git.scottworley.com Git - paperdoorknob/blob - paperdoorknob.py

   1 # paperdoorknob: Print glowfic
   2 #
   3 # This program is free software: you can redistribute it and/or modify it
   4 # under the terms of the GNU General Public License as published by the
   5 # Free Software Foundation, version 3.
   6
   7 from sys import stderr
   8
   9 from typing import Any, Iterable
  10
  11 from bs4 import BeautifulSoup
  12
  13 from args import spec_from_commandline_args
  14 from glowfic import chunkDOMs, flatURL, makeChunk
  15 from spec import Spec
  16
  17
  18 def parse(content: bytes) -> BeautifulSoup:
  19     return BeautifulSoup(content, 'html.parser')
  20
  21
  22 def ilen(it: Iterable[Any]) -> int:
  23     return sum(1 for _ in it)
  24
  25
  26 def process(spec: Spec) -> None:
  27     spec.texout.write(b'''\\documentclass{article}
  28 \\usepackage{booktabs}
  29 \\usepackage{graphicx}
  30 \\usepackage{longtable}
  31 \\usepackage{soul}
  32 \\usepackage{varwidth}
  33 \\usepackage{wrapstuff}
  34 ''')
  35     if spec.geometry is not None:
  36         spec.texout.write(b'\\usepackage[' +
  37                           spec.geometry.encode('UTF-8') +
  38                           b']{geometry}\n')
  39     spec.texout.write(b'\\begin{document}\n')
  40     url = flatURL(spec.url)
  41     html = parse(spec.htmlfilter(spec.fetcher.fetch(url)))
  42     num_chunks = ilen(chunkDOMs(html))
  43     for i, r in enumerate(chunkDOMs(html)):
  44         percent = 100.0 * i / num_chunks
  45         print(
  46             f'Processing chunk {i} of {num_chunks} ({percent:.1f}%)',
  47             end='\r',
  48             file=stderr)
  49         spec.domfilter(r)
  50         chunk = makeChunk(r, spec.images)
  51         spec.texout.write(spec.texfilter(spec.layout.renderChunk(chunk)))
  52     spec.texout.write(b'\\end{document}\n')
  53
  54
  55 def main() -> None:
  56     with spec_from_commandline_args() as spec:
  57         process(spec)
  58
  59
  60 if __name__ == '__main__':
  61     main()