How do I unparse restructured text back into an rst file?

Question

My question is a follow up to How to parse restructuredtext in python? where @mbdevpl provided code for parsing the file into a tree and doing some processing on the tree.

I want to autoformat Python code in RST files. Now that I've discovered the code blocks in the tree and changed their contents, how do I write the changes back over the original file? Right now I'm doing this with regexes.

Acrobatics

You could use a package like rst.readthedocs.io/en/latest – Marissamarist

doesn't seem to be a serious package – Acrobatics

Looks like a duplicate of Is there a reST Writer? – Singlestick

Answer 1

Considering the lack of a good solution in my own searching, I would suggest a paragraph from https://docutils.sourceforge.io/docs/dev/hacking.html#parsing-the-document, which is linked in your previous question.

To get an HTML document out of the node tree, we use a Writer, the HTML writer in this case (docutils/writers/html4css1.py).

The writer receives the node tree and returns the output document. For HTML output, we can test this using the rst2html.py tool:

So you need a writer of some sort. One for html exists, and perhaps you could pipe that output to html2rst

Consternate

Answer 2

An rST writer for Docutils is a long standing TODO item. To generate reStructuredText documents with Docutils, you may try the XML (Docutils native) writer and the xml2rst processor by Stefan Merten.

The latest version at http://www.merten-home.de/FreeSoftware/xml2rst is underdocumented, the documentation for a previous version points to additonal requirements. Unfortunately the package and custom requirements do not work with Python 3.x (required by current Docutils). The XSLT stylesheet at the core of the converter may still work independent of Python, see Other ways to use xml2rst.

Singlestick

Note: the code of xml2rst is broken and does not run because of unresolvable imports. – Adagietto

Thanks for pointing this out. I'll edit the answer. – Singlestick