<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<style>
#inputField {
font-size: 200%;
width: 100%;
box-sizing: border-box;
}
#patternsField > * {
border: 1px solid #eee;
border-radius: .4em;
margin: 1em 0;
padding: 1em;
font: bold 1em sans-serif;
}
#patternsField > .matches {
background-color: #4CAF50;
color: white;
}
</style>
</head>
<body>
<input id="inputField" value="Если ли в нолечие лапата?">
<div id="patternsField"></div>
<script>
function Soundex(expression) {
if(this instanceof Soundex === false)
return new Soundex(expression);
this.expression = expression;
this.calculate(this.expression);
}
Soundex.prototype = {
__proto__: Array.prototype,
_codes: {
"a":{
"0":[0,-1,-1],
"i":[[0,1,-1]],
"j":[[0,1,-1]],
"y":[[0,1,-1]],
"u":[[0,7,-1]]
},
"b":[[7,7,7]],
"c":{
"0":[5,5,5],
"1":[4,4,4],
"z":{"0":[4,4,4],"s":[[4,4,4]]},
"s":{"0":[4,4,4],"z":[[4,4,4]]},
"k":[[5,5,5],[45,45,45]],
"h":{"0":[5,5,5],"1":[4,4,4],"s":[[5,54,54]]}
},
"d":{
"0":[3,3,3],
"t":[[3,3,3]],
"z":{"0":[4,4,4],"h":[[4,4,4]],"s":[[4,4,4]]},
"s":{"0":[4,4,4],"h":[[4,4,4]],"z":[[4,4,4]]},
"r":{"s":[[4,4,4]],"z":[[4,4,4]]}
},
"e":{
"0":[0,-1,-1],
"i":[[0,1,-1]],
"j":[[0,1,-1]],
"y":[[0,1,-1]],
"u":[[1,1,-1]],
"w":[[1,1,-1]]
},
"f":{"0":[7,7,7],"b":[[7,7,7]]},
"g":[[5,5,5]],
"h":[[5,5,-1]],
"i":{
"0":[0,-1,-1],
"a":[[1,-1,-1]],
"e":[[1,-1,-1]],
"o":[[1,-1,-1]],
"u":[[1,-1,-1]]
},
"j":[[4,4,4]],
"k":{"0":[5,5,5],"h":[[5,5,5]],"s":[[5,54,54]]},
"l":[[8,8,8]],
"m":{"0":[6,6,6],"n":[[66,66,66]]},
"n":{"0":[6,6,6],"m":[[66,66,66]]},
"o":{"0":[0,-1,-1],"i":[[0,1,-1]],"j":[[0,1,-1]],"y":[[0,1,-1]]},
"p":{"0":[7,7,7],"f":[[7,7,7]],"h":[[7,7,7]]},
"q":[[5,5,5]],
"r":{
"0":[9,9,9],
"d": {"0":[93,93,93],"c":[[95,95,95]]},
"t": {"0":[93,93,93],"c":[[95,95,95]]},
"z":[[94,94,94],[94,94,94]],
"s":[[94,94,94],[94,94,94]]
},
"s":{
"0":[4,4,4],
"l":{"0":[2,43,43]},
"z":{
"0":[4,4,4],
"t":[[2,43,43]],
"c":{"z":[[2,4,4]],"s":[[2,4,4]]},
"d":[[2,43,43]]
},
"d":[[2,43,43]],
"t":{
"0":[2,43,43],
"r":{"z":[[2,4,4]],"s":[[2,4,4]]},
"n":[[26,46,46]],
"c":{"h":[[2,4,4]]},
"s":{"h":[[2,4,4]],"c":{"h":[[2,4,4]]}}
},
"c":{
"0":[2,4,4],
"h":{
"0":[4,4,4],
"t":{
"0":[2,43,43],
"s":{"c":{"h":[[2,4,4]]},"h":[[2,4,4]]},
"c":{"h":[[2,4,4]]}
},
"d":[[2,43,43]]
}
},
"h":{
"0":[4,4,4],
"t":{"0":[2,43,43],"c":{"h":[[2,4,4]]},"s":{"h":[[2,4,4]]}},
"c":{"h":[[2,4,4]]},
"d":[[2,43,43]]
}
},
"t":{
"0":[3,3,3],
"c":{"0":[4,4,4],"h":[[4,4,4]]},
"z":{"0":[4,4,4],"s":[[4,4,4]]},
"s":{"0":[4,4,4],"z":[[4,4,4]],"h":[[4,4,4]],"c":{"h":[[4,4,4]]}},
"t":{
"s":{"0":[4,4,4],"z":[[4,4,4]],"c":{"h":[[4,4,4]]}},
"c":{"h":[[4,4,4]]},
"z":[[4,4,4]]
},
"h":[[3,3,3]],
"r":{"z":[[4,4,4]],"s":[[4,4,4]]}
},
"u":{
"0":[0,-1,-1],
"e":[[0,-1,-1]],
"i":[[0,1,-1]],
"j":[[0,1,-1]],
"y":[[0,1,-1]]
},
"v":[[7,7,7]],
"w":[[7,7,7]],
"x":[[5,54,54]],
"y":[[1,-1,-1]],
"z":{
"0":[4,4,4],
"d":{"0":[2,43,43],"z":{"0":[2,4,4],"h":[[2,4,4]]}},
"h":{"0":[4,4,4],"d":{"0":[2,43,43],"z":{"h":[[2,4,4]]}}},
"s":{"0":[4,4,4],"h":[[4,4,4]],"c":{"h":[[4,4,4]]}}
}
},
_map: {
а: "a", б: "b", в: "v", г: "g", д: "d", е: "e", ё: "e", ж: "zh", з: "z",
и: "i", й: "i", к: "k", л: "l", м: "m", н: "n", о: "o", п: "p", р: "r",
с: "s", т: "t", у: "u", ф: "f", х: "h", ц: "c", ч: "ch", ш: "sh", щ: "sch",
ь: "'", ы: "y", ъ: "'", э: "e", ю: "yu", я: "ya"
},
_accented: {
a: /[AaªÀ-Åà-åĀ-ąǍǎȀ-ȃȦȧᴬᵃḀḁẚẠ-ảₐ⒜ⒶⓐＡａ]/ig,
b: /[BbᴮᵇḂ-ḇℬ⒝ⒷⓑＢｂ]/ig,
c: /[CcÇçĆ-čᶜ℀ℂ℃ℭⅭⅽ⒞Ⓒⓒ㏄-㏇Ｃｃ]/ig,
d: /[DdĎďDŽ-džDZ-dzᴰᵈḊ-ḓⅅⅆⅮⅾ⒟ⒹⓓＤｄ]/ig,
e: /[EeÈ-Ëè-ëĒ-ěȄ-ȇȨȩᴱᵉḘ-ḛẸ-ẽₑℯℰⅇ⒠ⒺⓔＥｅ]/ig,
f: /[FfᶠḞḟ℉ℱ℻⒡Ⓕⓕff-fflＦｆ]/ig,
g: /[GgĜ-ģǦǧǴǵᴳᵍḠḡℊ⒢ⒼⓖＧｇ]/ig,
h: /[HhĤĥȞȟʰᴴḢ-ḫẖℋ-ℎ⒣ⒽⓗＨｈ]/ig,
i: /[IiÌ-Ïì-ïĨ-İIJijǏǐȈ-ȋᴵᵢḬḭỈ-ịⁱℐℑℹⅈ⒤ⒾⓘＩｉ]/ig,
j: /[JjIJ-ĵLJ-njǰʲᴶⅉ⒥ⒿⓙⱼＪｊ]/ig,
k: /[KkĶķǨǩᴷᵏḰ-ḵK⒦ⓀⓚＫｋ]/ig,
l: /[LlĹ-ŀLJ-ljˡᴸḶḷḺ-ḽℒℓⅬⅼ⒧ⓁⓛＬｌ]/ig,
m: /[MmᴹᵐḾ-ṃ℠™ℳⅯⅿ⒨ⓂⓜＭｍ]/ig,
n: /[NnÑñŃ-ʼnNJ-njǸǹᴺṄ-ṋⁿℕ№⒩ⓃⓝＮｎ]/ig,
o: /[OoºÒ-Öò-öŌ-őƠơǑǒǪǫȌ-ȏȮȯᴼᵒỌ-ỏₒℴ⒪ⓄⓞＯｏ]/ig,
p: /[PpᴾᵖṔ-ṗℙ⒫Ⓟⓟ㎀㎊㎩Ｐｐ]/ig,
q: /[Qqℚ⒬Ⓠⓠ㏃Ｑｑ]/ig,
r: /[RrŔ-řȐ-ȓʳᴿᵣṘ-ṛṞṟ₨ℛ-ℝ⒭ⓇⓡＲｒ]/ig,
s: /[SsŚ-šſȘșˢṠ-ṣ₨℁℠⒮ⓈⓢstＳｓ]/ig,
t: /[TtŢ-ťȚțᵀᵗṪ-ṱẗ℡™⒯Ⓣⓣ㏏ſtstＴｔ]/ig,
u: /[UuÙ-Üù-üŨ-ųƯưǓǔȔ-ȗᵁᵘᵤṲ-ṷỤ-ủ℆⒰Ⓤⓤ㍳㍺Ｕｕ]/ig,
v: /[VvᵛᵥṼ-ṿ⒱ⓋⓥⱽＶｖ]/ig,
w: /[WwŴŵʷᵂẀ-ẉẘ⒲ⓌⓦＷｗ]/ig,
x: /[XxˣẊ-ẍₓ⒳Ⓧⓧ㏓Ｘｘ]/ig,
y: /[YyÝýÿŶ-ŸȲȳʸẎẏẙỲ-ỹ⒴ⓎⓨＹｙ]/ig,
z: /[ZzŹ-žDZ-dzᶻẐ-ẕℤℨ⒵ⓏⓩＺｚ]/ig
},
_cache: {},
word(str, isCyrillic = true) {
var length = str.length,
result = "",
i = 0,
previousCode = -1,
maxLength = 6,
code;
while(i < length) {
var current = last = this._codes[str[i]];
for(var j = k = 1; k <= maxLength; k++) {
if(!str[i + k] || !current[str[i + k]]) break;
current = current[str[i + k]];
if(current[0]) {
last = current;
j = k + 1;
}
}
if(i === 0)
code = last[0][0];
else
code = (isCyrillic || "1" in last === false ? last[0] : last[1])[
str[i + j] && this._codes[str[i + j]][0][0] === 0 ? 1 : 2
];
if(code !== -1 && code !== previousCode) result += code;
previousCode = code;
i += j;
}
while(result && result.length < 3)
result += "0";
return result;
},
calculate(str) {
str = str.trim().toLowerCase();
var cyrillicMatch = str.match(/[ёа-я]/g);
if(cyrillicMatch)
str = this.transliterate(str, cyrillicMatch);
for(var letter in this._accented)
str = str.replace(this._accented[letter], letter);
str = str
.replace(/[^\sa-z]/g, "")
.replace(/\s{2,}/g, " ");
if(!str) return [];
str = str.split(" ").map(word => {
if(word in this._cache === false)
this._cache[word] = this.word(word, cyrillicMatch !== null);
return this._cache[word];
});
Object.assign(this, str);
this.length = str.length;
},
transliterate(str, matches) {
for(const match of matches)
str = str.replace(match, this._map[match]);
return str;
},
toString() {
return this.join(" ");
}
};
function Rx(template, ...substitutions) {
if(this instanceof Rx === false) return new Rx(template, ...substitutions);
this.raw = template[0] + substitutions.map((v, i) => `(${v.source})${template[i + 1]}`).join("");
this.pattern = (template.slice(1).map((part, index) => `(${substitutions[index].source}) ${Soundex(part)}`));
this.pattern.unshift(Soundex(template[0]));
this.pattern = new RegExp(this.pattern.join(" ").trim());
}
Rx.prototype = {
test(expression) {
return this.pattern.test(Soundex(expression));
},
match(expression) {
return (this.pattern.exec(Soundex(expression)) || []).slice(1);
}
};
// пример
const patterns = [
Rx`Есть ли в наличии ${/[\s\S]+/}?`,
Rx`Есть ли ${/[\s\S]+/} в наличии?`,
];
inputField.addEventListener("input", () => {
patternsField.innerHTML = patterns.map(pattern => {
const ok = pattern.test(inputField.value);
return `<div class="${ok ? "matches" : ""}">${pattern.raw}</div>`;
}).join("");
});
inputField.dispatchEvent(new Event("input"));
</script>
</body>
</html>
[1]: Paramonov V., Shigarov A., Ruzhnikov G., Belykh P. Polyphon: An Algorithm for Phonetic String Matching in Russian Language // Communications in Computer and Information Science. Springer. 2016. Vol. 639, pp. 568-579.
