vheliotech-guide-de-montage/sphinx-tools/make_pdf.py

import sys
import subprocess
import re
import os

source_dir = sys.argv[1]
build_dir = sys.argv[2]
insert_toc_after_page = 1
max_bookmark_level = 3

index_pdf_filename = build_dir + '/weasyprint/index.pdf'
css_filename = source_dir + '/css/print-theme.css'

script_dir = os.path.dirname(__file__)

# Compile PDF
assert(subprocess.run(['weasyprint', build_dir + '/weasyprint/index.html', index_pdf_filename, '-s', css_filename]).returncode == 0)

# Generate table of content (TOC)
assert(subprocess.run(['sh', '-c', script_dir + '/../pdftoc-to-latex "' + index_pdf_filename + '" > "' + build_dir + '/weasyprint/toc.tex"']).returncode == 0)
subprocess.run(['pdflatex', '-interaction', 'nonstopmode', '-output-directory=' + build_dir + '/weasyprint', build_dir + '/weasyprint/toc.tex'])

# Count TOC pages
toc_pdfinfo = subprocess.run(['pdfinfo', build_dir + '/weasyprint/toc.pdf'], stdout=subprocess.PIPE)
assert(toc_pdfinfo.returncode == 0)
toc_pages_match = re.search('\\nPages:\s+([0-9]+)\\n', toc_pdfinfo.stdout.decode())
toc_num_pages = int(toc_pages_match.group(1))
print('toc.pdf: ' + str(toc_num_pages) + ' page(s)')

# Count pages in index.pdf
pdfinfo = subprocess.run(['pdfinfo', index_pdf_filename], stdout=subprocess.PIPE)
assert(pdfinfo.returncode == 0)
pages_match = re.search('\\nPages:\s+([0-9]+)\\n', pdfinfo.stdout.decode())
num_pages = int(pages_match.group(1))
print('index.pdf: ' + str(num_pages) + ' pages')

num_pages = num_pages + toc_num_pages # account for table of content that will be added later

# If needed, update the CSS file with the correct number of pages
with open(css_filename) as css_file:
	original_css = css_file.read()

modified_css = re.sub('content: counter\(page\) "/[0-9]+";', 'content: counter(page) "/'+str(num_pages)+'";', original_css)

if modified_css != original_css:
	with open(css_filename, 'w') as css_file:
		css_file.write(modified_css)
	
	# We need to compile again with the modified CSS (this won't impact the TOC)
	print('Number of pages has changed, rebuilding PDF...')
	assert(subprocess.run(['weasyprint', build_dir + '/weasyprint/index.html', index_pdf_filename, '-s', css_filename]).returncode == 0)

# Insert TOC in the PDF
assert(subprocess.run(['pdftk', 'A='+index_pdf_filename, 'B='+build_dir+'/weasyprint/toc.pdf', 'cat', 'A'+str(insert_toc_after_page), 'B', 'A'+str(insert_toc_after_page+1)+'-end', 'output', build_dir + '/weasyprint/vheliotech-without-bookmarks.pdf']).returncode == 0)

# Restore bookmarks
extract_bookmarks_from = index_pdf_filename
source_pdf_filename = build_dir + '/weasyprint/vheliotech-without-bookmarks.pdf'
output_filename = build_dir + '/weasyprint/vheliotech.pdf'

bookmarks_filename = extract_bookmarks_from.replace('.pdf', '.txt')
assert(bookmarks_filename != extract_bookmarks_from)

# extract PDF metadata into a text file
assert(subprocess.run(['pdftk', extract_bookmarks_from, 'dump_data', 'output', bookmarks_filename]).returncode == 0)

with open(bookmarks_filename) as bookmarks_file:
	metadata = bookmarks_file.read()

# Remove link icon character at the end of each bookmark name (these are added by sphinx but make no sense in a PDF bookmark)
metadata = metadata.replace('', '')

# Remove bookmarks for small titles, adjust page number for remaining ones
def filterBookmark(match):
	#print('bookmark: "' + match.group(0) + '"')
	level = int(match.group(2))
	if level > max_bookmark_level:
		return ''
	initial_page = int(match.group(3))
	final_page = initial_page + toc_num_pages if initial_page > insert_toc_after_page else initial_page
	return 'BookmarkBegin\nBookmarkTitle: '+match.group(1).replace(' ', ' ')+'\nBookmarkLevel: '+match.group(2)+'\nBookmarkPageNumber: '+str(final_page)+'\n'
metadata = re.sub('BookmarkBegin\nBookmarkTitle: (.*)\nBookmarkLevel: ([0-9]+)\nBookmarkPageNumber: ([0-9]+)\n', filterBookmark, metadata)

with open(bookmarks_filename, 'w') as bookmarks_file:
	 bookmarks_file.write(metadata)

# generate the output PDF
assert(subprocess.run(['pdftk', source_pdf_filename, 'update_info', bookmarks_filename, 'output', output_filename]).returncode == 0)

print('Generated file: ' + output_filename)