0

I'm trying to remove repeated chars from strings like I looooovvee this. It's awesomee. Very gooood.

to an output like: I love this. It's awesome, Very good.

I'm already using in PHP this instrunction $str=preg_replace("/(.)\1+/", "$1", $str);

But it outputs I love this. It's awesome. Very god.

The problem is in words that already should have repeated chars like 'good' or 'cool'

user1201779
  • 35
  • 5
  • 2
    you just need to create an AI that will proof read for you and determine which word needs to be used, based on the context. – Tyler Collins Mar 28 '16 at 17:10
  • 1
    @TylerCollins it is not called AI, it is called logic. – Evan Carslake Mar 28 '16 at 17:12
  • Perhaps some sort of spell checking library? – ryantxr Mar 28 '16 at 17:13
  • I actually feel that this is on par with [obscenity filters](http://blog.codinghorror.com/obscenity-filters-bad-idea-or-incredibly-intercoursing-bad-idea/) and the nightmares that come with those too. – sjagr Mar 28 '16 at 17:15

2 Answers2

0

Since you started with regex you could continue with regex.

Use negative look-behind & negative look-ahead to check if before or after the repeating letter there's anything that you wouldn't want to match.

$str = preg_replace('/(?<!g|c)(\w)\1+(?!d|l)/', '$1', $str);

Unfortunately this implies writing a list of possible prefixes and suffixes.

Mihai Stancu
  • 15,848
  • 2
  • 33
  • 51
0

I suppose you could store your allowed words (like "good" and "cool") in a Trie Dictionary.

Whenever you are check a word for duplicate chars, you should allow duplicate characters upto the point where the Dictionary still have some valid words with that prefix.

When the dictionary has no valid words for the prefix, you can remove the duplicate chars from that point on

Eg: if the word you are checking is "Goooood"

check "Go" in trie it will return "God" and "Good" as valid words
check "Goo" in trie, it will return "Good" as the valid word
check "Gooo" in trie, it will say there are no valid words

Therefore you keep upto "Goo" and remove the rest of the o's

abhaybhatia
  • 599
  • 1
  • 7
  • 16