Some clarification with RegEx

[beic]

Let's say I wanted to replace the string 'Sh' with the letter 'F' throughout a book:

<p class="chaptitle">The Shop</p>

<p class="indent">This should show the hashes.</p>

End result

<p class="chaptitle">The Fop</p>

<p class="indent">This fould fow the hafes.</p>

In order for it not to mess up any code, I assume I need to isolate it to within the <p> tags?

<p class="[\d\D]*?">[\d\D]*?</p>

This above matches all paragraphs regardless of class. Is this correct so far?

How would I then match all instances of 'Sh' within the <p> tags and what would the replacement text expression look like?

Thanks

[eschwartz]

See here where I recently discussed this: https://www.mobileread.com/forums/sho...18#post3360318

[beic]

Originally Posted by eschwartz

See here where I recently discussed this: https://www.mobileread.com/forums/sho...18#post3360318

(?<=>[^<]*)s\B(?=[^>]*<)

Thanks, I think that's what I'm looking for but when I paste it into the RegEx builder, the box goes red as if it's invalid. I also tried DreamWeaver's find and replace and it says RegEx syntax Error: Invalid Quantifier. This site returns the following error: "Lookbehinds need to be zero-width, thus quantifiers are not allowed".

[eschwartz]

Well, not all regex engines support variable-length lookbehinds. But I know for a fact that calibre's Editor does, because it was calibre that I tested it with!

Furthermore, calibre does not have a "RegEx builder", it has a single input box and it doesn't go red when there is an error -- it pops up an error dialog.
So clearly you did not try this regex using calibre's Editor, and given that this is the subforum for the calibre Editor, I cannot be expected to read your mind and know which piece of software you decided to use instead (without telling me).

So... try using calibre's Editor.

[beic]

Thank you! I was trying to run the regex in the conversion Search & Replace tool, not the book editor

In the end I used a Regex-Function because I had a lot of characters to replace.

Past this into the find box so it will only replace text within HTML tags

>[^<>]+<

Then run a simple replace Regex-Function

def replace(match, number, file_name, metadata, dictionaries, data, functions, *args, **kwargs):
    return match.group().replace('CH', 'Ċ').replace('GH', 'Ġ').replace('BH', 'Ḃ').replace('DH', 'Ḋ').replace('FH', 'Ḟ').replace('MH', 'Ṁ').replace('PH', 'Ṗ').replace('SH', 'Ṡ').replace('TH', 'Ṫ').replace('Ch', 'Ċ').replace('Gh', 'Ġ').replace('Bh', 'Ḃ').replace('Dh', 'Ḋ').replace('Fh', 'Ḟ').replace('Mh', 'Ṁ').replace('Ph', 'Ṗ').replace('Sh', 'Ṡ').replace('Th', 'Ṫ').replace('ch', 'ċ').replace('gh', 'ġ').replace('bh', 'ḃ').replace('dh', 'ḋ').replace('fh', 'ḟ').replace('mh', 'ṁ').replace('ph', 'ṗ').replace('sh', 'ṡ').replace('th', 'ṫ').replace('&', '⁊')

Hopefully this helps someone else.

[eschwartz]

Well, that might do it.

I have never actually used the S&R tool, because it makes more sense to use the Editor (or another Editor, before calibre had one). But for all I know it uses the builtin python stdlib re() instead of Matthew Barnett's enhanced regex module.
Anyway, I cannot guarantee how something might work if I don't use it and don't know that that is what the other guy is using.

...

Feel free to post your regex function to the relevant sticky in this subforum, if you feel other people might find it useful and want to make it more likely they will find it.

[kovidgoyal]

The conversion S&R uses the python re module. The regex module did not exist when it was written