Sith/core/markdown.py

133 lines
3.9 KiB
Python
Raw Permalink Normal View History

#
# Copyright 2023 © AE UTBM
# ae@utbm.fr / ae.info@utbm.fr
#
# This file is part of the website of the UTBM Student Association (AE UTBM),
# https://ae.utbm.fr.
#
# You can find the source code of the website at https://github.com/ae-utbm/sith
#
# LICENSED UNDER THE GNU GENERAL PUBLIC LICENSE VERSION 3 (GPLv3)
2024-09-23 08:25:27 +00:00
# SEE : https://raw.githubusercontent.com/ae-utbm/sith/master/LICENSE
# OR WITHIN THE LOCAL FILE "LICENSE"
#
#
2024-07-01 15:33:05 +00:00
from __future__ import annotations
import re
2024-07-01 15:33:05 +00:00
from typing import TYPE_CHECKING
2024-06-24 11:07:36 +00:00
2024-07-01 15:33:05 +00:00
import mistune
from django.urls import reverse
2024-07-01 15:33:05 +00:00
from mistune import HTMLRenderer, Markdown
if TYPE_CHECKING:
from mistune import InlineParser, InlineState
# match __text__, without linebreak in the text, nor backslash prepending an underscore
# Examples :
# - "__text__" : OK
# - "__te xt__" : OK
# - "__te_xt__" : nope (underscore in the middle)
# - "__te\_xt__" : Ok (the middle underscore is escaped)
# - "__te\nxt__" : nope (there is a linebreak in the text)
# - "\__text__" : nope (one of the underscores have a backslash prepended)
# - "\\__text__" : Ok (the backslash is ignored, because there is another backslash before)
UNDERLINED_RE = (
r"(?<!\\)(?:\\{2})*" # ignore if there is an odd number of backslashes before
r"_{2}" # two underscores
r"(?P<underlined>([^\\_]|\\.)+)" # the actual text
r"_{2}" # closing underscores
)
SITH_LINK_RE = (
r"\[(?P<page_name>[\w\s]+)\]" # [nom du lien]
r"\(page:\/\/" # (page://
r"(?P<page_slug>[a-zA-Z0-9][a-zA-Z0-9.-]*[a-zA-Z0-9])" # actual page name
r"\)" # )
)
CUSTOM_DIMENSIONS_IMAGE_RE = (
r"\[(?P<img_name>[\w\s]+)\]" # [nom du lien]
r"\(img:\/\/" # (img://
r"(?P<img_slug>[a-zA-Z0-9][a-zA-Z0-9.-]*[a-zA-Z0-9])" # actual page name
r"\)" # )
)
def parse_underline(_inline: InlineParser, m: re.Match, state: InlineState):
state.append_token({"type": "underline", "raw": m.group("underlined")})
return m.end()
def underline(md_instance: Markdown):
md_instance.inline.register(
2018-10-04 19:29:19 +00:00
"underline",
2024-07-01 15:33:05 +00:00
UNDERLINED_RE,
parse_underline,
before="emphasis",
)
md_instance.renderer.register("underline", lambda _, text: f"<u>{text}</u>")
def parse_sith_link(_inline: InlineParser, m: re.Match, state: InlineState):
page_name = m.group("page_name")
page_slug = m.group("page_slug")
state.append_token(
{
"type": "link",
"children": [{"type": "text", "raw": page_name}],
"attrs": {"url": reverse("core:page", kwargs={"page_name": page_slug})},
}
)
return m.end()
def sith_link(md_instance: Markdown):
md_instance.inline.register(
"sith_link",
SITH_LINK_RE,
parse_sith_link,
before="emphasis",
)
# no custom renderer here.
# we just add another parsing rule, but render it as if it was
# a regular markdown link
class SithRenderer(HTMLRenderer):
def image(self, text: str, url: str, title=None) -> str:
if "?" not in url:
return super().image(text, url, title)
new_url, params = url.rsplit("?", maxsplit=1)
m = re.match(r"^(?P<width>\d+(%|px)?)(x(?P<height>\d+(%|px)?))?$", params)
if not m:
return super().image(text, url, title)
width, height = m.group("width"), m.group("height")
if not width.endswith(("%", "px")):
width += "px"
style = f"width:{width};"
if height is not None:
if not height.endswith(("%", "px")):
height += "px"
style += f"height:{height};"
return super().image(text, new_url, title).replace("/>", f'style="{style}" />')
markdown = mistune.create_markdown(
renderer=SithRenderer(escape=True),
plugins=[
underline,
sith_link,
2018-10-04 19:29:19 +00:00
"strikethrough",
2024-07-01 15:33:05 +00:00
"footnotes",
"table",
"spoiler",
"subscript",
"superscript",
2018-10-04 19:29:19 +00:00
"url",
2024-07-01 15:33:05 +00:00
],
)