is the height difference here expected since we are using a different DOM element for the screenshot?

Yeah, I was expecting this would change some of the snapshots but I was surprised that it made the snapshot smaller.

It looks like the reason is that there are some negative margins involved. The stHeader div is a bit shorter than the stMarkdownContainer which has this negative margin.

marginBottom: isLabel ? "" : `-${theme.spacing.lg}` // = -1rem = -16px

Wouldn't it be possible to just pass the text_inside_heading string to has_text?

Yes, we could, this is just following the same pattern as the get_markdown function, is there a difference in requirements? I know the heading elements directly use the same markdown component. Do you remember the reason for adding this in get_markdown, it makes sense to me to handle them all the same way?

I will merge this for now, but can make another PR to change these and maybe also utilize re.escape() as suggested by Copilot which seems like it might be a good idea also if we continue to convert strings to regex expressions but could be applied to all of these helper functions in the same way.

The get_heading function converts string inputs to regex patterns automatically, but this can cause issues when the heading text contains regex special characters. For example, if a heading contains characters like ., *, +, ?, [, ], etc., they will be interpreted as regex metacharacters rather than literal text.

Consider using re.escape() when converting strings to patterns:

if isinstance(text_inside_heading, str):
    text_inside_heading = re.compile(re.escape(text_inside_heading))

This ensures that headings with special characters like "Price: $5.00" or "Question?" are matched literally.