Question for lxml experts, please

[jackie_w]

My lxml expertise is currently somewhat lacking. Is there a known technique, or sample calibre code I can look at, which can reliably identify matching start & end HTML tags?

My aims are two-fold:

1. to create something which will automatically find occurrences of <span class="italic">...</span> and <span class="bold">...</span> and replace them with 'naked' <i>...</i> and <b>...</b> tags.
2. to use this as a practical learning exercise to improve my parsing knowledge

P.S. I know Regex can easily be used to convert non-nested occurrences but if possible I'd like to create something which can also reliably handle the nested ones.

[kovidgoyal]

from calibre.ebooks.oeb.base import XHTML

for tag in root.iterdescendants('*'):
   if tag.name.endswith('}span'):
        cls = tag.get('class')
        if cls in {'bold', 'italic'}:
             tag.name = XHTML('b' if cls == 'bold' else 'i')

[jackie_w]

Is that it?!! Consider me informed ... and humbled

As ever, thank you for your help

[kovidgoyal]

You're welcome

These kinds of things are much easier to do with a fully parsed representation of the html, like lxml, rather than regexps.