CODEWITHSUNDEEP
jqueryregexunicodehighlightdiacritics
Regexme
Regexme

Reputation: 95

accent insensitive regex

My code:

jQuery.fn.extend({
 highlight: function(search){
  var regex = new RegExp('(<[^>]*>)|('+ search.replace(/[.+]i/,"$0") +')','ig');

  return this.html(this.html().replace(regex, function(a, b, c){
   return (a.charAt(0) == '<') ? a : '<strong class="highlight">' + c + '</strong>';
  }));
 }

});

I want to highlight letters with accents, ie:

$('body').highlight("cao");

should highlight: [ção] OR [çÃo] OR [cáo] OR expre[cão]tion OR [Cáo]tion

How can I do that?

Upvotes: 9

Views: 2681

Answers (2)

tchrist
tchrist

Reputation: 80384

The sole correct way to do this is to first run it through Unicode Normalization Form D, canonical decomposition.

You then strip our any Marks that result (\pM characters, or perhaps \p{Diacritic}, depending), and run your match against the de/un-marked version.

Do not under any circumstances hardcode a bunch of literals. Eek!

Boa sorte!

Upvotes: 5

casablanca
casablanca

Reputation: 70701

You need to come up with a table of alternative characters and dynamically generate a regex based on that. For example:

var alt = {
  'c': '[cCç]',
  'a': '[aAãÃá]',
  /* etc. */
};

highlight: function (search) {
  var pattern = '';
  for (var i = 0; i < search.length; i++) {
    var ch = search[i];
    if (alt.hasOwnProperty(ch))
      pattern += alt[ch];
    else
      pattern += ch;
  }

  ...
}

Then for search = 'cao' this will generate a pattern [cCç][aAãÃá]o.

Upvotes: 3

Related Questions