Simply thank you!

[Lucas-C]

Hi!

I just want to report a happy user 🙂

I have switched from using markdown-it (NodeJS) to mistletoe, and I really love it.

This morning, in 30min, I was able to quickly add support for ::: block containers, in a similar fashion as markdown-it-container:

from mistletoe import HtmlRenderer
from mistletoe.block_token import tokenize, BlockToken

class CustomHtmlRenderer(HtmlRenderer):
    def __init__(self):
        super().__init__(TripleCommaDiv)

    def render_triple_comma_div(self, token):
        inner = self.render_inner(token)
        return f'<div class="{token.classes}">{inner}</div>'

class TripleCommaDiv(BlockToken):
    """
    Simple <div> block. (["::: class1 class2", ..., ":::"])
    Block start is indicated by a line starting with at least three ":" characters.
    Same for the block end.
    The exact number of ":" characters does not matter at all.

    This aims to be compliant / cover the same functionality as markdown-it-container:
    https://www.npmjs.com/package/markdown-it-container

    Attributes:
        classes (str): CSS class names inserted in the "class" HTML attribute
    """

    @staticmethod
    def start(line):
        return line.startswith(":::")

    @classmethod
    def read(cls, lines):
        first_line = next(lines)
        classes = first_line.lstrip(":").strip()
        delimiter = cls._delimiter_from_line(first_line)
        child_lines = []
        for line in lines:
            if line.startswith(delimiter):
                if line[len(delimiter)] != ":":
                    # End block found:
                    break
                else:
                    print(f"WARN: Unexpected longer delimiter: '{line.rstrip()}' - Expected block end delimiter: {delimiter}")
            child_lines.append(line)
        children = tokenize(child_lines)
        return classes, children

    @staticmethod
    def _delimiter_from_line(line):
        level = 0
        while line[3 + level] == ":":
            level += 1
        return ":" * (3 + level)

    def __init__(self, match):
        self.classes, self.children = match

Example Markdown content that can be processed by this parser:

:::: class1

## Heading

Some parapgraph text.

::: class2 class3

## Sub-heading

Some other content.

:::

::::

The API of your library is really well crafted, and its excellent design makes it very easy to extend!

So I just wanted to say THANK YOU 👍

[Lucas-C]

Also, I have a question: how do you recommand to handle malformed Markdown? Should a BlockToken subclass raise an error in the read() method? Or would it better to try to produce a valid AST and just produce a warning, or a log line, describing the problem ?

[pbodnar]

Hi @Lucas-C, thanks for your message, I guess it would make @miyuchina happy, provided he still watches this project. :)

To your question, I think the common approach so far is to skip the "malformed markdown" and treat it as it would be a plain text instead - as it actually could be just a text. At least that is what I can see in the various examples in the CommonMark spec. But I would also say it also depends on the exact case we want to handle. So your suggestions could possibly be also valid, but I'm not just sure at the moment...

[tim-forrer]

@Lucas-C thank you also for providing this example for a custom block token - I found it really helpful!

Perhaps it could be a good idea to add this example to the developer guide, so that there is one example for a span token and one for a block token?

[TheCodeForge]

Perhaps it could be a good idea to add this example to the developer guide, so that there is one example for a span token and one for a block token?

229 opened

[pbodnar]

OK, so I think we can safely close this one now. :)