Detect code in docx

[lukasberbuer]

Requested feature

OpenXML doesn't offer an explicit way to identify code (blocks). But we could use some detection rules, e.g. the combination of:

• Style name lowercase includes "code"
• Paragraph uses monospaced fonts (is there already a checker somewhere in the code base?)

What do you think?
If that's of interest, I could prepare a PR. I guess the right place would be MsWordDocumentBackend._handle_text_elements?

[cau-git]

@lukasberbuer That is a great idea. You are very welcome to propose a PR with a heuristic for this.

[LucasArray]

Hi @cau-git , following up on your invitation here:

I have opened #3735, implementing the detection along the lines @lukasberbuer proposed. It uses both signals, a curated code style-name tier (matched through style inheritance, pandoc-style) and a guarded monospaced-font fallback, with consecutive code paragraphs merged into single CodeItems. _handle_text_elements was indeed the right place, thanks for the pointer.

Happy to adjust anything in review.