What is the best practice for code runner for markdown files?

[epogrebnyak]

I'm running the code from documentation to make it up-to date, wonder is there is a better way than I do.

My workflow:

• extract fenced code blocks with codedown
• save to local file and run it with poetry run python
• do the same for command line scripts (.sh or .bat )

Code for this is here: https://github.com/epogrebnyak/abacus/blob/main/justfile#L60-L72

The workflow is far form perfect:

• I have to preprocess files to extract code from content tabs
• the command line scripts do not stop on error, just show something bad on a screen
• invoking codedown with npx is rather slow
• many files are created in docs folder

This a working, but not elegeant solution, so my question is - how do you test you the code in docs, what is th best practice to do so?
Specifically - is there a code extractor that understands fenced code blocks with offset?

[epogrebnyak]

Thnking of adding something like:

"""Extract code blocks from markdown file  

If this filename is md.py can use as:

    poetry run python md.py | poetry run python

For more functionality: https://github.com/earldouglas/codedown 
"""

from markdown_it import MarkdownIt

def strip(text: str) -> str:
    return "\n".join([line.strip() for line in text.split("\n")])

def codeblock(text: str, info="python") -> str:
    md = MarkdownIt('commonmark')
    tokens = [t for t in md.parse(strip(text)) if t.type=="fence" and t.info==info]
    return "\n\n".join([t.content for t in tokens])


text = """

    ```python
    print("line with offset")
    ```

    ```python
    print("another offset")
    ```
"""
print(codeblock(text))