]>
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 | import itertools | |
9 | ||
10 | from typing import Iterable | |
11 | ||
12 | from bs4 import BeautifulSoup | |
13 | from bs4.element import Tag | |
14 | ||
15 | # We avoid the name "post" because the Glowfic community uses the term | |
16 | # inconsistently: | |
17 | # * The Glowfic software sometimes uses "post" to refer to a whole thread | |
18 | # (eg: in the URL), but more often uses "post" to refer to just the first | |
19 | # chunk in a thread. The non-first chunks are "replies". | |
20 | # * Readers and this software don't need to distinguish first-chunks and | |
21 | # non-first-chunks. | |
22 | # * Humans in the community tend to use "posts" to mean "chunks" ("replies" | |
23 | # in the Glowfic software's lexicon). | |
24 | ||
25 | ||
26 | def chunkDOMs(html: BeautifulSoup) -> Iterable[Tag]: | |
27 | def text() -> Tag: | |
28 | body = html.body | |
29 | assert body | |
30 | text = body.find_next("div", class_="post-post") | |
31 | assert isinstance(text, Tag) | |
32 | return text | |
33 | ||
34 | def the_replies() -> Iterable[Tag]: | |
35 | rs = html.find_all("div", class_="post-reply") | |
36 | assert all(isinstance(r, Tag) for r in rs) | |
37 | return rs | |
38 | ||
39 | return itertools.chain([text()], the_replies()) |