Regex in Regex function mode

[lindlind]

Doing a search and replace like the one below in Regex mode works perfectly fine:

Search: </span>\s*<br/>\s*
Replace: xx

But when doing the same thing in Regex function, nothing happens:

Search: </span>\s*<br/>\s*
Function: return match.group().replace('</span>\s*<br/>\s*', 'xx')

It seems as if everything works fine if I do not use \s or quantifiers in the function. How can I do this right?

[kovidgoyal]

Function: return 'xx'

[lindlind]

I think I was not very clear, it tried to make it as simple as possible, showing that I can search for the string in Regex mode but not in Regex function mode. This just made it impossible to understand what I need...

The initial search is actually wider (<div.*?</div>), and I will do many replacements within the function, one of them is this:

string.replace('</span>\s*<br/>\s*', 'xx')

But it seems I cannot use \s or quantifiers in the function. Is there a way to do this?

[lindlind]

additional note: I guess I should use something like re.sub instead, but is that possible?

[lomkiri]

Originally Posted by lindlind

string.replace('</span>\s*<br/>\s*', 'xx')
But it seems I cannot use \s or quantifiers in the function. Is there a way to do this?

\s, \s* etc. are part of the regex language, the python str.replace() does not understand them.
You'll have to import re or regex and use sub or match :

import regex
string = regex.sub(r'</span>\s*<br/>\s*', 'xx', string)

(calibre uses mainly the module regex, but you can get the help from the python module "re", regex is a superset of re : https://docs.python.org/3/library/re.html. You can also use re, if you prefer)

--------------

Edit (I haven't seen your last message when I posted this one) :

additional note: I guess I should use something like re.sub instead, but is that possible?

Yes, it is possible. You can import in the function any module that is part of the calibre package, and re and regex are in this case.
You may see my other message of today to get a practical exemple

[lindlind]

Thanks a million!