File size: 18,579 Bytes
2e1d581 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 |
"""
Fallback strategies for BackgroundFX Pro.
Implements robust fallback mechanisms when primary processing fails.
"""
import cv2
import numpy as np
import torch
from typing import Dict, List, Optional, Tuple, Any
from dataclasses import dataclass
from enum import Enum
import logging
import traceback
from ..utils.logger import setup_logger
from ..utils.device import DeviceManager
from ..utils.config import ConfigManager
from ..core.quality import QualityAnalyzer
logger = setup_logger(__name__)
class FallbackLevel(Enum):
"""Fallback hierarchy levels."""
NONE = 0
QUALITY_REDUCTION = 1
METHOD_SWITCH = 2
BASIC_PROCESSING = 3
MINIMAL_PROCESSING = 4
PASSTHROUGH = 5
@dataclass
class FallbackConfig:
"""Configuration for fallback strategies."""
max_retries: int = 3
quality_reduction_factor: float = 0.75
min_quality: float = 0.3
enable_caching: bool = True
cache_size: int = 10
timeout_seconds: float = 30.0
gpu_fallback_to_cpu: bool = True
progressive_downscale: bool = True
min_resolution: Tuple[int, int] = (320, 240)
class FallbackStrategy:
"""Intelligent fallback strategy manager."""
def __init__(self, config: Optional[FallbackConfig] = None):
self.config = config or FallbackConfig()
self.device_manager = DeviceManager()
self.quality_analyzer = QualityAnalyzer()
self.cache = {}
self.fallback_history = []
self.current_level = FallbackLevel.NONE
def execute_with_fallback(self, func, *args, **kwargs) -> Dict[str, Any]:
"""
Execute function with automatic fallback on failure.
Args:
func: Function to execute
*args: Function arguments
**kwargs: Function keyword arguments
Returns:
Result dictionary with status and output
"""
attempt = 0
last_error = None
original_args = args
original_kwargs = kwargs.copy()
while attempt < self.config.max_retries:
try:
# Log attempt
logger.info(f"Attempt {attempt + 1}/{self.config.max_retries} for {func.__name__}")
# Try execution
result = func(*args, **kwargs)
# Success - reset fallback level
self.current_level = FallbackLevel.NONE
return {
'success': True,
'result': result,
'attempts': attempt + 1,
'fallback_level': self.current_level
}
except Exception as e:
last_error = e
logger.warning(f"Attempt {attempt + 1} failed: {str(e)}")
# Apply fallback strategy
fallback_result = self._apply_fallback(
func, e, attempt,
original_args, original_kwargs
)
if fallback_result['handled']:
args = fallback_result.get('new_args', args)
kwargs = fallback_result.get('new_kwargs', kwargs)
else:
break
attempt += 1
# All attempts failed - apply final fallback
logger.error(f"All attempts failed for {func.__name__}")
return self._final_fallback(func, last_error, original_args)
def _apply_fallback(self, func, error: Exception,
attempt: int, original_args: tuple,
original_kwargs: dict) -> Dict[str, Any]:
"""Apply appropriate fallback strategy based on error type."""
error_type = type(error).__name__
self.fallback_history.append({
'function': func.__name__,
'error': error_type,
'attempt': attempt
})
# GPU memory error - switch to CPU
if 'CUDA' in str(error) or 'GPU' in str(error):
return self._handle_gpu_error(original_kwargs)
# Memory error - reduce quality
elif 'memory' in str(error).lower():
return self._handle_memory_error(original_args, original_kwargs)
# Timeout error - simplify processing
elif 'timeout' in str(error).lower():
return self._handle_timeout_error(original_kwargs)
# Model loading error - use simpler model
elif 'model' in str(error).lower():
return self._handle_model_error(original_kwargs)
# Generic error - progressive degradation
else:
return self._handle_generic_error(attempt, original_kwargs)
def _handle_gpu_error(self, kwargs: dict) -> Dict[str, Any]:
"""Handle GPU-related errors."""
logger.info("GPU error detected, falling back to CPU")
if self.config.gpu_fallback_to_cpu:
# Switch to CPU
self.device_manager.device = torch.device('cpu')
kwargs['device'] = 'cpu'
# Reduce batch size if present
if 'batch_size' in kwargs:
kwargs['batch_size'] = max(1, kwargs['batch_size'] // 2)
self.current_level = FallbackLevel.METHOD_SWITCH
return {
'handled': True,
'new_kwargs': kwargs
}
return {'handled': False}
def _handle_memory_error(self, args: tuple,
kwargs: dict) -> Dict[str, Any]:
"""Handle memory-related errors."""
logger.info("Memory error detected, reducing quality")
# Try to find image in args
image = None
image_idx = -1
for i, arg in enumerate(args):
if isinstance(arg, np.ndarray) and len(arg.shape) == 3:
image = arg
image_idx = i
break
if image is not None and self.config.progressive_downscale:
# Reduce image size
h, w = image.shape[:2]
new_h = int(h * self.config.quality_reduction_factor)
new_w = int(w * self.config.quality_reduction_factor)
# Ensure minimum resolution
new_h = max(new_h, self.config.min_resolution[1])
new_w = max(new_w, self.config.min_resolution[0])
if new_h < h or new_w < w:
resized = cv2.resize(image, (new_w, new_h))
args = list(args)
args[image_idx] = resized
self.current_level = FallbackLevel.QUALITY_REDUCTION
return {
'handled': True,
'new_args': tuple(args),
'new_kwargs': kwargs
}
# Reduce other memory-intensive parameters
if 'quality' in kwargs:
kwargs['quality'] = max(
self.config.min_quality,
kwargs['quality'] * self.config.quality_reduction_factor
)
return {
'handled': True,
'new_kwargs': kwargs
}
def _handle_timeout_error(self, kwargs: dict) -> Dict[str, Any]:
"""Handle timeout errors by simplifying processing."""
logger.info("Timeout detected, simplifying processing")
# Disable expensive operations
simplifications = {
'use_refinement': False,
'use_temporal': False,
'use_guided_filter': False,
'iterations': 1,
'num_samples': 1
}
for key, value in simplifications.items():
if key in kwargs:
kwargs[key] = value
self.current_level = FallbackLevel.BASIC_PROCESSING
return {
'handled': True,
'new_kwargs': kwargs
}
def _handle_model_error(self, kwargs: dict) -> Dict[str, Any]:
"""Handle model loading errors."""
logger.info("Model error detected, using simpler model")
# Switch to simpler model
if 'model_type' in kwargs:
model_hierarchy = ['large', 'base', 'small', 'tiny']
current = kwargs.get('model_type', 'base')
if current in model_hierarchy:
idx = model_hierarchy.index(current)
if idx < len(model_hierarchy) - 1:
kwargs['model_type'] = model_hierarchy[idx + 1]
self.current_level = FallbackLevel.METHOD_SWITCH
return {
'handled': True,
'new_kwargs': kwargs
}
# Disable model-based processing
kwargs['use_model'] = False
self.current_level = FallbackLevel.BASIC_PROCESSING
return {
'handled': True,
'new_kwargs': kwargs
}
def _handle_generic_error(self, attempt: int,
kwargs: dict) -> Dict[str, Any]:
"""Handle generic errors with progressive degradation."""
logger.info(f"Generic error, applying degradation level {attempt + 1}")
# Progressive degradation based on attempt
if attempt == 0:
# First attempt - minor quality reduction
self.current_level = FallbackLevel.QUALITY_REDUCTION
if 'quality' in kwargs:
kwargs['quality'] *= 0.8
elif attempt == 1:
# Second attempt - switch methods
self.current_level = FallbackLevel.METHOD_SWITCH
kwargs['method'] = 'basic'
else:
# Final attempt - minimal processing
self.current_level = FallbackLevel.MINIMAL_PROCESSING
kwargs['skip_refinement'] = True
kwargs['fast_mode'] = True
return {
'handled': True,
'new_kwargs': kwargs
}
def _final_fallback(self, func, error: Exception,
original_args: tuple) -> Dict[str, Any]:
"""Apply final fallback when all attempts fail."""
logger.error(f"Final fallback for {func.__name__}: {str(error)}")
self.current_level = FallbackLevel.PASSTHROUGH
# Try to return something useful
for arg in original_args:
if isinstance(arg, np.ndarray):
# Return original image/mask
return {
'success': False,
'result': arg,
'fallback_level': self.current_level,
'error': str(error)
}
# Return empty result
return {
'success': False,
'result': None,
'fallback_level': self.current_level,
'error': str(error)
}
class ProcessingFallback:
"""Specific fallback implementations for processing operations."""
def __init__(self):
self.logger = setup_logger(f"{__name__}.ProcessingFallback")
self.quality_analyzer = QualityAnalyzer()
def basic_segmentation(self, image: np.ndarray) -> np.ndarray:
"""
Basic segmentation using traditional CV methods.
Used as fallback when ML models fail.
Args:
image: Input image
Returns:
Binary mask
"""
try:
# Convert to grayscale
if len(image.shape) == 3:
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
else:
gray = image
# Apply GrabCut for basic foreground extraction
mask = np.zeros(gray.shape[:2], np.uint8)
bgd_model = np.zeros((1, 65), np.float64)
fgd_model = np.zeros((1, 65), np.float64)
# Initialize rectangle (center 80% of image)
h, w = gray.shape[:2]
rect = (int(w * 0.1), int(h * 0.1),
int(w * 0.8), int(h * 0.8))
# Apply GrabCut
cv2.grabCut(image, mask, rect, bgd_model, fgd_model,
5, cv2.GC_INIT_WITH_RECT)
# Extract foreground
mask2 = np.where((mask == 2) | (mask == 0), 0, 255).astype('uint8')
return mask2
except Exception as e:
self.logger.error(f"Basic segmentation failed: {e}")
# Return center blob as last resort
return self._center_blob_mask(image.shape[:2])
def _center_blob_mask(self, shape: Tuple[int, int]) -> np.ndarray:
"""Create a center ellipse mask as ultimate fallback."""
h, w = shape
mask = np.zeros((h, w), dtype=np.uint8)
# Create center ellipse
center = (w // 2, h // 2)
axes = (w // 3, h // 3)
cv2.ellipse(mask, center, axes, 0, 0, 360, 255, -1)
# Smooth edges
mask = cv2.GaussianBlur(mask, (21, 21), 10)
_, mask = cv2.threshold(mask, 127, 255, cv2.THRESH_BINARY)
return mask
def basic_matting(self, image: np.ndarray,
mask: np.ndarray) -> np.ndarray:
"""
Basic matting using morphological operations.
Args:
image: Input image
mask: Binary mask
Returns:
Alpha matte
"""
try:
# Ensure uint8
if mask.dtype != np.uint8:
mask = (mask * 255).astype(np.uint8)
# Morphological smoothing
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5, 5))
mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, kernel)
mask = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel)
# Edge softening
mask = cv2.GaussianBlur(mask, (5, 5), 2)
# Normalize to [0, 1]
alpha = mask.astype(np.float32) / 255.0
return alpha
except Exception as e:
self.logger.error(f"Basic matting failed: {e}")
return mask.astype(np.float32) / 255.0
def color_difference_keying(self, image: np.ndarray,
key_color: Optional[np.ndarray] = None,
threshold: float = 30) -> np.ndarray:
"""
Simple color difference keying for solid backgrounds.
Args:
image: Input image
key_color: Background color to remove
threshold: Color difference threshold
Returns:
Alpha matte
"""
try:
if key_color is None:
# Estimate background color from corners
h, w = image.shape[:2]
corners = [
image[0:10, 0:10],
image[0:10, w-10:w],
image[h-10:h, 0:10],
image[h-10:h, w-10:w]
]
key_color = np.mean([np.mean(c, axis=(0, 1)) for c in corners], axis=0)
# Calculate color difference
diff = np.sqrt(np.sum((image - key_color) ** 2, axis=2))
# Create mask
mask = (diff > threshold).astype(np.float32)
# Smooth edges
mask = cv2.GaussianBlur(mask, (5, 5), 2)
return mask
except Exception as e:
self.logger.error(f"Color keying failed: {e}")
return np.ones(image.shape[:2], dtype=np.float32)
def edge_based_segmentation(self, image: np.ndarray) -> np.ndarray:
"""
Edge-based segmentation as fallback.
Args:
image: Input image
Returns:
Binary mask
"""
try:
# Convert to grayscale
if len(image.shape) == 3:
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
else:
gray = image
# Edge detection
edges = cv2.Canny(gray, 50, 150)
# Close contours
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5))
closed = cv2.morphologyEx(edges, cv2.MORPH_CLOSE, kernel, iterations=2)
# Find contours
contours, _ = cv2.findContours(
closed, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE
)
# Create mask from largest contour
mask = np.zeros(gray.shape, dtype=np.uint8)
if contours:
largest = max(contours, key=cv2.contourArea)
cv2.drawContours(mask, [largest], -1, 255, -1)
return mask
except Exception as e:
self.logger.error(f"Edge segmentation failed: {e}")
return self._center_blob_mask(image.shape[:2])
def cached_result(self, cache_key: str,
fallback_func, *args, **kwargs) -> Any:
"""
Try to retrieve cached result or compute with fallback.
Args:
cache_key: Cache identifier
fallback_func: Function to call if not cached
*args, **kwargs: Function arguments
Returns:
Cached or computed result
"""
# Simple in-memory cache implementation
if not hasattr(self, '_cache'):
self._cache = {}
if cache_key in self._cache:
self.logger.info(f"Using cached result for {cache_key}")
return self._cache[cache_key]
try:
result = fallback_func(*args, **kwargs)
self._cache[cache_key] = result
# Limit cache size
if len(self._cache) > 100:
# Remove oldest entries
keys = list(self._cache.keys())
for key in keys[:20]:
del self._cache[key]
return result
except Exception as e:
self.logger.error(f"Cached computation failed: {e}")
return None |