How can I detect laughing words in a string?

try with this pattern:

\b(?:a*(?:ha)+h?|(?:l+o+)+l+)\b

or better if your regex flavour support atomic groups and possessive quantifiers:

\b(?>a*+(?:ha)++h?|(?:l+o+)++l+)\b

\b(a*ha+h[ha]*|o?l+o+l+[ol]*)\b

Matches:

hahahah
haha
lol
loll
loool
looooool
lolololol
lolololololo
ahaha
aaaahahahahahaha

Does not match:

looo
oool
oooo
llll
ha
l
o
lo
ol
ah
aah
aha
kill
lala
haunt
hauha
louol

In Python, I tried to do it in this way:

import re

re.sub(r"\b(?:a{0,2}h{1,2}a{0,2}){2,}h?\b", "<laugh>", "hahahahha! I love laughing")

>> <laugh>! I love laughing

To keep it simple, because the solutions posted may be overly complicated for what you want to do: if the only thing you count as "laughing words" are ha, haha, etc. and lol, lolol, lololol, etc., then the following regular expression will be sufficient:

\b(ha)+|l(ol)+\b

This assumes a regex dialect in which \b represents a word boundary, which you seem to be using.