X-Git-Url: http://git.scottworley.com/paperdoorknob/blobdiff_plain/e6adf6ced68d667429975110f2d1a1bd9c8d79b6..69f86f06f2bfed44d957f0b5eabbdeb47a0e9572:/glowfic.py?ds=inline diff --git a/glowfic.py b/glowfic.py index 901246b..a4f3ca9 100644 --- a/glowfic.py +++ b/glowfic.py @@ -5,6 +5,8 @@ # Free Software Foundation, version 3. +from abc import ABC, abstractmethod +from dataclasses import dataclass import itertools from typing import Iterable @@ -12,15 +14,27 @@ from typing import Iterable from bs4 import BeautifulSoup from bs4.element import Tag +from images import ImageStore +from texify import Texifier + + +@dataclass(frozen=True) +class Chunk: + icon: str | None + character: str | None + screen_name: str | None + author: str | None + content: Tag + # We avoid the name "post" because the Glowfic community uses the term # inconsistently: # * The Glowfic software sometimes uses "post" to refer to a whole thread -# (eg: in the URL), but more often uses "post" to refer to just the first -# chunk in a thread. The non-first chunks are "replies". +# (in the URL), sometimes uses "post" to refer to chunks (in the CSS), +# but mostly uses "post" to refer to just the first chunk in a thread +# (in the HTML and UI). The non-first chunks are "replies". # * Readers and this software don't need to distinguish first-chunks and # non-first-chunks. -# * Humans in the community tend to use "posts" to mean "chunks" ("replies" -# in the Glowfic software's lexicon). +# * Humans in the community tend to use "posts" to mean chunks. def chunkDOMs(html: BeautifulSoup) -> Iterable[Tag]: @@ -37,3 +51,81 @@ def chunkDOMs(html: BeautifulSoup) -> Iterable[Tag]: return rs return itertools.chain([text()], the_replies()) + + +def makeChunk(chunk_dom: Tag, image_store: ImageStore) -> Chunk: + + def getIcon() -> str | None: + icon_div = chunk_dom.find_next('div', class_='post-icon') + if icon_div is None: + return None + icon_img = icon_div.find_next('img') + if icon_img is None: + return None + assert isinstance(icon_img, Tag) + return image_store.get_image(icon_img.attrs['src']) + + def getTextByClass(css_class: str) -> str | None: + div = chunk_dom.find_next('div', class_=css_class) + if div is None: + return None + return div.text.strip() + + content = chunk_dom.find_next('div', class_='post-content') + assert isinstance(content, Tag) + + return Chunk(getIcon(), + getTextByClass('post-character'), + getTextByClass('post-screenname'), + getTextByClass('post-author'), + content) + + +def renderIcon(icon_path: str | None, image_size: float) -> bytes: + return b'\\includegraphics[width=%fmm,height=%fmm,keepaspectratio]{%s}' % ( + image_size, image_size, icon_path.encode('UTF-8')) if icon_path else b'' + + +class Layout(ABC): + + @abstractmethod + def renderChunk(self, chunk: Chunk) -> bytes: + raise NotImplementedError() + + +class ContentOnlyLayout(Layout): + + def __init__(self, texifier: Texifier) -> None: + self._texifier = texifier + + def renderChunk(self, chunk: Chunk) -> bytes: + return self._texifier.texify(chunk.content) + + +class BelowIconLayout(Layout): + + def __init__(self, texifier: Texifier, image_size: float) -> None: + self._texifier = texifier + self._image_size = image_size + + def renderChunk(self, chunk: Chunk) -> bytes: + icon_width = b'0.25\\textwidth' # TODO: Make this configurable + return b'''\\begin{wrapfigure}{l}{%s} +%s + +%s + +%s + +%s +\\end{wrapfigure} + +%s + +''' % ( + icon_width, + renderIcon(chunk.icon, self._image_size), + chunk.character.encode('UTF-8') if chunk.character else b'', + chunk.screen_name.encode('UTF-8') if chunk.screen_name else b'', + chunk.author.encode('UTF-8') if chunk.author else b'', + self._texifier.texify(chunk.content))