sebastien
/
Copies
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

minor modifs to grouping

master
Sébastien Miquel 2026-02-08 18:45:49 +01:00
parent e994655e58
commit b95e55d088
1 changed files with 4 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
grouping.py
View File

@ -11,26 +11,12 @@ from pdf2image import convert_from_path, pdfinfo_from_path
DPI = 200 # Good balance for readability and size DPI = 200 # Good balance for readability and size
A4_HEIGHT_INCHES = 11.69 A4_HEIGHT_INCHES = 11.69
FULL_PAGE_PX = int(A4_HEIGHT_INCHES * DPI) FULL_PAGE_PX = int(A4_HEIGHT_INCHES * DPI)
MAX_GROUP_HEIGHT = 2. * FULL_PAGE_PX MAX_GROUP_HEIGHT = 1.75 * FULL_PAGE_PX
MAX_GROUP_COUNT = 15 MAX_GROUP_COUNT = 8
SEPARATOR_HEIGHT = 20 SEPARATOR_HEIGHT = 20
LABEL_HEIGHT = 50 LABEL_HEIGHT = 50
MAX_FILE_SIZE_BYTES = 2.5 * 1024 * 1024 # 2MB MAX_FILE_SIZE_BYTES = 2.5 * 1024 * 1024 # 2MB
# def get_pdf_height(path):
# """Returns height in pixels at defined DPI without rendering."""
# try:
# info = pdfinfo_from_path(path)
# # info["Page size"] is usually "width height pts"
# # 1 pt = 1/72 inch
# # We assume single page PDFs as per prompt implication, or take the first page
# pts_height = float(info['Page size'].split(' ')[2]) if 'Page size' in info else 0
# return int((pts_height / 72.0) * DPI)
# except Exception as e:
# print(f"Error reading {path}: {e}")
# return 0
def get_pdf_height(path): def get_pdf_height(path):
"""Returns total height of all pages in pixels at defined DPI.""" """Returns total height of all pages in pixels at defined DPI."""
try: try:
@ -103,7 +89,7 @@ def group_files(file_list):
overhead = (SEPARATOR_HEIGHT + 30) if group['items'] else 0 overhead = (SEPARATOR_HEIGHT + 30) if group['items'] else 0
# Check Height Constraint # Check Height Constraint
if (group['current_height'] + height + overhead) <= MAX_GROUP_HEIGHT: if group['current_height'] + height + overhead <= MAX_GROUP_HEIGHT:
group['items'].append(item) group['items'].append(item)
group['current_height'] += height + overhead group['current_height'] += height + overhead
placed = True placed = True
@ -181,8 +167,6 @@ def create_jpg(identifier, group_index, group, root_dir):
try: try:
# Convert pdf to image # Convert pdf to image
imgs = convert_from_path(path, dpi=DPI) imgs = convert_from_path(path, dpi=DPI)
# if imgs:
# images.append((dd, imgs[0])) # Assume 1 page per pdf !! ??
if imgs: if imgs:
# Concatenate multi-page PDFs into one single image object # Concatenate multi-page PDFs into one single image object
combined_img = stitch_pdf_pages(imgs) combined_img = stitch_pdf_pages(imgs)
@ -256,7 +240,7 @@ def create_jpg(identifier, group_index, group, root_dir):
break break
quality -= 5 quality -= 5
print(f"Saved {output_path} ({os.path.getsize(output_path)/1024/1024:.2f} MB)") print(f"Saved {output_path} with {len(group)} ({os.path.getsize(output_path)/1024/1024:.2f} MB)")
def main(): def main():
if len(sys.argv) < 2: if len(sys.argv) < 2: