meltdown

A naive Markdown parser in pure Python.

WARNING: The library will never attempt to do any sanitization on the input. If used for user generated content it is highly recommended to use an external sanitization library.

Installation

meltdown currently only supports Python 3.12.

pip install meltdown

Usage

from meltdown import MarkdownParser, HtmlProducer

doc = MarkdownParser().parse("# Hello **friends**!")
html = HtmlProducer().produce(doc)

print(doc.dump())
print(html)

The default HtmlProducer is heavily inspired by pandoc, however, if you are unhappy you can easily write your own producer or if only some formattings are unwanted override the default methods.

In the following example we change the bold formatting form <strong> to <b>:

from meltdown import MarkdownParser, HtmlProducer
from meltdown.Nodes import *
from typing import Self

class CustomHtmlProducer(HtmlProducer):
    def visit_bold(self: Self, node: BoldNode):
        self._output += "<b>"
        for child in node.children:
            child.accept(self)
        self._output += "<b>"

doc = MarkdownParser().parse("# Hello **friends**!")
html = CustomHtmlProducer().produce(doc)

print(html)

Development

To quickly test some stuff there is a cli that consumes the library which you can run with:

usage: meltdown-dev-cli [-h] [--dump] filename

A cli for developers trying meltdown

positional arguments:
  filename    Name of the file being parsed and converted.

options:
  -h, --help  show this help message and exit
  --dump      Print the dump instead of the html.

uv run cli.py tests/blog/post-jit.md

Run all tests

uv run pytest

Most of our tests are snapshot tests which you can update with the following if you add new ones.

uv run pytest --inline-snapshot=create