Pixcribe / batch_processing_manager.py
DawnC's picture
Upload 5 files
f3a4ad9 verified
raw
history blame
15.9 kB
import time
import json
import csv
import zipfile
from io import BytesIO
from typing import List, Dict, Optional, Callable
from PIL import Image
import traceback
class BatchProcessingManager:
"""
Manages batch processing of multiple images with progress tracking,
error handling, and result export functionality.
Follows the Facade pattern by delegating actual image processing
to the PixcribePipeline instance.
"""
def __init__(self, pipeline=None):
"""
Initialize the Batch Processing Manager.
Args:
pipeline: Reference to PixcribePipeline instance for processing images
"""
self.pipeline = pipeline
self.results = {} # Store processing results indexed by image number
self.timing_data = [] # Track processing time for each image
def process_batch(
self,
images: List[Image.Image],
platform: str = 'instagram',
yolo_variant: str = 'l',
language: str = 'zh',
progress_callback: Optional[Callable] = None
) -> Dict:
"""
Process a batch of images with progress tracking.
Args:
images: List of PIL Image objects to process (max 10)
platform: Target social media platform
yolo_variant: YOLO model variant ('m', 'l', 'x')
language: Caption language ('zh', 'en')
progress_callback: Optional callback function for progress updates
Returns:
Dictionary containing batch processing summary and results
Raises:
ValueError: If images list is empty or exceeds 10 images
"""
# Validate input
if not images:
raise ValueError("Images list cannot be empty")
if len(images) > 10:
raise ValueError("Maximum 10 images allowed per batch")
# Initialize results storage
self.results = {}
self.timing_data = []
total_images = len(images)
# Record batch start time
batch_start_time = time.time()
print(f"\n{'='*60}")
print(f"Starting batch processing: {total_images} images")
print(f"Platform: {platform} | Variant: {yolo_variant} | Language: {language}")
print(f"{'='*60}\n")
# Process each image
for idx, image in enumerate(images):
image_start_time = time.time()
image_index = idx + 1
try:
print(f"[{image_index}/{total_images}] Processing image {image_index}...")
# Call pipeline's process_image method
result = self.pipeline.process_image(
image=image,
platform=platform,
yolo_variant=yolo_variant,
language=language
)
# Store successful result
self.results[image_index] = {
'status': 'success',
'result': result,
'image_index': image_index,
'error': None
}
print(f"βœ“ Image {image_index} processed successfully")
except Exception as e:
# Store error result
error_trace = traceback.format_exc()
self.results[image_index] = {
'status': 'failed',
'result': None,
'image_index': image_index,
'error': {
'type': type(e).__name__,
'message': str(e),
'traceback': error_trace
}
}
print(f"βœ— Image {image_index} failed: {str(e)}")
# Record processing time for this image
image_elapsed = time.time() - image_start_time
self.timing_data.append(image_elapsed)
# Calculate progress information
completed = image_index
percent = (completed / total_images) * 100
# Estimate remaining time based on average processing time
avg_time = sum(self.timing_data) / len(self.timing_data)
remaining_images = total_images - completed
estimated_remaining = avg_time * remaining_images
# Call progress callback if provided
if progress_callback:
progress_info = {
'current': completed,
'total': total_images,
'percent': percent,
'estimated_remaining': estimated_remaining,
'latest_result': self.results[image_index],
'image_index': image_index
}
progress_callback(progress_info)
# Calculate batch summary
batch_elapsed = time.time() - batch_start_time
total_processed = len(self.results)
total_failed = sum(1 for r in self.results.values() if r['status'] == 'failed')
total_success = total_processed - total_failed
print(f"\n{'='*60}")
print(f"Batch processing completed!")
print(f"Total: {total_processed} | Success: {total_success} | Failed: {total_failed}")
print(f"Total time: {batch_elapsed:.2f}s | Avg per image: {batch_elapsed/total_processed:.2f}s")
print(f"{'='*60}\n")
# Return batch summary
return {
'results': self.results,
'total_processed': total_processed,
'total_success': total_success,
'total_failed': total_failed,
'total_time': batch_elapsed,
'average_time_per_image': batch_elapsed / total_processed if total_processed > 0 else 0
}
def get_result(self, image_index: int) -> Optional[Dict]:
"""
Get processing result for a specific image.
Args:
image_index: Index of the image (1-based)
Returns:
Result dictionary or None if index doesn't exist
"""
return self.results.get(image_index)
def get_all_results(self) -> Dict:
"""
Get all processing results.
Returns:
Complete results dictionary
"""
return self.results
def clear_results(self):
"""Clear all stored results to free memory."""
self.results = {}
self.timing_data = []
print("βœ“ Batch results cleared")
def export_to_json(self, results: Dict, output_path: str) -> str:
"""
Export batch results to JSON format.
Args:
results: Results dictionary from process_batch
output_path: Path to save JSON file
Returns:
Path to the saved JSON file
"""
# Prepare export data
export_data = {
'batch_summary': {
'total_processed': results.get('total_processed', 0),
'total_success': results.get('total_success', 0),
'total_failed': results.get('total_failed', 0),
'total_time': results.get('total_time', 0),
'average_time_per_image': results.get('average_time_per_image', 0)
},
'images': []
}
# Process each image result
for img_idx, img_result in results.get('results', {}).items():
if img_result['status'] == 'success':
result_data = img_result['result']
image_export = {
'image_index': img_idx,
'status': 'success',
'captions': result_data.get('captions', []),
'detected_objects': [
det['class_name'] for det in result_data.get('detections', [])
],
'detected_brands': [
brand[0] if isinstance(brand, tuple) else brand
for brand in result_data.get('brands', [])
],
'scene_info': result_data.get('scene', {}),
'lighting': result_data.get('lighting', {})
}
else:
image_export = {
'image_index': img_idx,
'status': 'failed',
'error': img_result.get('error', {})
}
export_data['images'].append(image_export)
# Write to JSON file
with open(output_path, 'w', encoding='utf-8') as f:
json.dump(export_data, f, ensure_ascii=False, indent=2)
print(f"βœ“ Batch results exported to JSON: {output_path}")
return output_path
def export_to_csv(self, results: Dict, output_path: str) -> str:
"""
Export batch results to CSV format.
Args:
results: Results dictionary from process_batch
output_path: Path to save CSV file
Returns:
Path to the saved CSV file
"""
# Define CSV headers
headers = [
'image_index',
'status',
'caption_professional',
'caption_creative',
'caption_authentic',
'detected_objects',
'detected_brands',
'hashtags'
]
# Prepare rows
rows = []
for img_idx, img_result in results.get('results', {}).items():
if img_result['status'] == 'success':
result_data = img_result['result']
captions = result_data.get('captions', [])
# Extract captions by tone
caption_professional = ''
caption_creative = ''
caption_authentic = ''
all_hashtags = []
for cap in captions:
tone = cap.get('tone', '').lower()
caption_text = cap.get('caption', '')
hashtags = cap.get('hashtags', [])
if 'professional' in tone:
caption_professional = caption_text
elif 'creative' in tone:
caption_creative = caption_text
elif 'authentic' in tone or 'casual' in tone:
caption_authentic = caption_text
all_hashtags.extend(hashtags)
# Remove duplicates from hashtags
all_hashtags = list(set(all_hashtags))
row = {
'image_index': img_idx,
'status': 'success',
'caption_professional': caption_professional,
'caption_creative': caption_creative,
'caption_authentic': caption_authentic,
'detected_objects': ', '.join([
det['class_name'] for det in result_data.get('detections', [])
]),
'detected_brands': ', '.join([
brand[0] if isinstance(brand, tuple) else brand
for brand in result_data.get('brands', [])
]),
'hashtags': ' '.join([f'#{tag}' for tag in all_hashtags])
}
else:
row = {
'image_index': img_idx,
'status': 'failed',
'caption_professional': '',
'caption_creative': '',
'caption_authentic': '',
'detected_objects': '',
'detected_brands': '',
'hashtags': ''
}
rows.append(row)
# Write to CSV file
with open(output_path, 'w', newline='', encoding='utf-8') as f:
writer = csv.DictWriter(f, fieldnames=headers)
writer.writeheader()
writer.writerows(rows)
print(f"βœ“ Batch results exported to CSV: {output_path}")
return output_path
def export_to_zip(self, results: Dict, images: List[Image.Image], output_path: str) -> str:
"""
Export batch results to ZIP archive with images and text files.
Args:
results: Results dictionary from process_batch
images: List of original PIL Image objects
output_path: Path to save ZIP file
Returns:
Path to the saved ZIP file
"""
with zipfile.ZipFile(output_path, 'w', zipfile.ZIP_DEFLATED) as zipf:
for img_idx, img_result in results.get('results', {}).items():
if img_result['status'] == 'success':
# Save original image
image_filename = f"image_{img_idx:03d}.jpg"
# Convert PIL image to bytes
img_buffer = BytesIO()
images[img_idx - 1].save(img_buffer, format='JPEG', quality=95)
img_buffer.seek(0)
zipf.writestr(image_filename, img_buffer.read())
# Save caption text file
text_filename = f"image_{img_idx:03d}.txt"
text_content = self._format_result_as_text(img_result['result'])
zipf.writestr(text_filename, text_content)
print(f"βœ“ Added to ZIP: {image_filename} and {text_filename}")
print(f"βœ“ Batch results exported to ZIP: {output_path}")
return output_path
def _format_result_as_text(self, result: Dict) -> str:
"""
Format a single image result as plain text for ZIP export.
Args:
result: Single image processing result dictionary
Returns:
Formatted text string
"""
lines = []
lines.append("=" * 60)
lines.append("PIXCRIBE - AI GENERATED SOCIAL MEDIA CONTENT")
lines.append("=" * 60)
lines.append("")
# Captions section
captions = result.get('captions', [])
for i, cap in enumerate(captions, 1):
tone = cap.get('tone', 'Unknown').upper()
caption_text = cap.get('caption', '')
hashtags = cap.get('hashtags', [])
lines.append(f"CAPTION {i} - {tone} STYLE")
lines.append("-" * 60)
lines.append(caption_text)
lines.append("")
lines.append("Hashtags:")
lines.append(' '.join([f'#{tag}' for tag in hashtags]))
lines.append("")
lines.append("")
# Detected objects section
detections = result.get('detections', [])
if detections:
lines.append("DETECTED OBJECTS")
lines.append("-" * 60)
object_names = [det['class_name'] for det in detections]
lines.append(', '.join(object_names))
lines.append("")
# Detected brands section
brands = result.get('brands', [])
if brands:
lines.append("DETECTED BRANDS")
lines.append("-" * 60)
brand_names = [
brand[0] if isinstance(brand, tuple) else brand
for brand in brands
]
lines.append(', '.join(brand_names))
lines.append("")
# Scene information
scene_info = result.get('scene', {})
if scene_info:
lines.append("SCENE ANALYSIS")
lines.append("-" * 60)
if 'lighting' in scene_info:
lighting = scene_info['lighting'].get('top', 'Unknown')
lines.append(f"Lighting: {lighting}")
if 'mood' in scene_info:
mood = scene_info['mood'].get('top', 'Unknown')
lines.append(f"Mood: {mood}")
lines.append("")
lines.append("=" * 60)
lines.append("Generated by Pixcribe V5 - AI Social Media Caption Generator")
lines.append("=" * 60)
return '\n'.join(lines)
print("βœ“ BatchProcessingManager defined")