RunasSudo/pdf-segmented
1
Fork 0
Code Issues Pull Requests Activity

Improve performance of reading XCF metadata

Browse Source
This commit is contained in:
RunasSudo 2025-05-08 21:24:21 +10:00
parent d4b9c8f4bd
commit 7ae4b0f558
Signed by: RunasSudo
GPG Key ID: 7234E476BF21C61A
1 changed files with 18 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

50
pdf_segmented/input/xcf.py
View File

@ -35,45 +35,31 @@ def xcf_get_pages(input_file: str) -> InputPages:
height = None height = None
# Read metadata # Read metadata
proc = subprocess.run(['magick', 'identify', '-verbose', input_file], capture_output=True, encoding='utf-8', check=True) proc = subprocess.run(['magick', 'identify', '-format', '%n;%W;%H;%x;%y;%U|', input_file], capture_output=True, encoding='utf-8', check=True)
for line in proc.stdout.splitlines(): magick_layers_data = proc.stdout.strip('|').split('|')
if line == 'Image:':
num_layers += 1
elif line.startswith(' Geometry: '):
layer_width = float(line[len(' Geometry: '):line.index('x')])
layer_height = float(line[line.index('x')+1:line.index('+')])
if (width is not None and layer_width != width) or (height is not None and layer_height != height):
print('Error: Image with variable-dimension layers is not supported ({}x{} vs {}x{})'.format(layer_width, layer_height, width, height))
sys.exit(1)
width = layer_width
height = layer_height
elif line.startswith(' Resolution: '):
resolution_x = float(line[len(' Resolution: '):line.index('x')])
resolution_y = float(line[line.index('x')+1:])
if resolution_x != resolution_y:
raise Exception('Unexpected non-square DPI ({}x{})'.format(resolution_x, resolution_y))
if dpi is not None and resolution_x != dpi:
raise Exception('Unexpected variable DPI image ({} vs {})'.format(resolution_x, dpi))
dpi = resolution_x
elif line.startswith(' Units: '):
if line != ' Units: PixelsPerInch':
raise Exception('Unexpected Units (expected PixelsPerInch, got {})'.format(line[len(' Units: '):]))
if num_layers == 0: if any(d != magick_layers_data[0] for d in magick_layers_data[1:]):
raise Exception('Unexpected 0 layers') raise Exception('Different metadata between XCF layers')
if dpi is None:
raise Exception('Unexpected no DPI information') magick_layer_data = magick_layers_data[0].split(';')
if width is None: num_layers = int(magick_layer_data[0])
raise Exception('Unexpected no width information') width = int(magick_layer_data[1])
if height is None: height = int(magick_layer_data[2])
raise Exception('Unexpected no height information') resolution_x = float(magick_layer_data[3])
resolution_y = float(magick_layer_data[4])
resolution_units = magick_layer_data[5]
if resolution_units != 'PixelsPerInch':
raise Exception('Unexpected resolution units (expected PixelsPerInch, got {})'.format(resolution_units))
if resolution_x != resolution_y:
raise Exception('Unexpected non-square DPI ({}x{})'.format(resolution_x, resolution_y))
return InputPages( return InputPages(
file_name=input_file, file_name=input_file,
num_pages=num_layers, num_pages=num_layers,
width=width, width=width,
height=height, height=height,
dpi=dpi, dpi=resolution_x,
pages=_do_get_pages(input_file, num_layers) pages=_do_get_pages(input_file, num_layers)
) )