1 (edited by FlashT 2010-06-03 14:59)

Topic: [RESOLVED] Searching with accents...

Hello,
I just noticed that punbb search function has a problem with accents... i mean if in the forum there is a phrase with accents, and youre searching without them, it wont find the phrase... why didn't you just use fulltext search?
Best regards,
FlashT

Edit: Got it. Instructions on how to quick fix it, below.

Re: [RESOLVED] Searching with accents...

Thanks for the suggestion, we will think about it.

Re: [RESOLVED] Searching with accents...

Any suggestion how to fix it for now? As far as I understand search function generates some phrase indexes.... maybe just strip accents in generating that index?

4 (edited by FlashT 2010-06-03 14:56)

Re: [RESOLVED] Searching with accents...

Ok... a quickfix from my side:

Just remove accents when creating serach_words index and before processing search query with those functions:

function remove_accents($string) {
    if (seems_utf8($string)) {
        $chars = array(
        // Decompositions for Latin-1 Supplement
        chr(195).chr(128) => 'A', chr(195).chr(129) => 'A',
        chr(195).chr(130) => 'A', chr(195).chr(131) => 'A',
        chr(195).chr(132) => 'A', chr(195).chr(133) => 'A',
        chr(195).chr(135) => 'C', chr(195).chr(136) => 'E',
        chr(195).chr(137) => 'E', chr(195).chr(138) => 'E',
        chr(195).chr(139) => 'E', chr(195).chr(140) => 'I',
        chr(195).chr(141) => 'I', chr(195).chr(142) => 'I',
        chr(195).chr(143) => 'I', chr(195).chr(145) => 'N',
        chr(195).chr(146) => 'O', chr(195).chr(147) => 'O',
        chr(195).chr(148) => 'O', chr(195).chr(149) => 'O',
        chr(195).chr(150) => 'O', chr(195).chr(153) => 'U',
        chr(195).chr(154) => 'U', chr(195).chr(155) => 'U',
        chr(195).chr(156) => 'U', chr(195).chr(157) => 'Y',
        chr(195).chr(159) => 's', chr(195).chr(160) => 'a',
        chr(195).chr(161) => 'a', chr(195).chr(162) => 'a',
        chr(195).chr(163) => 'a', chr(195).chr(164) => 'a',
        chr(195).chr(165) => 'a', chr(195).chr(167) => 'c',
        chr(195).chr(168) => 'e', chr(195).chr(169) => 'e',
        chr(195).chr(170) => 'e', chr(195).chr(171) => 'e',
        chr(195).chr(172) => 'i', chr(195).chr(173) => 'i',
        chr(195).chr(174) => 'i', chr(195).chr(175) => 'i',
        chr(195).chr(177) => 'n', chr(195).chr(178) => 'o',
        chr(195).chr(179) => 'o', chr(195).chr(180) => 'o',
        chr(195).chr(181) => 'o', chr(195).chr(182) => 'o',
        chr(195).chr(182) => 'o', chr(195).chr(185) => 'u',
        chr(195).chr(186) => 'u', chr(195).chr(187) => 'u',
        chr(195).chr(188) => 'u', chr(195).chr(189) => 'y',
        chr(195).chr(191) => 'y',
        // Decompositions for Latin Extended-A
        chr(196).chr(128) => 'A', chr(196).chr(129) => 'a',
        chr(196).chr(130) => 'A', chr(196).chr(131) => 'a',
        chr(196).chr(132) => 'A', chr(196).chr(133) => 'a',
        chr(196).chr(134) => 'C', chr(196).chr(134) => 'c',
        chr(196).chr(136) => 'C', chr(196).chr(137) => 'c',
        chr(196).chr(138) => 'C', chr(196).chr(139) => 'c',
        chr(196).chr(140) => 'C', chr(196).chr(141) => 'c',
        chr(196).chr(142) => 'D', chr(196).chr(143) => 'd',
        chr(196).chr(144) => 'D', chr(196).chr(145) => 'd',
        chr(196).chr(146) => 'E', chr(196).chr(147) => 'e',
        chr(196).chr(148) => 'E', chr(196).chr(149) => 'e',
        chr(196).chr(150) => 'E', chr(196).chr(151) => 'e',
        chr(196).chr(152) => 'E', chr(196).chr(153) => 'e',
        chr(196).chr(154) => 'E', chr(196).chr(155) => 'e',
        chr(196).chr(156) => 'G', chr(196).chr(157) => 'g',
        chr(196).chr(158) => 'G', chr(196).chr(159) => 'g',
        chr(196).chr(160) => 'G', chr(196).chr(161) => 'g',
        chr(196).chr(162) => 'G', chr(196).chr(163) => 'g',
        chr(196).chr(164) => 'H', chr(196).chr(165) => 'h',
        chr(196).chr(166) => 'H', chr(196).chr(167) => 'h',
        chr(196).chr(168) => 'I', chr(196).chr(169) => 'i',
        chr(196).chr(170) => 'I', chr(196).chr(171) => 'i',
        chr(196).chr(172) => 'I', chr(196).chr(173) => 'i',
        chr(196).chr(174) => 'I', chr(196).chr(175) => 'i',
        chr(196).chr(176) => 'I', chr(196).chr(177) => 'i',
        chr(196).chr(178) => 'IJ',chr(196).chr(179) => 'ij',
        chr(196).chr(180) => 'J', chr(196).chr(181) => 'j',
        chr(196).chr(182) => 'K', chr(196).chr(183) => 'k',
        chr(196).chr(184) => 'k', chr(196).chr(185) => 'L',
        chr(196).chr(186) => 'l', chr(196).chr(187) => 'L',
        chr(196).chr(188) => 'l', chr(196).chr(189) => 'L',
        chr(196).chr(190) => 'l', chr(196).chr(191) => 'L',
        chr(197).chr(128) => 'l', chr(196).chr(129) => 'L',
        chr(197).chr(130) => 'l', chr(196).chr(131) => 'N',
        chr(197).chr(132) => 'n', chr(196).chr(133) => 'N',
        chr(197).chr(134) => 'n', chr(196).chr(135) => 'N',
        chr(197).chr(136) => 'n', chr(196).chr(137) => 'N',
        chr(197).chr(138) => 'n', chr(196).chr(139) => 'N',
        chr(197).chr(140) => 'O', chr(196).chr(141) => 'o',
        chr(197).chr(142) => 'O', chr(196).chr(143) => 'o',
        chr(197).chr(144) => 'O', chr(196).chr(145) => 'o',
        chr(197).chr(146) => 'OE',chr(197).chr(147) => 'oe',
        chr(197).chr(148) => 'R',chr(197).chr(149) => 'r',
        chr(197).chr(150) => 'R',chr(197).chr(151) => 'r',
        chr(197).chr(152) => 'R',chr(197).chr(153) => 'r',
        chr(197).chr(154) => 'S',chr(197).chr(155) => 's',
        chr(197).chr(156) => 'S',chr(197).chr(157) => 's',
        chr(197).chr(158) => 'S',chr(197).chr(159) => 's',
        chr(197).chr(160) => 'S', chr(197).chr(161) => 's',
        chr(197).chr(162) => 'T', chr(197).chr(163) => 't',
        chr(197).chr(164) => 'T', chr(197).chr(165) => 't',
        chr(197).chr(166) => 'T', chr(197).chr(167) => 't',
        chr(197).chr(168) => 'U', chr(197).chr(169) => 'u',
        chr(197).chr(170) => 'U', chr(197).chr(171) => 'u',
        chr(197).chr(172) => 'U', chr(197).chr(173) => 'u',
        chr(197).chr(174) => 'U', chr(197).chr(175) => 'u',
        chr(197).chr(176) => 'U', chr(197).chr(177) => 'u',
        chr(197).chr(178) => 'U', chr(197).chr(179) => 'u',
        chr(197).chr(180) => 'W', chr(197).chr(181) => 'w',
        chr(197).chr(182) => 'Y', chr(197).chr(183) => 'y',
        chr(197).chr(184) => 'Y', chr(197).chr(185) => 'Z',
        chr(197).chr(186) => 'z', chr(197).chr(187) => 'Z',
        chr(197).chr(188) => 'z', chr(197).chr(189) => 'Z',
        chr(197).chr(190) => 'z', chr(197).chr(191) => 's',
        // Euro Sign
        chr(226).chr(130).chr(172) => 'E');
       
        $string = strtr($string, $chars);
    } else {
        // Assume ISO-8859-1 if not UTF-8
        $chars['in'] = chr(128).chr(131).chr(138).chr(142).chr(154).chr(158)
            .chr(159).chr(162).chr(165).chr(181).chr(192).chr(193).chr(194)
            .chr(195).chr(196).chr(197).chr(199).chr(200).chr(201).chr(202)
            .chr(203).chr(204).chr(205).chr(206).chr(207).chr(209).chr(210)
            .chr(211).chr(212).chr(213).chr(214).chr(216).chr(217).chr(218)
            .chr(219).chr(220).chr(221).chr(224).chr(225).chr(226).chr(227)
            .chr(228).chr(229).chr(231).chr(232).chr(233).chr(234).chr(235)
            .chr(236).chr(237).chr(238).chr(239).chr(241).chr(242).chr(243)
            .chr(244).chr(245).chr(246).chr(248).chr(249).chr(250).chr(251)
            .chr(252).chr(253).chr(255);

        $chars['out'] = "EfSZszYcYuAAAAAACEEEEIIIINOOOOOOUUUUYaaaaaaceeeeiiiinoooooouuuuyy";

        $string = strtr($string, $chars['in'], $chars['out']);
        $double_chars['in'] = array(chr(140), chr(156), chr(198), chr(208), chr(222), chr(223), chr(230), chr(240), chr(254));
        $double_chars['out'] = array('OE', 'oe', 'AE', 'DH', 'TH', 'ss', 'ae', 'dh', 'th');
        $string = str_replace($double_chars['in'], $double_chars['out'], $string);
    }

    return $string;
}

function seems_utf8($Str) { # by bmorel at ssi dot fr
    for ($i=0; $i<strlen($Str); $i++) {
        if (ord($Str[$i]) < 0x80) continue; # 0bbbbbbb
        elseif ((ord($Str[$i]) & 0xE0) == 0xC0) $n=1; # 110bbbbb
        elseif ((ord($Str[$i]) & 0xF0) == 0xE0) $n=2; # 1110bbbb
        elseif ((ord($Str[$i]) & 0xF8) == 0xF0) $n=3; # 11110bbb
        elseif ((ord($Str[$i]) & 0xFC) == 0xF8) $n=4; # 111110bb
        elseif ((ord($Str[$i]) & 0xFE) == 0xFC) $n=5; # 1111110b
        else return false; # Does not match any model
        for ($j=0; $j<$n; $j++) { # n bytes matching 10bbbbbb follow ?
            if ((++$i == strlen($Str)) || ((ord($Str[$i]) & 0xC0) != 0x80))
            return false;
        }
    }
    return true;
}

Instructions:

1. Add above functions to include/functions.php
2. In file include/search_idx.php, function split_words, before "return $words" add this line:

$words = array_map('remove_accents',$words);

3. In search.php, find line "if ($action == 'search')" and below you should have "$keywords = (isset($_GET['keywords'])) ? forum_trim($_GET['keywords']) : null;";
relace it with:

$keywords = (isset($_GET['keywords'])) ? forum_trim(remove_accents($_GET['keywords'])) : null;

4. Rebuild search index in admin panel, to make it work on already posted posts.

Works like a charm...

The idea is to deaccentize all words in search index and search them without accents even when user search terms contains them.

The bad side: well... since fulltext serach does the same, it should do a job.

PS. PunBB already has some function to deaccentize words (used in friendly urls creation) but can't find it so I added my own