Extract acronyms patterns from string using regex

Question

I have this problem:

list_= ["blabla S.P.A words J.R words. , words","words words !! words s.r.l words. D.T. words","words words I.B.M. words words."]

I would like to have:

['S.P.A', 'J.R']
['s.r.l', 'D.T.']
['I.B.M.']

I found this amazing solution Finding Acronyms Using Regex In Python that returns:

['S.P.', 'J.']
['s.r.', 'D.T.']
['I.B.M.']

How can I use that solution in my situation?

Thank you

what about `(?<= )(\w+\.[^ ]+)` ? – KC. Oct 06 '18 at 02:00 — KC., Oct 06 '18 at 02:00

CertainPerformance · Answer 1 · 2018-10-05T06:28:23.380

3

You just need to make the final period optional. Also lookbehind for a space or the start of the string before the first letter to ensure it's not part of another word, and lookahead after the end for a space or the end of the string:

pattern = r'(?i)(?:^|(?<= ))(?:[a-z]\.)+[a-z]\.?(?= |$)'

edited Oct 05 '18 at 06:28

answered Oct 05 '18 at 05:46

CertainPerformance

356,069
52
309
320

OP didn't have any digits in his input, but if there are, just change the `\w`s to `[a-zA-Z]` – CertainPerformance Oct 05 '18 at 06:04
1

but you know what acronyms are letters, better to make them letters. +1 – The Scientific Method Oct 05 '18 at 06:06
1

it's also extracting `A.B.C` from `AAA.B.C` – Jean-François Fabre Oct 05 '18 at 06:07
@Jean-FrançoisFabre oh that is great how spot it. – The Scientific Method Oct 05 '18 at 06:10
1

@TheScientificMethod That's what the `(?i)` is for, figured it'd be nicer than repeating `[a-zA-z]` twice – CertainPerformance Oct 05 '18 at 06:13
thank you for your answer! I tried this code on this case "A.TT.L", I do not know if it is an acronym (I hope not, because I think I do not want it :) ) In this case returns "A.T" – frquestions Oct 05 '18 at 06:25
@frquestions Add lookahead after the end for a space or the end of the string `(?= |$)` – CertainPerformance Oct 05 '18 at 06:28
also tell OP to use `flags=re.I` or your pattern won't match uppercased letters. – Jean-François Fabre Oct 05 '18 at 06:36
@Jean-FrançoisFabre That's what the `(?i)` is for – CertainPerformance Oct 05 '18 at 06:36
there is still a small bug: A.TT.L => A.TTL – frquestions Oct 05 '18 at 13:53
@frquestions I don't think those count because there's more than one letter in between. If you *want* to allow more than one letter in between, use `[a-z]+` instead of `[a-z]`. – CertainPerformance Oct 05 '18 at 23:47

Extract acronyms patterns from string using regex

1 Answers1

Linked