minchin.pelican.readers.commonmark 2.0.1

CommonMark Reader for Pelican (via Markdown-IT)

Powered by Markdown-IT-Py

This plugin is intended to be a roughly drop-in replacement for Pelican’s built-in Markdown Reader (the “reader” is the part of Pelican that turns your source files into something Pelican can assemble into a website). As this uses a CommonMark implementation of Markdown (specifically, markdown-it-py, there are be subtle differences when compared to the output of Pelican’s built in Markdown Reader (which relies on Python-Markdown); if you are particular about your site output, it may require building the site with the two readers, and running a diff on the two outputs, and tweaking your source files (or adding Markdown plugins here) until the output is what you want.

When I set out to build this plugin, I (naively) thought I would stick to a “pure” CommonMark/Markdown implementation, but I quickly realized that I like the extensions to Markdown I use, and I wasn’t ready to give them up. That said, I’ve tried to keep them generally mild. The default configuration will automatically include all the plugins that I use by default, although you can add or remove from that list as you wish. Currently enabled CommonMark extensions are:

• front matter

• footnotes

• defintion list

• tables

Changes Required from “Vanilla” Pelican

This (Pelican) plugin uses the Markdown-IT front matter (plugin) by default. This expects front matter (i.e. metadata) to be at the top of the file, between lines of three dashes (e.g. ---). (“Vanilla” Pelican doesn’t require these marker lines.) Front matter is then phrased as YAML.

An example:

---
title: Front matter Test
Category: test
date: 2023-11-19 14:56
---

This is a test of the front matter plugin.

If you don’t want to (or can’t) update your source files, you can provide a customized COMMONMARK settings (in your pelicanconf.py) that doesn’t include the frontmatter plugin.

If the frontmatter plugin is not active, the plugin should parse metadata in the same matter as “vanilla” Pelican. Note that this has been less tested, as I personally use the front matter plugin.

To rely on “vanilla Pelican” front matter (generally, not recommended):

# pelicanconf.py

from minchin.pelican.readers.commonmark.constants import COMMONMARK_DEFAULT_CONFIG

# other Pelican settings

# start with default configuration
COMMONMARK = COMMONMARK_DEFAULT_CONFIG
# add remove Markdown-IT's front matter plugin
for i, plugin in enumerate(COMMONMARK["extensions"]):
    if plugin.__name__ == "front_matter_plugin":
        del a["extensions"][i]

Additional Features

In addition to the “base” CommonMark parser/render, this Reader offers the following additional features:

• Title from H1: if a post’s title isn’t defined in the metadata block, it will try and pull a title from the first H1 tag in the body of the entry.

• Remove duplicated H1 title: If the first H1 tag in the post matches the title as defined in the metadata block, it will remove the H1 tag. It is assumed that the theme will include the title as a H1 tag in the generated site.

• Relative links ready for Pelican: relative links included in the body of posts will have {filename} or {static} prefixed to them, so that Pelican can maintain these links even if the generated site has a different layout from your source files.

• Code block highlighting: Pygments is called to allow code block syntax highlighting. Generated site HTML will display code highlighting if you include (or link to) a Pygments CSS file.

• removes “tag only” lines from the body of your entries.

• “cleans” dates provided in front matter, so they are provided to Pelican as datetimes rather than strings.

Pelican Settings

COMMONMARK = {“extensions”: [<plugin classes>], “enable”: [<str of name of features>]}

(To be defined). Used to configure which CommonMark extensions are loaded by the plugin. The default is available at minchin.pelican.readers.commonmark.constants.COMMONMARK_DEFAULT_CONFIG.

This is a dictionary, expecting two keys: extensions and enable, each with a list as the key. For extensions, the list items should be the classes of the Markdown-IT plugins (aka “extensions”) you want to use. For enable, it should be the names (as strings) of the Markdown-IT features you want to enable (e.g. "table").

COMMONMARK_VERSION

Version of the plugin. Inserted by the plugin (if not provided).

COMMONMARK_DEV_URL

Homepage URL of the plugin. Inserted by the plugin (if not provided).

COMMONMARK_HTML_PARSER = “html.parser”

Will be set to “lxml” is it is installed. This is the parser that Beautiful Soup uses.

COMMONMARK_MARKDOWN_LOG_LEVEL = logging.WARNING

If you want to see the debugging for the Markdown-IT library change this to logging.DEBUG (but be warned that it is very verbose).

COMMONMARK_INLINE_TAG_SYMBOLS = “#”

Tag symbols used before inline tags. If a line contains only tags, it will be removed from the body of the entry.

Extended Abilities

I have written a markdown-it-py plugin to support “fancy” tasklists/checkboxes, but it is not activated by default.

This requires separate installation and activation within Pelican, which you might do like this:

# pelicanconf.py

from minchin.pelican.readers.commonmark.constants import COMMONMARK_DEFAULT_CONFIG
import minchin.md_it.fancy_tasklists

# other Pelican settings

# start with default configuration
COMMONMARK = COMMONMARK_DEFAULT_CONFIG
# add fancy tasklists
COMMONMARK["extensions"].append(
    minchin.md_it.fancy_tasklists.fancy_tasklists_plugin,
)

See sample rendered checkboxes.

Prior Art

This plugin relies on much work that has gone before, both explicitly for code and implicitely for the encouragement of this even being possible. This list is sadly incomplete, but in particlar:

• Johnathan Sundqvist’s Obisidian Plugin for Pelican (and forks) – in particular, for providing inspiration on how to deal with Wiki-style links