image/remove_metadata.py

"""
Image Metadata Remover Module

This module provides a comprehensive class-based metadata remover that strips ALL metadata
from images while preserving original quality, format, and aspect ratio.
Handles EXIF, IPTC, XMP, ICC profiles, and all other embedded metadata for security/privacy.
"""

from .image_base import ImageBase
import os
from custom_logger import logger_config
from PIL import Image
from PIL.ExifTags import TAGS, GPSTAGS
from typing import Dict
import base64

class RemoveMetadata(ImageBase):
	"""
	Comprehensive image metadata remover for security and privacy.
	
	Removes ALL metadata types including EXIF, IPTC, XMP, ICC profiles, thumbnails,
	GPS data, camera settings, and any other embedded information.
	"""
	def __init__(self):
		super().__init__("remove_metadata")
	
	def _extract_all_metadata(self, image: Image.Image):
		"""
		Extract ALL available metadata from the image for security analysis.
		
		Args:
			image: PIL Image object
			
		Returns:
			Dictionary containing comprehensive metadata extraction
		"""
		all_metadata = {}
		
		# 1. Extract EXIF data (most common and detailed)
		try:
			if hasattr(image, '_getexif') and image._getexif():
				exif_dict = image._getexif()
				all_metadata['exif_raw'] = exif_dict
				
				# Decode EXIF tags to human-readable format
				exif_decoded = {}
				for tag_id, value in exif_dict.items():
					tag = TAGS.get(tag_id, tag_id)
					
					# Special handling for GPS data
					if tag == 'GPSInfo':
						gps_data = {}
						for gps_tag_id, gps_value in value.items():
							gps_tag = GPSTAGS.get(gps_tag_id, gps_tag_id)
							gps_data[gps_tag] = gps_value
						exif_decoded[tag] = gps_data
					else:
						exif_decoded[tag] = value
				
				all_metadata['exif_decoded'] = exif_decoded
		except Exception as e:
			all_metadata['exif_error'] = str(e)
		
		# 2. Extract all image.info metadata (includes IPTC, XMP, ICC, etc.)
		if hasattr(image, 'info') and image.info:
			info_data = {}
			for key, value in image.info.items():
				# Handle binary data by encoding it
				if isinstance(value, bytes):
					try:
						# Try to decode as text first
						info_data[key] = value.decode('utf-8', errors='ignore')
					except:
						# If that fails, encode as base64 for viewing
						info_data[key] = f"<binary_data_base64>{base64.b64encode(value[:100]).decode()}</binary_data_base64>"
				else:
					info_data[key] = value
			all_metadata['info'] = info_data
		
		# 3. Extract ICC Profile (color management)
		try:
			if hasattr(image, 'info') and 'icc_profile' in image.info:
				icc_profile = image.info['icc_profile']
				all_metadata['icc_profile_size'] = len(icc_profile) if icc_profile else 0
				all_metadata['icc_profile_present'] = bool(icc_profile)
		except Exception as e:
			all_metadata['icc_error'] = str(e)
		
		# 4. Extract XMP metadata
		try:
			if hasattr(image, 'info') and 'xmp' in image.info:
				xmp_data = image.info['xmp']
				if isinstance(xmp_data, bytes):
					all_metadata['xmp'] = xmp_data.decode('utf-8', errors='ignore')
				else:
					all_metadata['xmp'] = xmp_data
		except Exception as e:
			all_metadata['xmp_error'] = str(e)
		
		# 5. Extract IPTC metadata
		try:
			if hasattr(image, 'info') and any(key.startswith('iptc') for key in image.info.keys()):
				iptc_data = {k: v for k, v in image.info.items() if k.startswith('iptc')}
				all_metadata['iptc'] = iptc_data
		except Exception as e:
			all_metadata['iptc_error'] = str(e)
		
		# 6. Extract PIL tag and tag_v2 (alternative EXIF access)
		try:
			if hasattr(image, 'tag') and image.tag:
				all_metadata['pil_tag'] = dict(image.tag)
		except Exception as e:
			all_metadata['pil_tag_error'] = str(e)
			
		try:
			if hasattr(image, 'tag_v2') and image.tag_v2:
				all_metadata['pil_tag_v2'] = dict(image.tag_v2)
		except Exception as e:
			all_metadata['pil_tag_v2_error'] = str(e)
		
		# 7. Extract quantization tables (JPEG compression info)
		try:
			if hasattr(image, 'quantization') and image.quantization:
				all_metadata['quantization_tables'] = len(image.quantization)
		except Exception as e:
			all_metadata['quantization_error'] = str(e)
		
		# 8. Extract embedded thumbnails
		try:
			if hasattr(image, 'info') and 'thumbnail' in image.info:
				all_metadata['thumbnail_present'] = True
		except:
			pass
		
		# 9. Extract text chunks (PNG specific)
		try:
			if image.format == 'PNG' and hasattr(image, 'text'):
				all_metadata['png_text_chunks'] = dict(image.text)
		except Exception as e:
			all_metadata['png_text_error'] = str(e)
		
		# 10. Extract basic image properties that might contain metadata
		basic_props = {
			'format': getattr(image, 'format', None),
			'mode': getattr(image, 'mode', None),
			'size': getattr(image, 'size', None),
			'filename': getattr(image, 'filename', None),
			'format_description': getattr(image, 'format_description', None)
		}
		all_metadata['basic_properties'] = basic_props
		
		return all_metadata
	
	def _get_quality_settings_for_format(self, format_name: str, original_mode: str) -> Dict:
		"""
		Get optimal save settings for maximum quality preservation per format.
		ALL metadata will be stripped regardless of format.
		
		Args:
			format_name: PIL format name
			original_mode: Original image color mode
			
		Returns:
			Dictionary of save parameters
		"""
		settings = {}
		
		if format_name == 'JPEG':
			settings = {
				'quality': 100,		 # Maximum quality
				'optimize': False,	  # Don't optimize to preserve quality
				'progressive': False,   # Standard baseline JPEG
				'subsampling': 0,	   # No chroma subsampling
				'exif': b'',		   # Explicitly remove EXIF
				'icc_profile': None	# Remove ICC profile
			}
			
		elif format_name == 'PNG':
			settings = {
				'optimize': False,	  # Don't optimize to preserve quality
				'compress_level': 1,	# Minimal compression
				'icc_profile': None,	# Remove ICC profile
				'pnginfo': None		 # Remove PNG info chunks
			}
			
		elif format_name == 'WEBP':
			settings = {
				'lossless': True,	   # Lossless compression
				'quality': 100,		 # Maximum quality
				'method': 6,			# Best compression method
				'icc_profile': None,	# Remove ICC profile
				'exif': b''			 # Remove EXIF
			}
			
		elif format_name == 'TIFF':
			settings = {
				'compression': None,	# No compression
				'icc_profile': None	 # Remove ICC profile
			}
			
		elif format_name in ['BMP', 'GIF']:
			settings = {}  # These formats have limited metadata anyway
			
		return settings
	
	def _clean_png_with_transparency(self, image: Image.Image, output_path: str) -> bool:
		"""
		Clean PNG image while preserving transparency and quality, removing ALL metadata.
		
		Args:
			image: PIL Image object
			output_path: Path for output file
			
		Returns:
			True if successful
		"""
		try:
			original_mode = image.mode
			
			# Create completely new clean image preserving only visual data
			if original_mode == 'P':
				# Palette mode - preserve only palette and pixel data
				clean_img = Image.new(original_mode, image.size)
				if image.getpalette():
					clean_img.putpalette(image.getpalette())
				clean_img.paste(image, (0, 0))
			else:
				# RGBA or LA mode - preserve only pixel data
				clean_img = Image.new(original_mode, image.size, (0, 0, 0, 0))
				clean_img.paste(image, (0, 0))
			
			# Ensure no metadata is carried over
			clean_img.info = {}
			
			# Save with quality preservation and NO metadata
			save_settings = self._get_quality_settings_for_format('PNG', original_mode)
			clean_img.save(output_path, format='PNG', **save_settings)
			
			return True
			
		except Exception as e:
			raise Exception(f"PNG transparency cleaning failed: {e}")
	
	def _clean_jpeg_image(self, image: Image.Image, output_path: str) -> bool:
		"""
		Clean JPEG image while preserving maximum quality, removing ALL metadata.
		
		Args:
			image: PIL Image object
			output_path: Path for output file
			
		Returns:
			True if successful
		"""
		try:
			# Convert to RGB if necessary and create clean copy
			if image.mode != 'RGB':
				clean_img = image.convert('RGB')
			else:
				# Create new image to ensure no metadata transfer
				clean_img = Image.new('RGB', image.size)
				clean_img.paste(image, (0, 0))
			
			# Ensure no metadata is carried over
			clean_img.info = {}
			
			# Save with maximum quality and absolutely NO metadata
			save_settings = self._get_quality_settings_for_format('JPEG', image.mode)
			clean_img.save(output_path, format='JPEG', **save_settings)
			
			return True
			
		except Exception as e:
			raise Exception(f"JPEG cleaning failed: {e}")
	
	def _clean_other_format(self, image: Image.Image, output_path: str, original_format: str) -> bool:
		"""
		Clean other image formats while preserving quality, removing ALL metadata.
		
		Args:
			image: PIL Image object
			output_path: Path for output file
			original_format: Original image format
			
		Returns:
			True if successful
		"""
		try:
			# Create completely new image without any metadata transfer
			clean_img = Image.new(image.mode, image.size)
			
			# Handle palette images (preserve only palette, not metadata)
			if image.mode == 'P' and image.getpalette():
				clean_img.putpalette(image.getpalette())
			
			# Paste only pixel data
			clean_img.paste(image, (0, 0))

			# Ensure absolutely no metadata is carried over
			clean_img.info = {}
			
			# Save with format-specific quality settings and no metadata
			save_settings = self._get_quality_settings_for_format(original_format, image.mode)
			clean_img.save(output_path, format=original_format, **save_settings)
			
			return True
			
		except Exception as e:
			raise Exception(f"{original_format} cleaning failed: {e}")
	
	def _verify_metadata_removal(self, output_path) -> bool:
		"""
		Comprehensive verification that ALL metadata has been removed.
		Returns:
			True if ALL metadata was successfully removed
		"""
		try:
			with Image.open(output_path) as img:
				remaining_metadata = self._extract_all_metadata(img)
				
				# Check for any remaining metadata beyond basic properties
				sensitive_keys = ['exif_raw', 'exif_decoded', 'info', 'icc_profile_present', 
								'xmp', 'iptc', 'pil_tag', 'pil_tag_v2', 'quantization_tables',
								'thumbnail_present', 'png_text_chunks']
				
				remaining_sensitive = {k: v for k, v in remaining_metadata.items() 
									 if k in sensitive_keys and v}
				
				if remaining_sensitive:
					logger_config.warning(f"WARNING: Sensitive metadata still present: {list(remaining_sensitive.keys())}")
					logger_config.debug(f"Remaining metadata: {remaining_sensitive}")
					return False
				else:
					logger_config.success("ALL metadata successfully removed - image is clean")
					return True
					
		except Exception as e:
			raise Exception(f"Verification failed: {e}")
	
	def _verify_image_quality(self, original_path: str, cleaned_path: str) -> bool:
		"""
		Verify that image quality and dimensions are preserved.
		
		Args:
			original_path: Path to original image
			cleaned_path: Path to cleaned image
			
		Returns:
			True if quality is preserved
		"""
		try:
			with Image.open(original_path) as original, Image.open(cleaned_path) as cleaned:
				# Check dimensions
				if original.size != cleaned.size:
					raise Exception(f"Dimension mismatch! Original: {original.size}, Cleaned: {cleaned.size}")
				
				# Check if file exists and has reasonable size
				if not os.path.exists(cleaned_path):
					return False
				
				cleaned_size = os.path.getsize(cleaned_path)
				if cleaned_size == 0:
					raise Exception("Cleaned file is empty")
				
				logger_config.success(f"Quality preserved - Size: {cleaned.size}, File size: {cleaned_size:,} bytes")
				return True
				
		except Exception as e:
			raise Exception(f"Quality verification failed: {e}")
	
	def process(self, input_file_name: str):
		"""
		Remove ALL metadata from an image for complete security and privacy protection.
		
		This method removes:
		- EXIF data (camera settings, GPS coordinates, timestamps)
		- IPTC data (keywords, captions, copyright)
		- XMP data (Adobe metadata)
		- ICC color profiles
		- Embedded thumbnails
		- Quantization tables
		- PNG text chunks
		- Any other embedded metadata
		
		Args:
			input_file_name: Input image name
			
		Returns:
			Tuple containing (output_path, extracted_metadata)
			
		Raises:
			FileNotFoundError: If input file doesn't exist
			ValueError: If file format is not supported
			Exception: If metadata removal fails
		"""
		try:
			self.input_file_name = input_file_name
			self.input_file_path = f'{self.input_dir}/{self.input_file_name}'
			# Validate input
			self._validate_input_file()
			
			# Generate output path
			output_path = self._generate_output_path()

			logger_config.info(f"Processing: {self.input_file_path}")
			
			# Open and analyze image
			with Image.open(self.input_file_path) as image:
				original_format = image.format
				original_mode = image.mode
				
				# Extract ALL metadata for security analysis
				extracted_metadata = self._extract_all_metadata(image)
				
				# Print comprehensive metadata analysis
				logger_config.info("=== COMPREHENSIVE METADATA ANALYSIS ===")
				if extracted_metadata:
					for key, value in extracted_metadata.items():
						if value:  # Only show non-empty metadata
							logger_config.info(f"{key}: {value}")
				else:
					logger_config.info("No metadata found in original image")
				logger_config.info("=== END METADATA ANALYSIS ===")
				
				# Choose appropriate cleaning method based on format
				success = False
				
				if original_format == 'PNG' and original_mode in ('RGBA', 'LA', 'P'):
					success = self._clean_png_with_transparency(image, output_path)
					
				elif original_format == 'JPEG' or output_path.lower().endswith(('.jpg', '.jpeg')):
					success = self._clean_jpeg_image(image, output_path)
					
				else:
					success = self._clean_other_format(image, output_path, original_format)
				
				if not success:
					raise Exception("Metadata cleaning failed")
			
			# Verify results
			self._verify_image_quality(self.input_file_path, output_path)
			
			# Comprehensive metadata removal verification
			self._verify_metadata_removal(output_path)
			
			logger_config.success(f"Cleaned image saved: {output_path}")
			logger_config.success("ALL METADATA REMOVED - Image is secure for sharing")
			return output_path, extracted_metadata
			
		except Exception as e:
			logger_config.error(f"Failed to clean image: {str(e)}")
			return None, {}

# Example usage
if __name__ == "__main__":
	cleaner = RemoveMetadata()
	cleaned_file, metadata = cleaner.process("image/input/test.png")
	print(f"Output: {cleaned_file}")
	print(f"Extracted metadata: {metadata}")
	print("Image is now secure for sharing - all metadata removed!")