git.scottworley.com Git - paperdoorknob/blob - glowfic.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
   8 from abc import ABC, abstractmethod
   9 from dataclasses import dataclass
  10 import itertools
  11
  12 from typing import Iterable
  13
  14 from bs4 import BeautifulSoup
  15 from bs4.element import Tag
  16
  17 from images import ImageStore
  18 from texify import Texifier
  19
  20
  21 @dataclass(frozen=True)
  22 class Chunk:
  23     icon: str | None
  24     character: str | None
  25     screen_name: str | None
  26     author: str | None
  27     content: Tag
  28
  29 # We avoid the name "post" because the Glowfic community uses the term
  30 # inconsistently:
  31 #  * The Glowfic software sometimes uses "post" to refer to a whole thread
  32 #    (in the URL), sometimes uses "post" to refer to chunks (in the CSS),
  33 #    but mostly uses "post" to refer to just the first chunk in a thread
  34 #    (in the HTML and UI).  The non-first chunks are "replies".
  35 #  * Readers and this software don't need to distinguish first-chunks and
  36 #    non-first-chunks.
  37 #  * Humans in the community tend to use "posts" to mean chunks.
  38
  39
  40 def chunkDOMs(html: BeautifulSoup) -> Iterable[Tag]:
  41     def text() -> Tag:
  42         body = html.body
  43         assert body
  44         text = body.find_next("div", class_="post-post")
  45         assert isinstance(text, Tag)
  46         return text
  47
  48     def the_replies() -> Iterable[Tag]:
  49         rs = html.find_all("div", class_="post-reply")
  50         assert all(isinstance(r, Tag) for r in rs)
  51         return rs
  52
  53     return itertools.chain([text()], the_replies())
  54
  55
  56 def makeChunk(chunk_dom: Tag, image_store: ImageStore) -> Chunk:
  57
  58     def getIcon() -> str | None:
  59         icon_div = chunk_dom.find_next('div', class_='post-icon')
  60         if icon_div is None:
  61             return None
  62         icon_img = icon_div.find_next('img')
  63         if icon_img is None:
  64             return None
  65         assert isinstance(icon_img, Tag)
  66         return image_store.get_image(icon_img.attrs['src'])
  67
  68     def getTextByClass(css_class: str) -> str | None:
  69         div = chunk_dom.find_next('div', class_=css_class)
  70         if div is None:
  71             return None
  72         return div.text.strip()
  73
  74     content = chunk_dom.find_next('div', class_='post-content')
  75     assert isinstance(content, Tag)
  76
  77     return Chunk(getIcon(),
  78                  getTextByClass('post-character'),
  79                  getTextByClass('post-screenname'),
  80                  getTextByClass('post-author'),
  81                  content)
  82
  83
  84 def renderIcon(icon_path: str | None) -> bytes:
  85     return b'\\includegraphics{%s}' % icon_path.encode(
  86         'UTF-8') if icon_path else b''
  87
  88
  89 class Layout(ABC):
  90
  91     @abstractmethod
  92     def renderChunk(self, chunk: Chunk) -> bytes:
  93         raise NotImplementedError()
  94
  95
  96 class ContentOnlyLayout(Layout):
  97
  98     def __init__(self, texifier: Texifier) -> None:
  99         self._texifier = texifier
 100
 101     def renderChunk(self, chunk: Chunk) -> bytes:
 102         return self._texifier.texify(chunk.content)
 103
 104
 105 class BelowIconLayout(Layout):
 106
 107     def __init__(self, texifier: Texifier) -> None:
 108         self._texifier = texifier
 109
 110     def renderChunk(self, chunk: Chunk) -> bytes:
 111         icon_width = b'0.25\\textwidth'  # TODO: Make this configurable
 112         return b'''\\begin{wrapfigure}{l}{%s}
 113 %s
 114 %s
 115 %s
 116 %s
 117 \\end{wrapfigure}
 118 %s
 119 ''' % (
 120             icon_width,
 121             renderIcon(chunk.icon),
 122             chunk.character.encode('UTF-8') if chunk.character else b'',
 123             chunk.screen_name.encode('UTF-8') if chunk.screen_name else b'',
 124             chunk.author.encode('UTF-8') if chunk.author else b'',
 125             self._texifier.texify(chunk.content))