I think the redundant second and third rules you used will cause you a problem here anyway even if you do need to extract this.

Do you understand correctly the role of square brackets in regular expressions?
It doesn't mean a sequence of characters! It means a single character from those listed inside the brackets...

So to me it looks like your closing rules are just plain incorrect.

Evzen Polenka said:

Do you understand correctly the role of square brackets in regular expressions?

Absolutely not! :-) I''m trying to learn and understand  from the help I get.

However, it seems to work quite well, which might be strange. 

How should I define a sequence of characters, and still more important question, where can I learn the secrets of opening and closing patterns? 

Evzen could you tell me what is the character for "OR", please? 

Thanks

As a cheat-sheet, I’m uploading a mini-guide in PDF.

Nora Díaz has posted some articles about regex, with good examples for Trados Studio:

http://noradiaz.blogspot.com/

Paul’s blog https://multifarious.filkin.com/ has some articles as well about regex.

☛  I’d recommend you to test first your regular expressions in sites like http://regexstorm.net/tester. It’s faster and let you learn in real-time as you change the regex and see the matches below. Try it!

PDF

Sophie Martin

You really should learn some basic regex.  There are hundreds of sites on the internet with tutorials and help.  Try starting with these as they are really relevant to translators:

http://noradiaz.blogspot.com/

Or this for more detail:

https://www.regular-expressions.info/

On your specific question.  You basically have two ways. You can use a pipe symbol, so a|b means match b if a fails.  Or you could use ab inside square brackets like this [ab] which means match a single character from within the list ab.

Essentially providing an OR type of operation.

If you mean inside square brackets, there is no OR.

[abc] means a or b or c

Jesús Prieto said:

If you mean inside square brackets, there is no OR.

[abc] means a or b or c

???
Isn't the word between the 'a', 'b' and 'c' in your sentence the word "or"? ;-)

Yes they were ORs. There is not a regex character meaning OR inside square backets, just like the character “|” in other places.

Sophie Martin

Thank you for the files you just sent me.  In my opinion you really need to emply a localization engineer to prepare these files for you properly.  Most of the content seems to be html chunks embedded into the php and a localization engineer would be able to prepare these files for translation using the html filetype, and what's left with a simple regex based filetype, then put them back together again afterwards.

If you try to do this all with a regex based filetype you're asking for problems. You can get there... so using these for the structure for example (and this is incomplete):

And then with some inline rules to exclude stuff you don't want... maybe this sort of thing for example:

This starts to take some shape, but it needs you to understand what you're doing and build on this.

But really... you should employ a localization engineer to manage this process for you as the results will be more reliable and they'll scale.