Regex: grabbing <h3><span> tag group

[meghane_e]

Need help with my regex expression to search <h3.*>(\w.+)</.*></h3> and replace it with <h3>\1</h3>
It was working and caught several of the sections I'm after, but stopped working on the rest. An example it failed on is the blurb below, which looks identical to first few (to me). I'm after the content "3. Escape". Complete blurb I'm looking at:

<h3 class="calibre_5"><span class="calibre3"><span class="bold"><a href="http://www.noname.org/forums/story/david-mcleod/thetranslator/3" class="calibre2"><span class="calibre_1"><span class="underline">3. Escape</span></span></a>
</span></span></h3>

It's grabbing all of the following instead of stopping at the first '<' after the content:

<h3 class="calibre_5"><span class="calibre3"><span class="bold"><a href="http://www.noname.org/forums/story/david-mcleod/thetranslator/3" class="calibre2"><span class="calibre_1"><span class="underline">3. Escape</span></span></a>

And the Replacement results in this. Note the </span>s and extra </h3>

<h3>3. Escape</h3>
</span></span></h3>

Thanks!

[theducks]

Why not use Diaps toolbag to delete spans with certain attributes (or naked) first to make things simpler?

[Brett Merkey]

The following regex seems to work fine with the code snippet you gave:

<h3.*?(\d+\. \w+).*?</h3>

All it does is look for a number, a period, a space and text between the h3 tags — everything else being ignored.

[meghane_e]

Originally Posted by theducks

Why not use Diaps toolbag to delete spans with certain attributes (or naked) first to make things simpler?

Thanks for quick feedback! Stupid question but if the tool you mention is the Plugins -> Edits Spans and Divs panel, how do I set it to strip nested spans? The ePub was generated from many Docx files (copied from web browser, files not all online anymore) converted to ePub files then merged into one ePub file. The point being it ends up with tons of nested spans and CSS auto-generated by the editor.

Maybe I'm having a more basic problem of what work flow to use here. I'm flip-flopping between whether the Editor is a better way to edit it, or should I write a script/program to transform the HTMZ first. Is there a thread that discusses which work flow methods work better in various conditions?

[theducks]

That is what it is called in Calibre (I also use it in Sigil)

Yes, it matches the tag pair for the conditions supplied.
It can be a bit tedious, but IMHO safer to do 1 condition at a time rather than do wild card (it will allow foot shooting )

span
style
calibre\d+ REGEX mode

[meghane_e]

Hmm, thank you for info. Not sure if I can use it right now, but sure know I'll probably need it later.

[JoeBloe]

A great site for testing Regex is regex101.com.
Try this:
(<h3.*>)(\w[^<]+)(<.+>)(</h3>)
The section [^<]+ tells it to take anything that is NOT a < whereas a . will keep going and take anything until it gets to the </tag> before the </h3>
Of course, you can do what you want with the grouping. I just use it on that site to better see if each section is doing what I want.

[meghane_e]

regex101.com looks like the most helpful test site I've come across! Good layout and usability! Thanks Joe! Still working my problem out though.

[meghane_e]

Thanks again everyone! Well, this gets me the content I want, at least on reg101.com and in the Editor:

Given code:

<h3 class="calibre_5"><span class="calibre3"><span class="bold">
<a href="http://www.noname.org/forums/story/david-mcleod/thetranslator/3" class="calibre2"><span class="calibre_1">
<span class="underline">3. Escape</span></span></a>
</span></span></h3>

The Find expression:
(<h3[^>]+>(<[^>]+>)+)([^<]+)(([^>]+>)+(.*<\/h3>))

Replace with \3

But that seems overly complicated?

Edit:Yay! (<h3.*>)(\w[^<]+)(<.+>)(<\/h3>) does work. There was a hidden newline in the generated code. Once I found it and took it out, it was happy