.npy"
self.index_file = self.cache_dir / "faiss_index.bin"
# In-memory cache
self.metadata = self._load_metadata()
self.documents = None
self.embeddings = None
def _load_metadata(self) -> Dict[str, Any]:
"""Load cache metadata"""
try:
if self.metadata_file.exists():
with open(self.metadata_file, 'r') as f:
return json.load(f)
return {
"version": "1.0",
"created": time.time(),
"last_updated": time.time(),
"document_count": 0,
"chunk_count": 0,
"file_hashes": {},
"embedder_config": None
}
except Exception as e:
logger.error(f"Error loading metadata: {e}")
return self._create_empty_metadata()
def _create_empty_metadata(self) -> Dict[str, Any]:
"""Create empty metadata structure"""
return {
"version": "1.0",
"created": time.time(),
"last_updated": time.time(),
"document_count": 0,
"chunk_count": 0,
"file_hashes": {},
"embedder_config": None
}
def _save_metadata(self):
"""Save metadata to file"""
try:
self.metadata["last_updated"] = time.time()
with open(self.metadata_file, 'w') as f:
json.dump(self.metadata, f, indent=2)
except Exception as e:
logger.error(f"Error saving metadata: {e}")
def _get_file_hash(self, file_path: Path) -> str:
"""Get hash of file for change detection"""
try:
with open(file_path, 'rb') as f:
content = f.read()
return hashlib.md5(content).hexdigest()
except Exception as e:
logger.error(f"Error hashing file {file_path}: {e}")
return ""
def _get_corpus_hash(self, pdf_files: List[Path]) -> str:
"""Get combined hash of all files in corpus"""
file_hashes = []
for pdf_file in sorted(pdf_files):
file_hash = self._get_file_hash(pdf_file)
file_hashes.append(f"{pdf_file.name}:{file_hash}")
combined = "|".join(file_hashes)
return hashlib.md5(combined.encode()).hexdigest()
def is_cache_valid(self, pdf_files: List[Path], embedder_config: Dict[str, Any]) -> bool:
"""
Check if cache is valid for given files and embedder config
Args:
pdf_files: List of PDF files in corpus
embedder_config: Current embedder configuration
Returns:
True if cache is valid and can be used
"""
try:
# Check if cache files exist
if not all(f.exists() for f in [self.documents_file, self.embeddings_file]):
logger.info("Cache files missing, cache invalid")
return False
# Check if metadata exists
if not self.metadata or self.metadata.get("document_count", 0) == 0:
logger.info("No metadata or empty cache, cache invalid")
return False
# Check embedder configuration hash
current_config_hash = create_embedder_config_hash(embedder_config)
cached_config_hash = self.metadata.get("embedder_config_hash")
if current_config_hash != cached_config_hash:
logger.info("Embedder configuration changed, cache invalid")
return False
# Check file count
if len(pdf_files) != self.metadata.get("document_count", 0):
logger.info(f"Document count changed: {len(pdf_files)} vs {self.metadata.get('document_count', 0)}")
return False
# Quick check: if no files have changed timestamps, cache is likely valid
all_files_unchanged = True
for pdf_file in pdf_files:
if not pdf_file.exists():
logger.info(f"File missing: {pdf_file.name}")
return False
# Check modification time first (faster than hashing)
cached_mtime = self.metadata.get("file_mtimes", {}).get(pdf_file.name)
current_mtime = pdf_file.stat().st_mtime
if cached_mtime != current_mtime:
all_files_unchanged = False
break
if all_files_unchanged:
logger.info("Cache validation successful (no timestamp changes)")
return True
# If timestamps changed, check file hashes (slower but accurate)
logger.info("Timestamps changed, checking file hashes...")
changed_files = []
for pdf_file in pdf_files:
current_hash = self._get_file_hash(pdf_file)
cached_hash = self.metadata.get("file_hashes", {}).get(pdf_file.name)
if current_hash != cached_hash:
changed_files.append(pdf_file.name)
if changed_files:
logger.info(f"Files changed: {', '.join(changed_files)}")
return False
logger.info("Cache validation successful (hashes match)")
return True
except Exception as e:
logger.error(f"Error validating cache: {e}")
return False
def load_documents(self) -> Optional[List[Any]]:
"""Load processed documents from cache"""
try:
if self.documents is None and self.documents_file.exists():
with open(self.documents_file, 'rb') as f:
self.documents = pickle.load(f)
logger.info(f"Loaded {len(self.documents)} documents from cache")
return self.documents
except Exception as e:
logger.error(f"Error loading documents: {e}")
return None
def load_embeddings(self) -> Optional[np.ndarray]:
"""Load embeddings from cache"""
try:
if self.embeddings is None and self.embeddings_file.exists():
self.embeddings = np.load(self.embeddings_file)
logger.info(f"Loaded embeddings with shape {self.embeddings.shape}")
return self.embeddings
except Exception as e:
logger.error(f"Error loading embeddings: {e}")
return None
def load_knowledge_base(self) -> Tuple[Optional[List[Any]], Optional[np.ndarray]]:
"""Load both documents and embeddings from cache"""
try:
documents = self.load_documents()
embeddings = self.load_embeddings()
if documents is not None and embeddings is not None:
logger.info(f"Loaded knowledge base: {len(documents)} documents, embeddings shape {embeddings.shape}")
return documents, embeddings
else:
logger.warning("Failed to load complete knowledge base from cache")
return None, None
except Exception as e:
logger.error(f"Error loading knowledge base: {e}")
return None, None
def is_valid(self) -> bool:
"""Check if cache has valid data"""
try:
return (self.documents_file.exists() and
self.embeddings_file.exists() and
self.metadata.get("chunk_count", 0) > 0)
except:
return False
def save_knowledge_base(self, documents: List[Any], embeddings: np.ndarray,
pdf_files: List[Path], embedder_config: Dict[str, Any]):
"""
Save processed documents and embeddings to cache
Args:
documents: List of processed document objects
embeddings: Numpy array of embeddings
pdf_files: List of source PDF files
embedder_config: Embedder configuration used
"""
try:
logger.info(f"Saving knowledge base: {len(documents)} documents, {embeddings.shape} embeddings")
# Save documents
with open(self.documents_file, 'wb') as f:
pickle.dump(documents, f)
# Save embeddings
np.save(self.embeddings_file, embeddings)
# Collect file metadata
file_hashes = {}
file_mtimes = {}
for pdf_file in pdf_files:
file_hashes[pdf_file.name] = self._get_file_hash(pdf_file)
file_mtimes[pdf_file.name] = pdf_file.stat().st_mtime
# Update metadata
self.metadata.update({
"document_count": len(pdf_files),
"chunk_count": len(documents),
"embedder_config": embedder_config,
"embedder_config_hash": create_embedder_config_hash(embedder_config),
"file_hashes": file_hashes,
"file_mtimes": file_mtimes
})
self._save_metadata()
# Cache in memory
self.documents = documents
self.embeddings = embeddings
logger.info("Knowledge base saved successfully")
except Exception as e:
logger.error(f"Error saving knowledge base: {e}")
raise
def get_cache_info(self) -> Dict[str, Any]:
"""Get information about cached data"""
return {
"cache_valid": self.documents_file.exists() and self.embeddings_file.exists(),
"document_count": self.metadata.get("document_count", 0),
"chunk_count": self.metadata.get("chunk_count", 0),
"last_updated": self.metadata.get("last_updated", 0),
"cache_size_mb": self._get_cache_size_mb(),
"embedder_config": self.metadata.get("embedder_config")
}
def _get_cache_size_mb(self) -> float:
"""Get total cache size in MB"""
try:
total_size = 0
for file_path in [self.metadata_file, self.documents_file, self.embeddings_file]:
if file_path.exists():
total_size += file_path.stat().st_size
return total_size / (1024 * 1024)
except:
return 0.0
def clear_cache(self):
"""Clear all cached data"""
try:
for file_path in [self.metadata_file, self.documents_file, self.embeddings_file, self.index_file]:
if file_path.exists():
file_path.unlink()
self.metadata = self._create_empty_metadata()
self.documents = None
self.embeddings = None
logger.info("Cache cleared successfully")
except Exception as e:
logger.error(f"Error clearing cache: {e}")
raise
def save_faiss_index(self, index_data: bytes):
"""Save FAISS index to cache"""
try:
with open(self.index_file, 'wb') as f:
f.write(index_data)
logger.info("FAISS index saved to cache")
except Exception as e:
logger.error(f"Error saving FAISS index: {e}")
def load_faiss_index(self) -> Optional[bytes]:
"""Load FAISS index from cache"""
try:
if self.index_file.exists():
with open(self.index_file, 'rb') as f:
return f.read()
return None
except Exception as e:
logger.error(f"Error loading FAISS index: {e}")
return None
def create_embedder_config_hash(system_or_config) -> Dict[str, Any]:
"""Extract embedder configuration for cache validation"""
try:
# Handle both system object and dict inputs
if isinstance(system_or_config, dict):
# Already a config dict, return as-is
return system_or_config
else:
# System object, extract config
embedder = system_or_config.get_component('embedder')
# Get key configuration parameters
config = {
"model_name": getattr(embedder, 'model_name', 'unknown'),
"model_type": type(embedder).__name__,
"device": getattr(embedder, 'device', 'unknown'),
"normalize_embeddings": getattr(embedder, 'normalize_embeddings', True)
}
# Add batch processor config if available
if hasattr(embedder, 'batch_processor'):
config["batch_size"] = getattr(embedder.batch_processor, 'batch_size', 32)
return config
except Exception as e:
logger.error(f"Error creating embedder config hash: {e}")
return {"error": str(e)} |