Regex for this?

Question

These two are is probably fairly simple, but I could only get a regex for parts of what I want. 
 
 I'm looking for something handles the chapter numbers and the section numbers. 
 For example "10－18" or "5－140". Up to digits on the left and up to three on the right. 
 Next are things like "6.", "6.2", "5.11.4". I can't get this one without making winding up with separate expressions. Something like "5.11.4" ends up giving me 3 or four different candidates. 
 Any help would be appreciated. 
 
 Best regards, Keenan

Paul Filkin · Accepted Answer

Hi Keenan Cooper 
 ok - I tested in Studio and created a segmentation rule that does this when the file is opened in Studio: 
 
 Now you can what Jerzy suggested and filter out the numbers only segments which is a lot cleaner for you when translating. The rules I used were these... I guess you could make it cleaner but this works using the rules I previously suggested: 
 Before break: 
 ^(?:(?:\d+\.\d+\.\d+)|(?:\d+\.\d+)|(?:\d+\.)) 
 Break characters: 
 Tab 
 After break: 
 Anything 
 This is based on my assumption of your file of course. This is the example file I used: 
 segmentation.docx

Paul Filkin · Answer

Hi Keenan Cooper 
 In fact looking at this you could also achieve this: 
 
 Now the work is even easier. Just create another segmentation rule on the TM like this: 
 Before break: 
 Anything 
 Break characters: 
 Tab 
 After break: 
 (\d{1,2}－\d{1,3}) 
 That's it... hope this helps anyway. I do think your segmentation rules is going to be the best solution here.

Paul Filkin · Answer

Hello Keenan Cooper , 
 
What do you want the regex for? Search and replace, regex autosuggest, verification, terminjector etc. 
 
Regards 
 
Paul

Paul Filkin · Answer

Hi Keenan Cooper , 
 
ok, the first one is probably this: 
 
\d{1,2}－\d{1,3} 
 
The second group of numbers I think you'd be better off excluding from Studio altogether with some segmentation rules. I haven't tested in a Studio TM yet, but this would pick them up with regex autosuggest I think: 
 
(?:\d+\.\d+\.\d+)|(?:\d+\.\d+)|(?:\d+\.) 
 
The tricky part is that you might always get multiple suggestions if they are in the same segment. But in principle, without testing, this is where I'd start. 
 
Regards 
 
Paul

Paul Filkin · Answer

Keenan Cooper said: or have them all show up using ctrl+, 
 This part will only ever work when Studio offers the ability to customise the number recognition patterns. I don't believe this is on the table yet.

Paul Filkin · Answer

Hi Keenan Cooper 
 
Sorry, I think I forgot an important part for regex autosuggest (I'm not at my desk)... 
 
(\d{1,2}－\d{1,3}) 
 
((?:\d+\.\d+\.\d+)|(?:\d+\.\d+)|(?:\d+\.)) 
 
Then $0 in the target 
 
Regards 
 
Paul

Paul Filkin · Answer

So for this file I guess this would work too as the before break segmentation rule: 
 
^[0-9.]+

Paul Filkin · Answer

Hi Keenan Cooper , If you use the shortened rule I suggested then it will segment no matter how long your chapter number gets: ^[0-9.]+ 
 The only reason I suggested the other one was to try and make this more specific for the regex autosuggest provider. But even this without the ^ would probably work for the autosuggest solution too like this: 
 [0-9.]+ 
 But segmentation and filtering out is far better.

Trados Studio > 5. Regex and XPath

Regex for this?