Unable to extract text inside <code> tag

[FernandoSSI]

Bug

When I try to extract the content of an API documentation page using docling, I realized that it cannot extract the text that is inside the <code> tag in html.

Steps to reproduce

url: https://api.slack.com/methods/conversations.acceptSharedInvite

The <code> blocks in the following image, which contain the path of the http request, cannot be extracted in the result markdown:

the result markdown of this part in the above picture is:

Docling version

Version: 2.28.4

Python version

Version: 3.12.15

[ceberam]

Thanks @FernandoSSI for your interest in docling!
The initial purpose of our docling's HTML parser was to keep the relevant textual and tabular elements of the document for AI applications and therefore we only considered some tags for conversion, such as <p>, <h1>, or <table>. The full list can be found in the backend implementation here.

We recently support formatted text and we think addressing the code tag would be feasible and a good enhancement. We would then like to seralize this text properly, for instance, with the backticks in Markdown. We got a similar request a while ago with the pre tag in issue #222

We will create a PR with this enhancement soon.

[ceberam]

Thanks @FernandoSSI for your fix suggestion. Your PR passes all checks and tests.
Before merging, could we ask you to add a test that contains such <code> tag? We don't have any in our regression tests and therefore we cannot test if your patch solves the issue.
You could simply add a dummy HTML document in folder tests/data/html, run the backend test (poetry run pytest tests/test_backend_html.py) and ensure that the generated conversion in folder is as expected (tests/data/groundtruth/docling_v2).