Remove accents/diacritics in a string in JavaScript

前端 未结 29 2532
轻奢々
轻奢々 2020-11-21 13:29

How do I remove accentuated characters from a string? Especially in IE6, I had something like this:

accentsTidy = function(s){
    var r=s.toLowerCase();
           


        
相关标签:
29条回答
  • 2020-11-21 14:04

    I have forked billy's code http://jsfiddle.net/billybraga/UHmnf/ (from his post) into this: http://jsfiddle.net/infralabs/dJX58/

    I corrected transcription of ſ and ß characters and also added coversion of these ones: Þþ, Ðð, Ŋŋ, IJij, Œœ.

    The modified snippet is below:

    var defaultDiacriticsRemovalMap = [{
        'base': "A",
            'letters': /(A|Ⓐ|A|À|Á|Â|Ầ|Ấ|Ẫ|Ẩ|Ã|Ā|Ă|Ằ|Ắ|Ẵ|Ẳ|Ȧ|Ǡ|Ä|Ǟ|Ả|Å|Ǻ|Ǎ|Ȁ|Ȃ|Ạ|Ậ|Ặ|Ḁ|Ą|Ⱥ|Ɐ|[\u0041\u24B6\uFF21\u00C0\u00C1\u00C2\u1EA6\u1EA4\u1EAA\u1EA8\u00C3\u0100\u0102\u1EB0\u1EAE\u1EB4\u1EB2\u0226\u01E0\u00C4\u01DE\u1EA2\u00C5\u01FA\u01CD\u0200\u0202\u1EA0\u1EAC\u1EB6\u1E00\u0104\u023A\u2C6F])/g
    }, {
        'base': "AA",
            'letters': /(Ꜳ|[\uA732])/g
    }, {
        'base': "AE",
            'letters': /(Æ|Ǽ|Ǣ|[\u00C6\u01FC\u01E2])/g
    }, {
        'base': "AO",
            'letters': /(Ꜵ|[\uA734])/g
    }, {
        'base': "AU",
            'letters': /(Ꜷ|[\uA736])/g
    }, {
        'base': "AV",
            'letters': /(Ꜹ|Ꜻ|[\uA738\uA73A])/g
    }, {
        'base': "AY",
            'letters': /(Ꜽ|[\uA73C])/g
    }, {
        'base': "B",
            'letters': /(B|Ⓑ|B|Ḃ|Ḅ|Ḇ|Ƀ|Ƃ|Ɓ|[\u0042\u24B7\uFF22\u1E02\u1E04\u1E06\u0243\u0182\u0181])/g
    }, {
        'base': "C",
            'letters': /(C|Ⓒ|C|Ć|Ĉ|Ċ|Č|Ç|Ḉ|Ƈ|Ȼ|Ꜿ|[\u0043\u24B8\uFF23\u0106\u0108\u010A\u010C\u00C7\u1E08\u0187\u023B\uA73E])/g
    }, {
        'base': "D",
            'letters': /(D|Ⓓ|D|Ḋ|Ď|Ḍ|Ḑ|Ḓ|Ḏ|Đ|Ƌ|Ɗ|Ɖ|Ꝺ|Ð|[\u0044\u24B9\uFF24\u1E0A\u010E\u1E0C\u1E10\u1E12\u1E0E\u0110\u018B\u018A\u0189\uA779\u00D0])/g
    }, {
        'base': "DZ",
            'letters': /(DZ|DŽ|[\u01F1\u01C4])/g
    }, {
        'base': "Dz",
            'letters': /(Dz|Dž|[\u01F2\u01C5])/g
    }, {
        'base': "E",
            'letters': /(E|Ⓔ|E|È|É|Ê|Ề|Ế|Ễ|Ể|Ẽ|Ē|Ḕ|Ḗ|Ĕ|Ė|Ë|Ẻ|Ě|Ȅ|Ȇ|Ẹ|Ệ|Ȩ|Ḝ|Ę|Ḙ|Ḛ|Ɛ|Ǝ|[\u0045\u24BA\uFF25\u00C8\u00C9\u00CA\u1EC0\u1EBE\u1EC4\u1EC2\u1EBC\u0112\u1E14\u1E16\u0114\u0116\u00CB\u1EBA\u011A\u0204\u0206\u1EB8\u1EC6\u0228\u1E1C\u0118\u1E18\u1E1A\u0190\u018E])/g
    }, {
        'base': "F",
            'letters': /(F|Ⓕ|F|Ḟ|Ƒ|Ꝼ|[\u0046\u24BB\uFF26\u1E1E\u0191\uA77B])/g
    }, {
        'base': "G",
            'letters': /(G|Ⓖ|G|Ǵ|Ĝ|Ḡ|Ğ|Ġ|Ǧ|Ģ|Ǥ|Ɠ|Ꞡ|Ᵹ|Ꝿ|[\u0047\u24BC\uFF27\u01F4\u011C\u1E20\u011E\u0120\u01E6\u0122\u01E4\u0193\uA7A0\uA77D\uA77E])/g
    }, {
        'base': "H",
            'letters': /(H|Ⓗ|H|Ĥ|Ḣ|Ḧ|Ȟ|Ḥ|Ḩ|Ḫ|Ħ|Ⱨ|Ⱶ|Ɥ|[\u0048\u24BD\uFF28\u0124\u1E22\u1E26\u021E\u1E24\u1E28\u1E2A\u0126\u2C67\u2C75\uA78D])/g
    }, {
        'base': "I",
            'letters': /(I|Ⓘ|I|Ì|Í|Î|Ĩ|Ī|Ĭ|İ|Ï|Ḯ|Ỉ|Ǐ|Ȉ|Ȋ|Ị|Į|Ḭ|Ɨ|[\u0049\u24BE\uFF29\u00CC\u00CD\u00CE\u0128\u012A\u012C\u0130\u00CF\u1E2E\u1EC8\u01CF\u0208\u020A\u1ECA\u012E\u1E2C\u0197])/g
    }, {
        'base': "J",
            'letters': /(J|Ⓙ|J|Ĵ|Ɉ|[\u004A\u24BF\uFF2A\u0134\u0248])/g
    }, {
        'base': "K",
            'letters': /(K|Ⓚ|K|Ḱ|Ǩ|Ḳ|Ķ|Ḵ|Ƙ|Ⱪ|Ꝁ|Ꝃ|Ꝅ|Ꞣ|[\u004B\u24C0\uFF2B\u1E30\u01E8\u1E32\u0136\u1E34\u0198\u2C69\uA740\uA742\uA744\uA7A2])/g
    }, {
        'base': "L",
            'letters': /(L|Ⓛ|L|Ŀ|Ĺ|Ľ|Ḷ|Ḹ|Ļ|Ḽ|Ḻ|Ł|Ƚ|Ɫ|Ⱡ|Ꝉ|Ꝇ|Ꞁ|[\u004C\u24C1\uFF2C\u013F\u0139\u013D\u1E36\u1E38\u013B\u1E3C\u1E3A\u0141\u023D\u2C62\u2C60\uA748\uA746\uA780])/g
    }, {
        'base': "LJ",
            'letters': /(LJ|[\u01C7])/g
    }, {
        'base': "Lj",
            'letters': /(Lj|[\u01C8])/g
    }, {
        'base': "M",
            'letters': /(M|Ⓜ|M|Ḿ|Ṁ|Ṃ|Ɱ|Ɯ|[\u004D\u24C2\uFF2D\u1E3E\u1E40\u1E42\u2C6E\u019C])/g
    }, {
        'base': "N",
            'letters': /(N|Ⓝ|N|Ǹ|Ń|Ñ|Ṅ|Ň|Ṇ|Ņ|Ṋ|Ṉ|Ƞ|Ɲ|Ꞑ|Ꞥ|Ŋ|[\u004E\u24C3\uFF2E\u01F8\u0143\u00D1\u1E44\u0147\u1E46\u0145\u1E4A\u1E48\u0220\u019D\uA790\uA7A4\u014A])/g
    }, {
        'base': "NJ",
            'letters': /(NJ|[\u01CA])/g
    }, {
        'base': "Nj",
            'letters': /(Nj|[\u01CB])/g
    }, {
        'base': "O",
            'letters': /(O|Ⓞ|O|Ò|Ó|Ô|Ồ|Ố|Ỗ|Ổ|Õ|Ṍ|Ȭ|Ṏ|Ō|Ṑ|Ṓ|Ŏ|Ȯ|Ȱ|Ö|Ȫ|Ỏ|Ő|Ǒ|Ȍ|Ȏ|Ơ|Ờ|Ớ|Ỡ|Ở|Ợ|Ọ|Ộ|Ǫ|Ǭ|Ø|Ǿ|Ɔ|Ɵ|Ꝋ|Ꝍ|[\u004F\u24C4\uFF2F\u00D2\u00D3\u00D4\u1ED2\u1ED0\u1ED6\u1ED4\u00D5\u1E4C\u022C\u1E4E\u014C\u1E50\u1E52\u014E\u022E\u0230\u00D6\u022A\u1ECE\u0150\u01D1\u020C\u020E\u01A0\u1EDC\u1EDA\u1EE0\u1EDE\u1EE2\u1ECC\u1ED8\u01EA\u01EC\u00D8\u01FE\u0186\u019F\uA74A\uA74C])/g
    }, {
        'base': "OE",
            'letters': /(Œ|[\u0152])/g
    }, {
        'base': "OI",
            'letters': /(Ƣ|[\u01A2])/g
    }, {
        'base': "OO",
            'letters': /(Ꝏ|[\uA74E])/g
    }, {
        'base': "OU",
            'letters': /(Ȣ|[\u0222])/g
    }, {
        'base': "P",
            'letters': /(P|Ⓟ|P|Ṕ|Ṗ|Ƥ|Ᵽ|Ꝑ|Ꝓ|Ꝕ|[\u0050\u24C5\uFF30\u1E54\u1E56\u01A4\u2C63\uA750\uA752\uA754])/g
    }, {
        'base': "Q",
            'letters': /(Q|Ⓠ|Q|Ꝗ|Ꝙ|Ɋ|[\u0051\u24C6\uFF31\uA756\uA758\u024A])/g
    }, {
        'base': "R",
            'letters': /(R|Ⓡ|R|Ŕ|Ṙ|Ř|Ȑ|Ȓ|Ṛ|Ṝ|Ŗ|Ṟ|Ɍ|Ɽ|Ꝛ|Ꞧ|Ꞃ|[\u0052\u24C7\uFF32\u0154\u1E58\u0158\u0210\u0212\u1E5A\u1E5C\u0156\u1E5E\u024C\u2C64\uA75A\uA7A6\uA782])/g
    }, {
        'base': "S",
            'letters': /(S|Ⓢ|S|ẞ|Ś|Ṥ|Ŝ|Ṡ|Š|Ṧ|Ṣ|Ṩ|Ș|Ş|Ȿ|Ꞩ|Ꞅ|[\u0053\u24C8\uFF33\u1E9E\u015A\u1E64\u015C\u1E60\u0160\u1E66\u1E62\u1E68\u0218\u015E\u2C7E\uA7A8\uA784])/g
    }, {
        'base': "T",
            'letters': /(T|Ⓣ|T|Ṫ|Ť|Ṭ|Ț|Ţ|Ṱ|Ṯ|Ŧ|Ƭ|Ʈ|Ⱦ|Ꞇ|[\u0054\u24C9\uFF34\u1E6A\u0164\u1E6C\u021A\u0162\u1E70\u1E6E\u0166\u01AC\u01AE\u023E\uA786])/g
    }, {
        'base': "TH",
            'letters': /(Þ|[\u00DE])/g
    }, {
        'base': "TZ",
            'letters': /(Ꜩ|[\uA728])/g
    }, {
        'base': "U",
            'letters': /(U|Ⓤ|U|Ù|Ú|Û|Ũ|Ṹ|Ū|Ṻ|Ŭ|Ü|Ǜ|Ǘ|Ǖ|Ǚ|Ủ|Ů|Ű|Ǔ|Ȕ|Ȗ|Ư|Ừ|Ứ|Ữ|Ử|Ự|Ụ|Ṳ|Ų|Ṷ|Ṵ|Ʉ|[\u0055\u24CA\uFF35\u00D9\u00DA\u00DB\u0168\u1E78\u016A\u1E7A\u016C\u00DC\u01DB\u01D7\u01D5\u01D9\u1EE6\u016E\u0170\u01D3\u0214\u0216\u01AF\u1EEA\u1EE8\u1EEE\u1EEC\u1EF0\u1EE4\u1E72\u0172\u1E76\u1E74\u0244])/g
    }, {
        'base': "V",
            'letters': /(V|Ⓥ|V|Ṽ|Ṿ|Ʋ|Ꝟ|Ʌ|[\u0056\u24CB\uFF36\u1E7C\u1E7E\u01B2\uA75E\u0245])/g
    }, {
        'base': "VY",
            'letters': /(Ꝡ|[\uA760])/g
    }, {
        'base': "W",
            'letters': /(W|Ⓦ|W|Ẁ|Ẃ|Ŵ|Ẇ|Ẅ|Ẉ|Ⱳ|[\u0057\u24CC\uFF37\u1E80\u1E82\u0174\u1E86\u1E84\u1E88\u2C72])/g
    }, {
        'base': "X",
            'letters': /(X|Ⓧ|X|Ẋ|Ẍ|[\u0058\u24CD\uFF38\u1E8A\u1E8C])/g
    }, {
        'base': "Y",
            'letters': /(Y|Ⓨ|Y|Ỳ|Ý|Ŷ|Ỹ|Ȳ|Ẏ|Ÿ|Ỷ|Ỵ|Ƴ|Ɏ|Ỿ|[\u0059\u24CE\uFF39\u1EF2\u00DD\u0176\u1EF8\u0232\u1E8E\u0178\u1EF6\u1EF4\u01B3\u024E\u1EFE])/g
    }, {
        'base': "Z",
            'letters': /(Z|Ⓩ|Z|Ź|Ẑ|Ż|Ž|Ẓ|Ẕ|Ƶ|Ȥ|Ɀ|Ⱬ|Ꝣ|[\u005A\u24CF\uFF3A\u0179\u1E90\u017B\u017D\u1E92\u1E94\u01B5\u0224\u2C7F\u2C6B\uA762])/g
    }, {
        'base': "a",
            'letters': /(a|ⓐ|a|ẚ|à|á|â|ầ|ấ|ẫ|ẩ|ã|ā|ă|ằ|ắ|ẵ|ẳ|ȧ|ǡ|ä|ǟ|ả|å|ǻ|ǎ|ȁ|ȃ|ạ|ậ|ặ|ḁ|ą|ⱥ|ɐ|[\u0061\u24D0\uFF41\u1E9A\u00E0\u00E1\u00E2\u1EA7\u1EA5\u1EAB\u1EA9\u00E3\u0101\u0103\u1EB1\u1EAF\u1EB5\u1EB3\u0227\u01E1\u00E4\u01DF\u1EA3\u00E5\u01FB\u01CE\u0201\u0203\u1EA1\u1EAD\u1EB7\u1E01\u0105\u2C65\u0250])/g
    }, {
        'base': "aa",
            'letters': /(ꜳ|[\uA733])/g
    }, {
        'base': "ae",
            'letters': /(æ|ǽ|ǣ|[\u00E6\u01FD\u01E3])/g
    }, {
        'base': "ao",
            'letters': /(ꜵ|[\uA735])/g
    }, {
        'base': "au",
            'letters': /(ꜷ|[\uA737])/g
    }, {
        'base': "av",
            'letters': /(ꜹ|ꜻ|[\uA739\uA73B])/g
    }, {
        'base': "ay",
            'letters': /(ꜽ|[\uA73D])/g
    }, {
        'base': "b",
            'letters': /(b|ⓑ|b|ḃ|ḅ|ḇ|ƀ|ƃ|ɓ|[\u0062\u24D1\uFF42\u1E03\u1E05\u1E07\u0180\u0183\u0253])/g
    }, {
        'base': "c",
            'letters': /(c|ⓒ|c|ć|ĉ|ċ|č|ç|ḉ|ƈ|ȼ|ꜿ|ↄ|[\u0063\u24D2\uFF43\u0107\u0109\u010B\u010D\u00E7\u1E09\u0188\u023C\uA73F\u2184])/g
    }, {
        'base': "d",
            'letters': /(d|ⓓ|d|ḋ|ď|ḍ|ḑ|ḓ|ḏ|đ|ƌ|ɖ|ɗ|ꝺ|ð|[\u0064\u24D3\uFF44\u1E0B\u010F\u1E0D\u1E11\u1E13\u1E0F\u0111\u018C\u0256\u0257\uA77A\u00F0])/g
    }, {
        'base': "dz",
            'letters': /(dz|dž|[\u01F3\u01C6])/g
    }, {
        'base': "e",
            'letters': /(e|ⓔ|e|è|é|ê|ề|ế|ễ|ể|ẽ|ē|ḕ|ḗ|ĕ|ė|ë|ẻ|ě|ȅ|ȇ|ẹ|ệ|ȩ|ḝ|ę|ḙ|ḛ|ɇ|ɛ|ǝ|[\u0065\u24D4\uFF45\u00E8\u00E9\u00EA\u1EC1\u1EBF\u1EC5\u1EC3\u1EBD\u0113\u1E15\u1E17\u0115\u0117\u00EB\u1EBB\u011B\u0205\u0207\u1EB9\u1EC7\u0229\u1E1D\u0119\u1E19\u1E1B\u0247\u025B\u01DD])/g
    }, {
        'base': "f",
            'letters': /(f|ⓕ|f|ḟ|ƒ|ꝼ|[\u0066\u24D5\uFF46\u1E1F\u0192\uA77C])/g
    }, {
        'base': "g",
            'letters': /(g|ⓖ|g|ǵ|ĝ|ḡ|ğ|ġ|ǧ|ģ|ǥ|ɠ|ꞡ|ᵹ|ꝿ|[\u0067\u24D6\uFF47\u01F5\u011D\u1E21\u011F\u0121\u01E7\u0123\u01E5\u0260\uA7A1\u1D79\uA77F])/g
    }, {
        'base': "h",
            'letters': /(h|ⓗ|h|ĥ|ḣ|ḧ|ȟ|ḥ|ḩ|ḫ|ẖ|ħ|ⱨ|ⱶ|ɥ|[\u0068\u24D7\uFF48\u0125\u1E23\u1E27\u021F\u1E25\u1E29\u1E2B\u1E96\u0127\u2C68\u2C76\u0265])/g
    }, {
        'base': "hv",
            'letters': /(ƕ|[\u0195])/g
    }, {
        'base': "i",
            'letters': /(i|ⓘ|i|ì|í|î|ĩ|ī|ĭ|ï|ḯ|ỉ|ǐ|ȉ|ȋ|ị|į|ḭ|ɨ|ı|[\u0069\u24D8\uFF49\u00EC\u00ED\u00EE\u0129\u012B\u012D\u00EF\u1E2F\u1EC9\u01D0\u0209\u020B\u1ECB\u012F\u1E2D\u0268\u0131])/g
    }, {
        'base': "ij",
            'letters': /(ij|[\u0133])/g
    }, {
        'base': "j",
            'letters': /(j|ⓙ|j|ĵ|ǰ|ɉ|[\u006A\u24D9\uFF4A\u0135\u01F0\u0249])/g
    }, {
        'base': "k",
            'letters': /(k|ⓚ|k|ḱ|ǩ|ḳ|ķ|ḵ|ƙ|ⱪ|ꝁ|ꝃ|ꝅ|ꞣ|[\u006B\u24DA\uFF4B\u1E31\u01E9\u1E33\u0137\u1E35\u0199\u2C6A\uA741\uA743\uA745\uA7A3])/g
    }, {
        'base': "l",
            'letters': /(l|ⓛ|l|ŀ|ĺ|ľ|ḷ|ḹ|ļ|ḽ|ḻ|ł|ƚ|ɫ|ⱡ|ꝉ|ꞁ|ꝇ|[\u006C\u24DB\uFF4C\u0140\u013A\u013E\u1E37\u1E39\u013C\u1E3D\u1E3B\u0142\u019A\u026B\u2C61\uA749\uA781\uA747])/g
    }, {
        'base': "lj",
            'letters': /(lj|[\u01C9])/g
    }, {
        'base': "m",
            'letters': /(m|ⓜ|m|ḿ|ṁ|ṃ|ɱ|ɯ|[\u006D\u24DC\uFF4D\u1E3F\u1E41\u1E43\u0271\u026F])/g
    }, {
        'base': "n",
            'letters': /(n|ⓝ|n|ǹ|ń|ñ|ṅ|ň|ṇ|ņ|ṋ|ṉ|ƞ|ɲ|ʼn|ꞑ|ꞥ|ŋ|[\u006E\u24DD\uFF4E\u01F9\u0144\u00F1\u1E45\u0148\u1E47\u0146\u1E4B\u1E49\u019E\u0272\u0149\uA791\uA7A5\u014B])/g
    }, {
        'base': "nj",
            'letters': /(nj|[\u01CC])/g
    }, {
        'base': "o",
            'letters': /(o|ⓞ|o|ò|ó|ô|ồ|ố|ỗ|ổ|õ|ṍ|ȭ|ṏ|ō|ṑ|ṓ|ŏ|ȯ|ȱ|ö|ȫ|ỏ|ő|ǒ|ȍ|ȏ|ơ|ờ|ớ|ỡ|ở|ợ|ọ|ộ|ǫ|ǭ|ø|ǿ|ɔ|ꝋ|ꝍ|ɵ|[\u006F\u24DE\uFF4F\u00F2\u00F3\u00F4\u1ED3\u1ED1\u1ED7\u1ED5\u00F5\u1E4D\u022D\u1E4F\u014D\u1E51\u1E53\u014F\u022F\u0231\u00F6\u022B\u1ECF\u0151\u01D2\u020D\u020F\u01A1\u1EDD\u1EDB\u1EE1\u1EDF\u1EE3\u1ECD\u1ED9\u01EB\u01ED\u00F8\u01FF\u0254\uA74B\uA74D\u0275])/g
    }, {
        'base': "oe",
            'letters': /(œ|[\u0153])/g
    }, {
        'base': "oi",
            'letters': /(ƣ|[\u01A3])/g
    }, {
        'base': "ou",
            'letters': /(ȣ|[\u0223])/g
    }, {
        'base': "oo",
            'letters': /(ꝏ|[\uA74F])/g
    }, {
        'base': "p",
            'letters': /(p|ⓟ|p|ṕ|ṗ|ƥ|ᵽ|ꝑ|ꝓ|ꝕ|[\u0070\u24DF\uFF50\u1E55\u1E57\u01A5\u1D7D\uA751\uA753\uA755])/g
    }, {
        'base': "q",
            'letters': /(q|ⓠ|q|ɋ|ꝗ|ꝙ|[\u0071\u24E0\uFF51\u024B\uA757\uA759])/g
    }, {
        'base': "r",
            'letters': /(r|ⓡ|r|ŕ|ṙ|ř|ȑ|ȓ|ṛ|ṝ|ŗ|ṟ|ɍ|ɽ|ꝛ|ꞧ|ꞃ|[\u0072\u24E1\uFF52\u0155\u1E59\u0159\u0211\u0213\u1E5B\u1E5D\u0157\u1E5F\u024D\u027D\uA75B\uA7A7\uA783])/g
    }, {
        'base': "s",
            'letters': /(s|ⓢ|s|ś|ṥ|ŝ|ṡ|š|ṧ|ṣ|ṩ|ș|ş|ȿ|ꞩ|ꞅ|ẛ|ſ|[\u0073\u24E2\uFF53\u015B\u1E65\u015D\u1E61\u0161\u1E67\u1E63\u1E69\u0219\u015F\u023F\uA7A9\uA785\u1E9B\u017F])/g
    }, {
        'base': "ss",
            'letters': /(ß|[\u00DF])/g
    }, {
        'base': "t",
            'letters': /(t|ⓣ|t|ṫ|ẗ|ť|ṭ|ț|ţ|ṱ|ṯ|ŧ|ƭ|ʈ|ⱦ|ꞇ|[\u0074\u24E3\uFF54\u1E6B\u1E97\u0165\u1E6D\u021B\u0163\u1E71\u1E6F\u0167\u01AD\u0288\u2C66\uA787])/g
    }, {
        'base': "th",
            'letters': /(þ|[\u00FE])/g
    }, {
        'base': "tz",
            'letters': /(ꜩ|[\uA729])/g
    }, {
        'base': "u",
            'letters': /(u|ⓤ|u|ù|ú|û|ũ|ṹ|ū|ṻ|ŭ|ü|ǜ|ǘ|ǖ|ǚ|ủ|ů|ű|ǔ|ȕ|ȗ|ư|ừ|ứ|ữ|ử|ự|ụ|ṳ|ų|ṷ|ṵ|ʉ|[\u0075\u24E4\uFF55\u00F9\u00FA\u00FB\u0169\u1E79\u016B\u1E7B\u016D\u00FC\u01DC\u01D8\u01D6\u01DA\u1EE7\u016F\u0171\u01D4\u0215\u0217\u01B0\u1EEB\u1EE9\u1EEF\u1EED\u1EF1\u1EE5\u1E73\u0173\u1E77\u1E75\u0289])/g
    }, {
        'base': "v",
            'letters': /(v|ⓥ|v|ṽ|ṿ|ʋ|ꝟ|ʌ|[\u0076\u24E5\uFF56\u1E7D\u1E7F\u028B\uA75F\u028C])/g
    }, {
        'base': "vy",
            'letters': /(ꝡ|[\uA761])/g
    }, {
        'base': "w",
            'letters': /(w|ⓦ|w|ẁ|ẃ|ŵ|ẇ|ẅ|ẘ|ẉ|ⱳ|[\u0077\u24E6\uFF57\u1E81\u1E83\u0175\u1E87\u1E85\u1E98\u1E89\u2C73])/g
    }, {
        'base': "x",
            'letters': /(x|ⓧ|x|ẋ|ẍ|[\u0078\u24E7\uFF58\u1E8B\u1E8D])/g
    }, {
        'base': "y",
            'letters': /(y|ⓨ|y|ỳ|ý|ŷ|ỹ|ȳ|ẏ|ÿ|ỷ|ẙ|ỵ|ƴ|ɏ|ỿ|[\u0079\u24E8\uFF59\u1EF3\u00FD\u0177\u1EF9\u0233\u1E8F\u00FF\u1EF7\u1E99\u1EF5\u01B4\u024F\u1EFF])/g
    }, {
        'base': "z",
            'letters': /(z|ⓩ|z|ź|ẑ|ż|ž|ẓ|ẕ|ƶ|ȥ|ɀ|ⱬ|ꝣ|[\u007A\u24E9\uFF5A\u017A\u1E91\u017C\u017E\u1E93\u1E95\u01B6\u0225\u0240\u2C6C\uA763])/g
    }];
    
    0 讨论(0)
  • 2020-11-21 14:05

    tested fastest way to remove diacritics from the string that covers almost all characters is this. found highest performance among all the other methods

    Diacritics Map

    var diacriticsMap = {
    '\u0041': 'A','\u24B6': 'A','\uFF21': 'A','\u00C0': 'A','\u00C1': 'A','\u00C2': 'A','\u1EA6': 'A','\u1EA4': 'A','\u1EAA': 'A','\u1EA8': 'A',
    '\u00C3': 'A','\u0100': 'A','\u0102': 'A','\u1EB0': 'A','\u1EAE': 'A','\u1EB4': 'A','\u1EB2': 'A','\u0226': 'A','\u01E0': 'A','\u00C4': 'A',
    '\u01DE': 'A','\u1EA2': 'A','\u00C5': 'A','\u01FA': 'A','\u01CD': 'A','\u0200': 'A','\u0202': 'A','\u1EA0': 'A','\u1EAC': 'A','\u1EB6': 'A',
    '\u1E00': 'A','\u0104': 'A','\u023A': 'A','\u2C6F': 'A',
    
    '\uA732': 'AA',
    '\u00C6': 'AE','\u01FC': 'AE','\u01E2': 'AE',
    '\uA734': 'AO',
    '\uA736': 'AU',
    '\uA738': 'AV','\uA73A': 'AV',
    '\uA73C': 'AY',
    '\u0042': 'B','\u24B7': 'B','\uFF22': 'B','\u1E02': 'B','\u1E04': 'B','\u1E06': 'B','\u0243': 'B','\u0182': 'B','\u0181': 'B',
    
    '\u0043': 'C','\u24B8': 'C','\uFF23': 'C','\u0106': 'C','\u0108': 'C','\u010A': 'C','\u010C': 'C','\u00C7': 'C','\u1E08': 'C','\u0187': 'C',
    '\u023B': 'C','\uA73E': 'C',
    
    '\u0044': 'D','\u24B9': 'D','\uFF24': 'D','\u1E0A': 'D','\u010E': 'D','\u1E0C': 'D','\u1E10': 'D','\u1E12': 'D','\u1E0E': 'D','\u0110': 'D',
    '\u018B': 'D','\u018A': 'D','\u0189': 'D','\uA779': 'D',
    
    '\u01F1': 'DZ','\u01C4': 'DZ',
    '\u01F2': 'Dz','\u01C5': 'Dz',
    
    '\u0045': 'E','\u24BA': 'E','\uFF25': 'E','\u00C8': 'E','\u00C9': 'E','\u00CA': 'E','\u1EC0': 'E','\u1EBE': 'E','\u1EC4': 'E','\u1EC2': 'E',
    '\u1EBC': 'E','\u0112': 'E','\u1E14': 'E','\u1E16': 'E','\u0114': 'E','\u0116': 'E','\u00CB': 'E','\u1EBA': 'E','\u011A': 'E','\u0204': 'E',
    '\u0206': 'E','\u1EB8': 'E','\u1EC6': 'E','\u0228': 'E','\u1E1C': 'E','\u0118': 'E','\u1E18': 'E','\u1E1A': 'E','\u0190': 'E','\u018E': 'E',
    
    '\u0046': 'F','\u24BB': 'F','\uFF26': 'F','\u1E1E': 'F','\u0191': 'F','\uA77B': 'F',
    
    '\u0047': 'G','\u24BC': 'G','\uFF27': 'G','\u01F4': 'G','\u011C': 'G','\u1E20': 'G','\u011E': 'G','\u0120': 'G','\u01E6': 'G','\u0122': 'G',
    '\u01E4': 'G','\u0193': 'G','\uA7A0': 'G','\uA77D': 'G','\uA77E': 'G',
    
    '\u0048': 'H','\u24BD': 'H','\uFF28': 'H','\u0124': 'H','\u1E22': 'H','\u1E26': 'H','\u021E': 'H','\u1E24': 'H','\u1E28': 'H','\u1E2A': 'H',
    '\u0126': 'H','\u2C67': 'H','\u2C75': 'H','\uA78D': 'H',
    
    '\u0049': 'I','\u24BE': 'I','\uFF29': 'I','\u00CC': 'I','\u00CD': 'I','\u00CE': 'I','\u0128': 'I','\u012A': 'I','\u012C': 'I','\u0130': 'I',
    '\u00CF': 'I','\u1E2E': 'I','\u1EC8': 'I','\u01CF': 'I','\u0208': 'I','\u020A': 'I','\u1ECA': 'I','\u012E': 'I','\u1E2C': 'I','\u0197': 'I',
    
    '\u004A': 'J','\u24BF': 'J','\uFF2A': 'J','\u0134': 'J','\u0248': 'J',
    
    '\u004B': 'K','\u24C0': 'K','\uFF2B': 'K','\u1E30': 'K','\u01E8': 'K','\u1E32': 'K','\u0136': 'K','\u1E34': 'K','\u0198': 'K','\u2C69': 'K',
    '\uA740': 'K','\uA742': 'K','\uA744': 'K','\uA7A2': 'K',
    
    '\u004C': 'L','\u24C1': 'L','\uFF2C': 'L','\u013F': 'L','\u0139': 'L','\u013D': 'L','\u1E36': 'L','\u1E38': 'L','\u013B': 'L','\u1E3C': 'L',
    '\u1E3A': 'L','\u0141': 'L','\u023D': 'L','\u2C62': 'L','\u2C60': 'L','\uA748': 'L','\uA746': 'L','\uA780': 'L',
    
    '\u01C7': 'LJ',
    '\u01C8': 'Lj',
    '\u004D': 'M','\u24C2': 'M','\uFF2D': 'M','\u1E3E': 'M','\u1E40': 'M','\u1E42': 'M','\u2C6E': 'M','\u019C': 'M',
    
    '\u004E': 'N','\u24C3': 'N','\uFF2E': 'N','\u01F8': 'N','\u0143': 'N','\u00D1': 'N','\u1E44': 'N','\u0147': 'N','\u1E46': 'N','\u0145': 'N',
    '\u1E4A': 'N','\u1E48': 'N','\u0220': 'N','\u019D': 'N','\uA790': 'N','\uA7A4': 'N',
    
    '\u01CA': 'NJ',
    '\u01CB': 'Nj',
    
    '\u004F': 'O','\u24C4': 'O','\uFF2F': 'O','\u00D2': 'O','\u00D3': 'O','\u00D4': 'O','\u1ED2': 'O','\u1ED0': 'O','\u1ED6': 'O','\u1ED4': 'O',
    '\u00D5': 'O','\u1E4C': 'O','\u022C': 'O','\u1E4E': 'O','\u014C': 'O','\u1E50': 'O','\u1E52': 'O','\u014E': 'O','\u022E': 'O','\u0230': 'O',
    '\u00D6': 'O','\u022A': 'O','\u1ECE': 'O','\u0150': 'O','\u01D1': 'O','\u020C': 'O','\u020E': 'O','\u01A0': 'O','\u1EDC': 'O','\u1EDA': 'O',
    '\u1EE0': 'O','\u1EDE': 'O','\u1EE2': 'O','\u1ECC': 'O','\u1ED8': 'O','\u01EA': 'O','\u01EC': 'O','\u00D8': 'O','\u01FE': 'O','\u0186': 'O',
    '\u019F': 'O','\uA74A': 'O','\uA74C': 'O',
    
    '\u01A2': 'OI',
    '\uA74E': 'OO',
    '\u0222': 'OU',
    '\u0050': 'P','\u24C5': 'P','\uFF30': 'P','\u1E54': 'P','\u1E56': 'P','\u01A4': 'P','\u2C63': 'P','\uA750': 'P','\uA752': 'P','\uA754': 'P',
    '\u0051': 'Q','\u24C6': 'Q','\uFF31': 'Q','\uA756': 'Q','\uA758': 'Q','\u024A': 'Q',
    
    '\u0052': 'R','\u24C7': 'R','\uFF32': 'R','\u0154': 'R','\u1E58': 'R','\u0158': 'R','\u0210': 'R','\u0212': 'R','\u1E5A': 'R','\u1E5C': 'R',
    '\u0156': 'R','\u1E5E': 'R','\u024C': 'R','\u2C64': 'R','\uA75A': 'R','\uA7A6': 'R','\uA782': 'R',
    
    '\u0053': 'S','\u24C8': 'S','\uFF33': 'S','\u1E9E': 'S','\u015A': 'S','\u1E64': 'S','\u015C': 'S','\u1E60': 'S','\u0160': 'S','\u1E66': 'S',
    '\u1E62': 'S','\u1E68': 'S','\u0218': 'S','\u015E': 'S','\u2C7E': 'S','\uA7A8': 'S','\uA784': 'S',
    
    '\u0054': 'T','\u24C9': 'T','\uFF34': 'T','\u1E6A': 'T','\u0164': 'T','\u1E6C': 'T','\u021A': 'T','\u0162': 'T','\u1E70': 'T','\u1E6E': 'T',
    '\u0166': 'T','\u01AC': 'T','\u01AE': 'T','\u023E': 'T','\uA786': 'T',
    
    '\uA728': 'TZ',
    
    '\u0055': 'U','\u24CA': 'U','\uFF35': 'U','\u00D9': 'U','\u00DA': 'U','\u00DB': 'U','\u0168': 'U','\u1E78': 'U','\u016A': 'U','\u1E7A': 'U',
    '\u016C': 'U','\u00DC': 'U','\u01DB': 'U','\u01D7': 'U','\u01D5': 'U','\u01D9': 'U','\u1EE6': 'U','\u016E': 'U','\u0170': 'U','\u01D3': 'U',
    '\u0214': 'U','\u0216': 'U','\u01AF': 'U','\u1EEA': 'U','\u1EE8': 'U','\u1EEE': 'U','\u1EEC': 'U','\u1EF0': 'U','\u1EE4': 'U','\u1E72': 'U',
    '\u0172': 'U','\u1E76': 'U','\u1E74': 'U','\u0244': 'U',
    
    '\u0056': 'V','\u24CB': 'V','\uFF36': 'V','\u1E7C': 'V','\u1E7E': 'V','\u01B2': 'V','\uA75E': 'V','\u0245': 'V',
    '\uA760': 'VY',
    '\u0057': 'W','\u24CC': 'W','\uFF37': 'W','\u1E80': 'W','\u1E82': 'W','\u0174': 'W','\u1E86': 'W','\u1E84': 'W','\u1E88': 'W','\u2C72': 'W',
    '\u0058': 'X','\u24CD': 'X','\uFF38': 'X','\u1E8A': 'X','\u1E8C': 'X',
    
    '\u0059': 'Y','\u24CE': 'Y','\uFF39': 'Y','\u1EF2': 'Y','\u00DD': 'Y','\u0176': 'Y','\u1EF8': 'Y','\u0232': 'Y','\u1E8E': 'Y','\u0178': 'Y',
    '\u1EF6': 'Y','\u1EF4': 'Y','\u01B3': 'Y','\u024E': 'Y','\u1EFE': 'Y',
    
    '\u005A': 'Z','\u24CF': 'Z','\uFF3A': 'Z','\u0179': 'Z','\u1E90': 'Z','\u017B': 'Z','\u017D': 'Z','\u1E92': 'Z','\u1E94': 'Z','\u01B5': 'Z',
    '\u0224': 'Z','\u2C7F': 'Z','\u2C6B': 'Z','\uA762': 'Z',
    
    '\u0061': 'a','\u24D0': 'a','\uFF41': 'a','\u1E9A': 'a','\u00E0': 'a','\u00E1': 'a','\u00E2': 'a','\u1EA7': 'a','\u1EA5': 'a','\u1EAB': 'a',
    '\u1EA9': 'a','\u00E3': 'a','\u0101': 'a','\u0103': 'a','\u1EB1': 'a','\u1EAF': 'a','\u1EB5': 'a','\u1EB3': 'a','\u0227': 'a','\u01E1': 'a',
    '\u00E4': 'a','\u01DF': 'a','\u1EA3': 'a','\u00E5': 'a','\u01FB': 'a','\u01CE': 'a','\u0201': 'a','\u0203': 'a','\u1EA1': 'a','\u1EAD': 'a',
    '\u1EB7': 'a','\u1E01': 'a','\u0105': 'a','\u2C65': 'a','\u0250': 'a',
    
    '\uA733': 'aa',
    '\u00E6': 'ae','\u01FD': 'ae','\u01E3': 'ae',
    '\uA735': 'ao',
    '\uA737': 'au',
    '\uA739': 'av','\uA73B': 'av',
    '\uA73D': 'ay',
    '\u0062': 'b','\u24D1': 'b','\uFF42': 'b','\u1E03': 'b','\u1E05': 'b','\u1E07': 'b','\u0180': 'b','\u0183': 'b','\u0253': 'b',
    
    '\u0063': 'c','\u24D2': 'c','\uFF43': 'c','\u0107': 'c','\u0109': 'c','\u010B': 'c','\u010D': 'c','\u00E7': 'c','\u1E09': 'c','\u0188': 'c',
    '\u023C': 'c','\uA73F': 'c','\u2184': 'c',
    
    '\u0064': 'd','\u24D3': 'd','\uFF44': 'd','\u1E0B': 'd','\u010F': 'd','\u1E0D': 'd','\u1E11': 'd','\u1E13': 'd','\u1E0F': 'd','\u0111': 'd',
    '\u018C': 'd','\u0256': 'd','\u0257': 'd','\uA77A': 'd',
    
    '\u01F3': 'dz','\u01C6': 'dz',
    
    '\u0065': 'e','\u24D4': 'e','\uFF45': 'e','\u00E8': 'e','\u00E9': 'e','\u00EA': 'e','\u1EC1': 'e','\u1EBF': 'e','\u1EC5': 'e','\u1EC3': 'e',
    '\u1EBD': 'e','\u0113': 'e','\u1E15': 'e','\u1E17': 'e','\u0115': 'e','\u0117': 'e','\u00EB': 'e','\u1EBB': 'e','\u011B': 'e','\u0205': 'e',
    '\u0207': 'e','\u1EB9': 'e','\u1EC7': 'e','\u0229': 'e','\u1E1D': 'e','\u0119': 'e','\u1E19': 'e','\u1E1B': 'e','\u0247': 'e','\u025B': 'e',
    '\u01DD': 'e',
    
    '\u0066': 'f','\u24D5': 'f','\uFF46': 'f','\u1E1F': 'f','\u0192': 'f','\uA77C': 'f',
    
    '\u0067': 'g','\u24D6': 'g','\uFF47': 'g','\u01F5': 'g','\u011D': 'g','\u1E21': 'g','\u011F': 'g','\u0121': 'g','\u01E7': 'g','\u0123': 'g',
    '\u01E5': 'g','\u0260': 'g','\uA7A1': 'g','\u1D79': 'g','\uA77F': 'g',
    
    '\u0068': 'h','\u24D7': 'h','\uFF48': 'h','\u0125': 'h','\u1E23': 'h','\u1E27': 'h','\u021F': 'h','\u1E25': 'h','\u1E29': 'h','\u1E2B': 'h',
    '\u1E96': 'h','\u0127': 'h','\u2C68': 'h','\u2C76': 'h','\u0265': 'h',
    
    '\u0195': 'hv',
    
    '\u0069': 'i','\u24D8': 'i','\uFF49': 'i','\u00EC': 'i','\u00ED': 'i','\u00EE': 'i','\u0129': 'i','\u012B': 'i','\u012D': 'i','\u00EF': 'i',
    '\u1E2F': 'i','\u1EC9': 'i','\u01D0': 'i','\u0209': 'i','\u020B': 'i','\u1ECB': 'i','\u012F': 'i','\u1E2D': 'i','\u0268': 'i','\u0131': 'i',
    
    '\u006A': 'j','\u24D9': 'j','\uFF4A': 'j','\u0135': 'j','\u01F0': 'j','\u0249': 'j',
    
    '\u006B': 'k','\u24DA': 'k','\uFF4B': 'k','\u1E31': 'k','\u01E9': 'k','\u1E33': 'k','\u0137': 'k','\u1E35': 'k','\u0199': 'k','\u2C6A': 'k',
    '\uA741': 'k','\uA743': 'k','\uA745': 'k','\uA7A3': 'k',
    
    '\u006C': 'l','\u24DB': 'l','\uFF4C': 'l','\u0140': 'l','\u013A': 'l','\u013E': 'l','\u1E37': 'l','\u1E39': 'l','\u013C': 'l','\u1E3D': 'l',
    '\u1E3B': 'l','\u017F': 'l','\u0142': 'l','\u019A': 'l','\u026B': 'l','\u2C61': 'l','\uA749': 'l','\uA781': 'l','\uA747': 'l',
    
    '\u01C9': 'lj',
    '\u006D': 'm','\u24DC': 'm','\uFF4D': 'm','\u1E3F': 'm','\u1E41': 'm','\u1E43': 'm','\u0271': 'm','\u026F': 'm',
    
    '\u006E': 'n','\u24DD': 'n','\uFF4E': 'n','\u01F9': 'n','\u0144': 'n','\u00F1': 'n','\u1E45': 'n','\u0148': 'n','\u1E47': 'n','\u0146': 'n',
    '\u1E4B': 'n','\u1E49': 'n','\u019E': 'n','\u0272': 'n','\u0149': 'n','\uA791': 'n','\uA7A5': 'n',
    
    '\u01CC': 'nj',
    
    '\u006F': 'o','\u24DE': 'o','\uFF4F': 'o','\u00F2': 'o','\u00F3': 'o','\u00F4': 'o','\u1ED3': 'o','\u1ED1': 'o','\u1ED7': 'o','\u1ED5': 'o',
    '\u00F5': 'o','\u1E4D': 'o','\u022D': 'o','\u1E4F': 'o','\u014D': 'o','\u1E51': 'o','\u1E53': 'o','\u014F': 'o','\u022F': 'o','\u0231': 'o',
    '\u00F6': 'o','\u022B': 'o','\u1ECF': 'o','\u0151': 'o','\u01D2': 'o','\u020D': 'o','\u020F': 'o','\u01A1': 'o','\u1EDD': 'o','\u1EDB': 'o',
    '\u1EE1': 'o','\u1EDF': 'o','\u1EE3': 'o','\u1ECD': 'o','\u1ED9': 'o','\u01EB': 'o','\u01ED': 'o','\u00F8': 'o','\u01FF': 'o','\u0254': 'o',
    '\uA74B': 'o','\uA74D': 'o','\u0275': 'o',
    
    '\u01A3': 'oi',
    '\u0223': 'ou',
    '\uA74F': 'oo',
    '\u0070': 'p','\u24DF': 'p','\uFF50': 'p','\u1E55': 'p','\u1E57': 'p','\u01A5': 'p','\u1D7D': 'p','\uA751': 'p','\uA753': 'p','\uA755': 'p',
    '\u0071': 'q','\u24E0': 'q','\uFF51': 'q','\u024B': 'q','\uA757': 'q','\uA759': 'q',
    
    '\u0072': 'r','\u24E1': 'r','\uFF52': 'r','\u0155': 'r','\u1E59': 'r','\u0159': 'r','\u0211': 'r','\u0213': 'r','\u1E5B': 'r','\u1E5D': 'r',
    '\u0157': 'r','\u1E5F': 'r','\u024D': 'r','\u027D': 'r','\uA75B': 'r','\uA7A7': 'r','\uA783': 'r',
    
    '\u0073': 's','\u24E2': 's','\uFF53': 's','\u00DF': 's','\u015B': 's','\u1E65': 's','\u015D': 's','\u1E61': 's','\u0161': 's','\u1E67': 's',
    '\u1E63': 's','\u1E69': 's','\u0219': 's','\u015F': 's','\u023F': 's','\uA7A9': 's','\uA785': 's','\u1E9B': 's',
    
    '\u0074': 't','\u24E3': 't','\uFF54': 't','\u1E6B': 't','\u1E97': 't','\u0165': 't','\u1E6D': 't','\u021B': 't','\u0163': 't','\u1E71': 't',
    '\u1E6F': 't','\u0167': 't','\u01AD': 't','\u0288': 't','\u2C66': 't','\uA787': 't',
    
    '\uA729': 'tz',
    
    '\u0075': 'u','\u24E4': 'u','\uFF55': 'u','\u00F9': 'u','\u00FA': 'u','\u00FB': 'u','\u0169': 'u','\u1E79': 'u','\u016B': 'u','\u1E7B': 'u',
    '\u016D': 'u','\u00FC': 'u','\u01DC': 'u','\u01D8': 'u','\u01D6': 'u','\u01DA': 'u','\u1EE7': 'u','\u016F': 'u','\u0171': 'u','\u01D4': 'u',
    '\u0215': 'u','\u0217': 'u','\u01B0': 'u','\u1EEB': 'u','\u1EE9': 'u','\u1EEF': 'u','\u1EED': 'u','\u1EF1': 'u','\u1EE5': 'u','\u1E73': 'u',
    '\u0173': 'u','\u1E77': 'u','\u1E75': 'u','\u0289': 'u',
    
    '\u0076': 'v','\u24E5': 'v','\uFF56': 'v','\u1E7D': 'v','\u1E7F': 'v','\u028B': 'v','\uA75F': 'v','\u028C': 'v',
    '\uA761': 'vy',
    '\u0077': 'w','\u24E6': 'w','\uFF57': 'w','\u1E81': 'w','\u1E83': 'w','\u0175': 'w','\u1E87': 'w','\u1E85': 'w','\u1E98': 'w','\u1E89': 'w',
    '\u2C73': 'w',
    '\u0078': 'x','\u24E7': 'x','\uFF58': 'x','\u1E8B': 'x','\u1E8D': 'x',
    
    '\u0079': 'y','\u24E8': 'y','\uFF59': 'y','\u1EF3': 'y','\u00FD': 'y','\u0177': 'y','\u1EF9': 'y','\u0233': 'y','\u1E8F': 'y','\u00FF': 'y',
    '\u1EF7': 'y','\u1E99': 'y','\u1EF5': 'y','\u01B4': 'y','\u024F': 'y','\u1EFF': 'y',
    
    '\u007A': 'z','\u24E9': 'z','\uFF5A': 'z','\u017A': 'z','\u1E91': 'z','\u017C': 'z','\u017E': 'z','\u1E93': 'z','\u1E95': 'z','\u01B6': 'z',
    '\u0225': 'z','\u0240': 'z','\u2C6C': 'z','\uA763': 'z',
    };
    

    Use

    function removeDicretics(str) {
        return str.replace(/[^\u0000-\u007E]/g, function (weirdo) {
            return diacriticsMap[weirdo] || weirdo;
        });
    }
    // Call
    var cleanStr = removeDicretics("bułłśhìt"); // bullshit
    
    0 讨论(0)
  • 2020-11-21 14:07

    Shortened code based on the excellent solution by Ian Elliott:

    accentsTidy = function(s){
        var r = s.toLowerCase();
        non_asciis = {'a': '[àáâãäå]', 'ae': 'æ', 'c': 'ç', 'e': '[èéêë]', 'i': '[ìíîï]', 'n': 'ñ', 'o': '[òóôõö]', 'oe': 'œ', 'u': '[ùúûűü]', 'y': '[ýÿ]'};
        for (i in non_asciis) { r = r.replace(new RegExp(non_asciis[i], 'g'), i); }
        return r;
    };
    

    Edit: Corrected non-working code

    0 讨论(0)
  • 2020-11-21 14:07

    Here's a very simple solution without too much code using a very simple map of diacritics that includes some or all that map to ascii equivalents containing more than one character, i.e. Æ => AE, ffi => ffi, etc... Also included some very basic functional tests

    var diacriticsMap = {
      '\u00C0': 'A',  // À => A
      '\u00C1': 'A',   // Á => A
      '\u00C2': 'A',   // Â => A
      '\u00C3': 'A',   // Ã => A
      '\u00C4': 'A',   // Ä => A
      '\u00C5': 'A',   // Å => A
      '\u00C6': 'AE', // Æ => AE
      '\u00C7': 'C',   // Ç => C
      '\u00C8': 'E',   // È => E
      '\u00C9': 'E',   // É => E
      '\u00CA': 'E',   // Ê => E
      '\u00CB': 'E',   // Ë => E
      '\u00CC': 'I',   // Ì => I
      '\u00CD': 'I',   // Í => I
      '\u00CE': 'I',   // Î => I
      '\u00CF': 'I',   // Ï => I
      '\u0132': 'IJ', // IJ => IJ
      '\u00D0': 'D',   // Ð => D
      '\u00D1': 'N',   // Ñ => N
      '\u00D2': 'O',   // Ò => O
      '\u00D3': 'O',   // Ó => O
      '\u00D4': 'O',   // Ô => O
      '\u00D5': 'O',   // Õ => O
      '\u00D6': 'O',   // Ö => O
      '\u00D8': 'O',   // Ø => O
      '\u0152': 'OE', // Œ => OE
      '\u00DE': 'TH', // Þ => TH
      '\u00D9': 'U',   // Ù => U
      '\u00DA': 'U',   // Ú => U
      '\u00DB': 'U',   // Û => U
      '\u00DC': 'U',   // Ü => U
      '\u00DD': 'Y',   // Ý => Y
      '\u0178': 'Y',   // Ÿ => Y
      '\u00E0': 'a',   // à => a
      '\u00E1': 'a',   // á => a
      '\u00E2': 'a',   // â => a
      '\u00E3': 'a',   // ã => a
      '\u00E4': 'a',   // ä => a
      '\u00E5': 'a',   // å => a
      '\u00E6': 'ae', // æ => ae
      '\u00E7': 'c',   // ç => c
      '\u00E8': 'e',   // è => e
      '\u00E9': 'e',   // é => e
      '\u00EA': 'e',   // ê => e
      '\u00EB': 'e',   // ë => e
      '\u00EC': 'i',   // ì => i
      '\u00ED': 'i',   // í => i
      '\u00EE': 'i',   // î => i
      '\u00EF': 'i',   // ï => i
      '\u0133': 'ij', // ij => ij
      '\u00F0': 'd',   // ð => d
      '\u00F1': 'n',   // ñ => n
      '\u00F2': 'o',   // ò => o
      '\u00F3': 'o',   // ó => o
      '\u00F4': 'o',   // ô => o
      '\u00F5': 'o',   // õ => o
      '\u00F6': 'o',   // ö => o
      '\u00F8': 'o',   // ø => o
      '\u0153': 'oe', // œ => oe
      '\u00DF': 'ss', // ß => ss
      '\u00FE': 'th', // þ => th
      '\u00F9': 'u',   // ù => u
      '\u00FA': 'u',   // ú => u
      '\u00FB': 'u',   // û => u
      '\u00FC': 'u',   // ü => u
      '\u00FD': 'y',   // ý => y
      '\u00FF': 'y',   // ÿ => y
      '\uFB00': 'ff', // ff => ff
      '\uFB01': 'fi',   // fi => fi
      '\uFB02': 'fl', // fl => fl
      '\uFB03': 'ffi',  // ffi => ffi
      '\uFB04': 'ffl',  // ffl => ffl
      '\uFB05': 'ft', // ſt => ft
      '\uFB06': 'st'  // st => st
    };
    
    function replaceDiacritics(str) {
      var returnStr = '';
      if(str) {
        for (var i = 0; i < str.length; i++) {
          if (diacriticsMap[str[i]]) {
            returnStr += diacriticsMap[str[i]];
          } else {
            returnStr += str[i];
          }
        }
      }
      return returnStr;
    }
    
    function testStripDiacritics(input, expected) {
      var coChar = replaceDiacritics(input);
      console.log('The character passed in was ' + input);
      console.log('The character that came out was ' + coChar);
      console.log('The character expected was' + expected);
    }
    
    testStripDiacritics('À','A');
    testStripDiacritics('A','A');
    testStripDiacritics('Æ','AE');
    testStripDiacritics('AE','AE');
    testStripDiacritics('ÇhÀrlËšYŸZŽ','ChArlEsYYZZ');
    
    0 讨论(0)
  • 2020-11-21 14:08

    I used string.js's latinise() method, which lets you do it like this:

    var output = S(input).latinise().toString();
    
    0 讨论(0)
提交回复
热议问题