]>
Commit | Line | Data |
---|---|---|
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, image_size: float) -> str | None: | |
85 | params = f'width={image_size}mm,height={image_size}mm,keepaspectratio' | |
86 | return f'\\includegraphics[{params}]{{{icon_path}}}' if icon_path else None | |
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) + b'\n' | |
103 | ||
104 | ||
105 | class BelowIconLayout(Layout): | |
106 | ||
107 | def __init__(self, texifier: Texifier, image_size: float) -> None: | |
108 | self._texifier = texifier | |
109 | self._image_size = image_size | |
110 | ||
111 | def renderChunk(self, chunk: Chunk) -> bytes: | |
112 | meta = [ | |
113 | renderIcon(chunk.icon, self._image_size), | |
114 | chunk.character, | |
115 | chunk.screen_name, | |
116 | chunk.author, | |
117 | ] | |
118 | ||
119 | return b'''\\wrapstuffclear | |
120 | \\begin{wrapstuff}[l] | |
121 | \\fbox{ | |
122 | \\begin{varwidth}{0.5\\textwidth} | |
123 | \\smash{\\parbox[t][0pt]{0pt}{ | |
124 | \\setlength{\\fboxrule}{0.2pt} | |
125 | \\setlength{\\fboxsep}{0pt} | |
126 | \\vspace{-3.4pt} | |
127 | \\fbox{\\hspace{107mm}} | |
128 | }\\\\*} | |
129 | \\vspace{-1em} | |
130 | \\begin{center} | |
131 | %s | |
132 | \\end{center} | |
133 | \\end{varwidth} | |
134 | } | |
135 | \\end{wrapstuff} | |
136 | ||
137 | \\strut | |
138 | ||
139 | \\noindent %s | |
140 | ''' % ( | |
141 | b'\\\\*'.join(x.encode('UTF-8') for x in meta if x is not None), | |
142 | self._texifier.texify(chunk.content)) | |
143 | ||
144 | ||
145 | class BesideIconLayout(Layout): | |
146 | ||
147 | def __init__(self, texifier: Texifier, image_size: float) -> None: | |
148 | self._texifier = texifier | |
149 | self._image_size = image_size | |
150 | ||
151 | def renderChunk(self, chunk: Chunk) -> bytes: | |
152 | icon = renderIcon(chunk.icon, self._image_size) | |
153 | meta = [ | |
154 | chunk.character, | |
155 | chunk.screen_name, | |
156 | chunk.author, | |
157 | ] | |
158 | ||
159 | # Why is \textwidth not the width of the text? | |
160 | # Why is the width of the text .765\textwidth? | |
161 | return b'''\\noindent\\fbox{ | |
162 | %s | |
163 | \\parbox[b]{.765\\textwidth}{ | |
164 | \\begin{center} | |
165 | %s | |
166 | \\end{center} | |
167 | } | |
168 | }\\\\* | |
169 | \\vspace{-0.75em}\\\\* | |
170 | \\noindent %s | |
171 | ||
172 | \\strut | |
173 | ||
174 | ''' % ( | |
175 | icon.encode('UTF-8') if icon else b'', | |
176 | b'\\\\*'.join(x.encode('UTF-8') for x in meta if x is not None), | |
177 | self._texifier.texify(chunk.content)) |