Use myst-parser in a sphinx extension to parse md to docutils nodes

[ewjoachim]

I'm trying to refactor a part of Sphinx-GitHub-Changelog, a sphinx extension that generates a changelog based on GitHub releases.

I can have either the GFM or the HTML of a GitHub release body (whichever is best). I'd like to use Myst-Parser to turn those into docutils nodes that my sphinx directive should return.

Does Myst-Parser expose an API to do that? I've spent some time on the doc and couldn't find it. There's the MystParser.parse(), but I don't really have a document to provide (my directive is expected to return nodes, not write them on the document directly) and even if I did, it would need to have myst config, which I don't know how to pass. I'm (as usual) confused by the Docutils API.

Anyone has an idea ?

[holmboe]

Is the Python code in https://myst-parser.readthedocs.io/en/latest/docutils.html#single-page-builds of any help to you?

[ewjoachim]

Thanks for the suggestion! I've read it, but didn't play with it. From the first glance, I don't see how to output docutils nodes.

[agoose77]

myst-parser renders MarkdownIt tokens and populates an existing docutils node (document). See

MyST-Parser/myst_parser/mdit_to_docutils/base.py

Line 100 in d3d7fbb

class DocutilsRenderer(RendererProtocol):

for the base-class of the renderer that is used.

Does this provide a helpful place to start? :)

[ewjoachim]

Thank you for your answer, I'll do my best to get going from there.