How to create spaces between characters only for acronysms in Python

Question

I am trying to add spaces between characters only for acronyms (all consecutive all-caps words) in Python.

INPUT:

"The PNUD, UN, UCALP and USA and U N."

DESIRED OUTPUT:

"The P N U D, U N, U C A L P and U S A and U N."

I have this solution so far, but I am looking for something more efficient/elegant:

import re
data = "The PNUD, UN, UCALP and USA and U N."
result = re.sub(r'(?=(?!^)[^[a-z]|s+|W]*)', ' ', data)
result = re.sub(r's+(W)', 'g<1>', result)
print(result)

Asked By: Esteban Gorupicz

||

Source

Answer 1

I think the following regex is a lot more trivial solution for this problem

re.sub('([A-Z])(?=[A-Z])', '\1 ', s)

I’m just using a positive lookahead and a backreference.

Answered By: aa_nador

Answer 2

Another solution re.sub with lambda function:

import re

data = "The PNUD, UN, UCALP and USA and U N."

result = re.sub(r"b[A-Z]{2,}b", lambda g: " ".join(g.group(0)), data)
print(result)

Prints:

The P N U D, U N, U C A L P and U S A and U N.

EDIT: Small benchmark

import re
import regex
from timeit import timeit


pat1 = re.compile(r"b[A-Z]{2,}b")
pat2 = re.compile(r"([A-Z])(?=[A-Z])")
pat3 = re.compile(r"[A-Z](?=[A-Z])")  # the same without capturing group

# using regex module instead of re
pat4 = regex.compile(r"b[[:upper:]]{2,}b")


data = "The PNUD, UN, UCALP and USA and U N."


def fn1():
    return pat1.sub(lambda g: " ".join(g.group(0)), data)


def fn2():
    return pat2.sub(r"g<1> ", data)


def fn3():
    return pat3.sub(r"g<0> ", data)


def fn4():
    return pat4.sub(lambda g: " ".join(g.group(0)), data)


t1 = timeit(fn1, number=10_000)
t2 = timeit(fn2, number=10_000)
t3 = timeit(fn3, number=10_000)
t4 = timeit(fn4, number=10_000)

print(t1)
print(t2)
print(t3)
print(t4)

Prints:

0.03805400081910193
0.10581987909972668
0.10386284696869552
0.044628452975302935

Answered By: Andrej Kesely

Answer 3

You can use a single call to re.sub and match a single uppercase char and assert another one to the right.

In the replacement use the match followed by a space using g<0>

[A-Z](?=[A-Z])

Regex demo

Example

result = re.sub('[A-Z](?=[A-Z])', r'g<0> ', data)

Answered By: The fourth bird

Answer 4

Using Lambda with regex instead of re is slightly slower but allows you to match also Unicode chars from non-English languages, which is caeteris paribus a more generalizable answer.

import re
import regex
from timeit import timeit

pat1 = re.compile(r"b[A-Z]+b")
pat2 = re.compile(r"([A-Z])(?=[A-Z])")
pat3 = regex.compile(r"b[[:upper:]]+b")

data = "The PNUD, UN, UCALP and USA and U N."

def fn1():
    return pat1.sub(lambda g: " ".join(g.group(0)), data)

def fn2():
    return pat2.sub(r"g<1> ", data)

def fn3():
    return pat3.sub(lambda g: " ".join(g.group(0)), data)


t1 = timeit(fn1, number=10_000)
t2 = timeit(fn2, number=10_000)
t3 = timeit(fn3, number=10_000)

print(fn1())
print(fn2())
print(fn3())
print(t1)
print(t2)
print(t3)

DEMO

Answered By: Esteban Gorupicz

How to create spaces between characters only for acronysms in Python

Question:

Answers: