Markdown to Markdown using Mistune (experimental and incomplete)

md2md.py

from mistune import create_markdown
from mistune.renderers import BaseRenderer
import re

ESCAPE_CHAR = re.compile(r'(?<!\\)([\\`*_()\[\]#+-])')
UL_BULLET = re.compile(r'(?<=^)(\*)( +)', re.MULTILINE)


def indent(text, level, tab_length=4):
    ''' Indent block of text by level '''
    return '\n'.join(f'{" "*tab_length*level}{line}' for line in text.split('\n'))


class MdRenderer(BaseRenderer):
    NAME = 'md'
    IS_TREE = False

    def text(self, text):
        # TODO: escaping is probably more agressive than it needs to be.
        return ESCAPE_CHAR.sub(r'\\\1', text)

    def link(self, link, text=None, title=None):
        if link == text or ('@' in text and link.startswith('mailto:') and link[7:] == text):
            # Autolink
            return f'<{text}>'
        text = link if text is None else text
        title = f' "{title}"' if title is not None else ''
        return f'[{text}]({link}{title})'

    def image(self, src, alt="", title=None):
        title = f' "{title}"' if title is not None else ''
        return f'![{alt}]({src}{title})'

    def emphasis(self, text):
        return f'*{text}*'

    def strong(self, text):
        return f'**{text}**'

    def codespan(self, text):
        # TODO: account for double backticks in code span.
        if '`' in text:
            return f'`` {text} ``'
        return f'`{text}`'

    def linebreak(self):
        return '  \n'

    def inline_html(self, html):
        return html

    def paragraph(self, text):
        return f'{text}\n\n'

    def heading(self, text, level):
        return f'{"#"*level} {text}\n\n'

    def newline(self):
        return '\n'

    def thematic_break(self):
        return '- - -\n\n'

    def block_text(self, text):
        return text

    def block_code(self, code, info=None):
        info = info or ''
        code = code.rstrip('\n')
        return f'```{info}\n{code}\n```\n\n'

    def block_quote(self, text):
        return '\n'.join([f'> {line}' for line in text.strip().splitlines()]) + '\n'

    def block_html(self, html):
        return f'{html}\n\n'

    def block_error(self, html):
        # TODO: this is non-standard. Maybe ignore?
        return f'<div class="error">{html}</div>\n\n'

    def list(self, text, ordered, level, start=None):
        if ordered:
            # convert `*` to `1.` in each list item using `start`.
            # TODO: make this increment. But how?
            start = f'{start}. ' if start is not None else '1. '
            text = UL_BULLET.sub(f'{start:<4}', text)
        if level == 1:
            # Adjust blank lines for level 1 lists
            # TODO: fix some edge cases with nested lists
            text = text.lstrip('\n')
            text += '\n\n'
        return text

    def list_item(self, text, level):
        if '\n' in text:
            # Indent all lines after the first line.
            firstline, therest = text.split('\n', 1)
            text = '\n'.join([firstline, indent(therest, 1)])
        # The linebreak goes at the front for nested items
        return f'\n*   {text}'


md2md = create_markdown(escape=False, renderer=MdRenderer())


if __name__ == '__main__':
    import sys
    if len(sys.argv) > 1:
        if sys.argv[1] == '-':
            src = sys.stdin.read()
        else:
            with open(sys.argv[1]) as f:
                src = f.read()
        print(md2md(src))
    else:
        print(f'usage: {sys.argv[0]} FILE|-')

lepture/mistune#208 may be relevant to this.

Also, note that this uses f-strings (see PEP 498), which are only supported in Python 3.6+.

Hi @waylan ! Here at SJTU-CSE/awesome-cs#31 , I made use of your script to automatically validate if items in Markdown list are sorted in dict order, and pretty-printed the validator result out. Thank you for this awesome work!