]> git.scottworley.com Git - paperdoorknob/blob - paperdoorknob.py
0fd1fdeceb6a5b9ff64b1a421fcc3b8f71a57c5a
[paperdoorknob] / 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 typing import Any, Iterable
8
9 from bs4 import BeautifulSoup
10
11 from args import spec_from_commandline_args
12 from glowfic import chunkDOMs, flatURL, makeChunk
13 from spec import Spec
14
15
16 def parse(content: bytes) -> BeautifulSoup:
17 return BeautifulSoup(content, 'html.parser')
18
19
20 def ilen(it: Iterable[Any]) -> int:
21 return sum(1 for _ in it)
22
23
24 def process(spec: Spec) -> None:
25 spec.texout.write(b'''\\documentclass{article}
26 \\usepackage{booktabs}
27 \\usepackage{graphicx}
28 \\usepackage{longtable}
29 \\usepackage{soul}
30 \\usepackage{varwidth}
31 \\usepackage{wrapstuff}
32 ''')
33 if spec.geometry is not None:
34 spec.texout.write(b'\\usepackage[' +
35 spec.geometry.encode('UTF-8') +
36 b']{geometry}\n')
37 spec.texout.write(b'''\\begin{document}
38 \\newcommand{\\href}[2]{#2\\footnote{#1}}
39 ''')
40 url = flatURL(spec.url)
41 spec.log('Fetching HTML...\r')
42 html = spec.fetcher.fetch(url)
43 spec.log('Parsing HTML...\r')
44 dom = parse(spec.htmlfilter(html))
45 spec.log('Counting chunks...\r')
46 num_chunks = ilen(chunkDOMs(dom))
47 for i, r in enumerate(chunkDOMs(dom)):
48 percent = 100.0 * i / num_chunks
49 spec.log(f'Processing chunk {i} of {num_chunks} ({percent:.1f}%)\r')
50 spec.domfilter(r)
51 chunk = makeChunk(r, spec.images)
52 spec.texout.write(spec.texfilter(spec.layout.renderChunk(chunk)))
53 spec.log('')
54 spec.texout.write(b'\\end{document}\n')
55
56
57 def main() -> None:
58 with spec_from_commandline_args() as spec:
59 process(spec)
60
61
62 if __name__ == '__main__':
63 main()