string-to-unicode-variant
Version:
Javascript function to convert a string into different kind of ⓤⓝⓘⓒⓞⓓⓔ variants.
613 lines (611 loc) • 21.4 kB
JavaScript
var __create = Object.create;
var __defProp = Object.defineProperty;
var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
var __getOwnPropNames = Object.getOwnPropertyNames;
var __getProtoOf = Object.getPrototypeOf;
var __hasOwnProp = Object.prototype.hasOwnProperty;
var __commonJS = (cb, mod) => function __require() {
return mod || (0, cb[__getOwnPropNames(cb)[0]])((mod = { exports: {} }).exports, mod), mod.exports;
};
var __copyProps = (to, from, except, desc) => {
if (from && typeof from === "object" || typeof from === "function") {
for (let key of __getOwnPropNames(from))
if (!__hasOwnProp.call(to, key) && key !== except)
__defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
}
return to;
};
var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
// If the importer is in node compatibility mode or this is not an ESM
// file that has been converted to a CommonJS file using a Babel-
// compatible transform (i.e. "__esModule" has not been set), then set
// "default" to the CommonJS "module.exports" for node compatibility.
isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
mod
));
// toUnicodeVariant.js
var require_toUnicodeVariant = __commonJS({
"toUnicodeVariant.js"(exports, module) {
function toUnicodeVariant2(str, variant, combinings) {
const string = String.fromCodePoint;
const offsets = {
m: [120432, 120822],
b: [119808, 120782],
i: [119860, 48],
bi: [119912, 48],
c: [119964, 48],
bc: [120016, 48],
g: [120068, 48],
d: [120120, 120792],
bg: [120172, 48],
s: [120224, 120802],
bs: [120276, 120812],
is: [120328, 48],
bis: [120380, 48],
o: [9398, 9311],
on: [127312, 9311],
p: [127248, 120822],
q: [127280, 48],
qn: [127344, 48],
w: [65313, 65296],
//
f: [127462, 120822],
nd: [120432, 9351],
nc: [120432, 127233],
ndc: [120432, 9460],
r: [120432, 9460]
};
const variantOffsets = {
"monospace": "m",
"bold": "b",
"italic": "i",
"bold italic": "bi",
"script": "c",
"bold script": "bc",
"gothic": "g",
"gothic bold": "bg",
"doublestruck": "d",
"sans": "s",
"bold sans": "bs",
"italic sans": "is",
"bold italic sans": "bis",
"parenthesis": "p",
"circled": "o",
"circled negative": "on",
"squared": "q",
"squared negative": "qn",
"fullwidth": "w",
//
"flags": "f",
"numbers dot": "nd",
"numbers comma": "nc",
"numbers double circled": "ndc",
"roman": "r"
};
const special = {
m: {
" ": 8192,
"-": 8211
},
i: {
"h": 8462
},
c: {
"B": 8492,
"E": 8496,
"F": 8497,
"H": 8459,
"I": 8464,
"L": 8466,
"M": 8499,
"R": 8475,
"e": 120046,
"g": 120048,
"o": 120056
},
g: {
"C": 8493,
"H": 8460,
"I": 8465,
"R": 8476,
"Z": 8488
},
d: {
"C": 8450,
"H": 8461,
"N": 8469,
"P": 8473,
"Q": 8474,
"R": 8477,
"Z": 8484
},
o: {
"0": 9450,
"10": 9321,
"11": 9322,
"12": 9323,
"13": 9324,
"14": 9325,
"15": 9326,
"16": 9327,
"17": 9328,
"18": 9329,
"19": 9330,
"20": 9331
},
on: {
"0": 9471,
"11": 9451,
"12": 9452,
"13": 9453,
"14": 9454,
"15": 9455,
"16": 9456,
"17": 9457,
"18": 9458,
"19": 9459,
"20": 9460
},
p: {
"1": 9332,
"2": 9333,
"3": 9334,
"4": 9335,
"5": 9336,
"6": 9337,
"7": 9338,
"8": 9339,
"9": 9340,
"10": 9341,
"11": 9342,
"12": 9343,
"13": 9344,
"14": 9345,
"15": 9346,
"16": 9347,
"17": 9348,
"18": 9349,
"19": 9350,
"20": 9351
},
q: {
"hv": 127306,
"mv": 127307,
"sd": 127308,
"ss": 127309,
"ppv": 127310,
"wc": 127311,
"cl": 127377,
"cool": 127378,
"free": 127379,
"id": 127380,
"new": 127381,
"ng": 127382,
"ok": 127383,
"sos": 127384,
"up!": 127385,
"vs": 127386,
"3d": 127387,
"2ndscr": 127388,
"2k": 127389,
"4k": 127390,
"8k": 127391,
"5.1": 127392,
"7.1": 127393,
"22.2": 127394,
"60p": 127395,
"120p": 127396,
"d": 127397,
"hc": 127398,
"hdr": 127399,
"hi-res": 127400,
"loss-less": 127401,
"shv": 127402,
"uhd": 127403,
"vod": 127404
},
qn: {
"ic": 127371,
"pa": 127372,
"sa": 127373,
"ab": 127374,
"wc": 127375
},
w: {
"!": 65281,
'"': 65282,
"#": 65283,
"$": 65284,
"%": 65285,
"&": 65286,
"'": 65287,
"(": 65288,
")": 65289,
"*": 65290,
"+": 65291,
",": 65292,
"-": 65293,
".": 65294,
"/": 65295,
":": 65306,
";": 65307,
"<": 65308,
"=": 65309,
">": 65310,
"?": 65311,
"@": 65312,
"\\": 65340,
"[": 65339,
"]": 65341,
"^": 65342,
"_": 65343,
"`": 65344,
"{": 65371,
"|": 65372,
"}": 65373,
"~": 65374,
"\u2985": 65375,
"\u2986": 65376,
"\uFFE0": 65504,
"\uFFE1": 65505,
"\xA6": 65508,
"\uFFE5": 65509,
"\uFFE6": 65510,
"\uFF70": 65392,
"\uFF61": 65392,
"\uFF64": 65380,
"\uFF65": 65381,
"\uFFE3": 65507,
"\xAC": 65506
},
f: {},
nd: {
"0": 127232,
"10": 9361,
"11": 9362,
"12": 9363,
"13": 9364,
"14": 9365,
"15": 9366,
"16": 9367,
"17": 9368,
"18": 9369,
"19": 9370,
"20": 9371
},
ndc: {
"0": 120822,
"10": 9470
},
r: {
"I": 8544,
"II": 8545,
"III": 8546,
"IV": 8547,
"V": 8548,
"VI": 8549,
"VII": 8550,
"VIII": 8551,
"IX": 8552,
"X": 8553,
"XI": 8554,
"XII": 8555,
"L": 8556,
"C": 8557,
"D": 8558,
"M": 8559,
"i": 8560,
"ii": 8561,
"iii": 8562,
"iv": 8563,
"v": 8564,
"vi": 8565,
"vii": 8566,
"viii": 8567,
"ix": 8568,
"x": 8569,
"xi": 8570,
"xii": 8571,
"l": 8572,
"c": 8573,
"d": 8574,
"m": 8575
}
};
for (var i = 97; i <= 122; i++) {
special["p"][String.fromCharCode(i)] = 9372 + (i - 97);
special["w"][String.fromCharCode(i)] = 65345 + (i - 97);
}
;
["on", "q", "qn", "f"].forEach((t) => {
for (var i2 = 97; i2 <= 122; i2++) {
special[t][String.fromCharCode(i2)] = offsets[t][0] + (i2 - 97);
}
});
const diacritics = {
"strike": { "short": "s", "code": 822 },
"strike-curly": { "short": "sc", "code": 820 },
"underline": { "short": "u", "code": 818 },
"underline-curly": { "short": "uc", "code": 816 },
"underline-sm": { "short": "u-sm", "code": 800 },
"underline-double": { "short": "ud", "code": 819 },
"underline-double-sm": { "short": "ud-sm", "code": 839 },
"overline": { "short": "o", "code": 773 },
"overline-curly": { "short": "oc", "code": 771 },
"overline-sm": { "short": "o-sm", "code": 772 },
"overline-double": { "short": "od", "code": 831 },
"slash": { "short": "sl", "code": 824 },
"cross-above": { "short": "ca", "code": 829 },
"plus-below": { "short": "pb", "code": 799 },
"a-above": { "short": "a-a", "code": 867 },
"c-above": { "short": "c-a", "code": 872 },
"d-above": { "short": "d-a", "code": 873 },
"e-above": { "short": "e-a", "code": 868 },
"h-above": { "short": "h-a", "code": 874 },
"i-above": { "short": "i-a", "code": 869 },
"m-above": { "short": "m-a", "code": 875 },
"o-above": { "short": "o-a", "code": 870 },
"r-above": { "short": "r-a", "code": 876 },
"u-above": { "short": "u-a", "code": 871 },
"v-above": { "short": "v-a", "code": 878 },
"x-above": { "short": "x-a", "code": 879 },
"halo-breve": { "short": "hb", "code": 1160 },
"halo-grave": { "short": "hg", "code": 1161 },
"enclose-circle": { "short": "en-c", "code": 8413 },
"enclose-backslash": { "short": "en-cb", "code": 8416 },
"enclose-circle-backslash": { "short": "en-cb", "code": 8416 },
"enclose-square": { "short": "en-s", "code": 8414 },
"enclose-diamond": { "short": "en-d", "code": 8415 },
"enclose-screen": { "short": "en-scr", "code": 8418 },
"enclose-keycap": { "short": "en-key", "code": 8419 },
//diacritics supporting special chars
"diaeresis": { "code": 776 },
"caron": { "code": 780 },
"perispomeni": { "code": 834 },
"tilde": { "code": 771 },
"tildesm": { "code": 732 },
"circumflex": { "code": 770 },
"ringabove": { "code": 778 },
"dotabove": { "code": 775 },
"dotbelow": { "code": 803 },
"grave": { "code": 832 },
"gravedouble": { "code": 783 },
"acute": { "code": 833 },
"breve": { "code": 774 },
"breveinverted": { "code": 785 },
"commabelow": { "code": 806 },
"macron": { "code": 772 },
"cedilla": { "code": 807 },
"ogonek": { "code": 808 },
"solidus": { "code": 824 },
"solidussm": { "code": 823 },
//spacing combinings
"space-zero": { "code": 65279 },
"space-hair": { "code": 8202 },
"space-thin": { "code": 8201 },
"space": { "code": 32 },
"space-en": { "code": 8192 },
"space-figure": { "code": 8199 },
"space-cjk": { "code": 12288 },
"space-em": { "code": 8193 },
"space-ogham": { "code": 5760 },
//combining grapheme joiner
"CGJ": { "code": 847 }
};
const special_chars = {
"\xE4": { "char": "a", "combine": string(diacritics.diaeresis.code) },
"\xE2": { "char": "a", "combine": string(diacritics.circumflex.code) },
"\xE1": { "char": "a", "combine": string(diacritics.acute.code) },
"\xE5": { "char": "a", "combine": string(diacritics.ringabove.code) },
"\u0103": { "char": "a", "combine": string(diacritics.breve.code) },
"\u01DF": { "char": "a", "combine": string(diacritics.diaeresis.code) + string(diacritics.macron.code) },
"\xE3": { "char": "a", "combine": string(diacritics.tilde.code) },
"\u0101": { "char": "a", "combine": string(diacritics.macron.code) },
"\u0227": { "char": "a", "combine": string(diacritics.dotabove.code) },
"\u0203": { "char": "a", "combine": string(diacritics.breveinverted.code) },
"\u1E05": { "char": "b", "combine": string(diacritics.dotbelow.code) },
"\u010D": { "char": "c", "combine": string(diacritics.caron.code) },
"\u0107": { "char": "c", "combine": string(diacritics.acute.code) },
"\xE7": { "char": "c", "combine": string(diacritics.cedilla.code) },
"\u1E09": { "char": "c", "combine": string(diacritics.cedilla.code) + string(diacritics.acute.code) },
"\u010B": { "char": "c", "combine": string(diacritics.dotabove.code) },
"\u0109": { "char": "c", "combine": string(diacritics.circumflex.code) },
"\xE8": { "char": "e", "combine": string(diacritics.grave.code) },
"\xE9": { "char": "e", "combine": string(diacritics.acute.code) },
"\u0113": { "char": "e", "combine": string(diacritics.macron.code) },
"\u0115": { "char": "e", "combine": string(diacritics.breve.code) },
"\xEB": { "char": "e", "combine": string(diacritics.diaeresis.code) },
"\u011B": { "char": "e", "combine": string(diacritics.caron.code) },
"\u0119": { "char": "e", "combine": string(diacritics.ogonek.code) },
"\u0207": { "char": "e", "combine": string(diacritics.breveinverted.code) },
"\u0205": { "char": "e", "combine": string(diacritics.gravedouble.code) },
"\xEA": { "char": "e", "combine": string(diacritics.circumflex.code) },
"\u011F": { "char": "g", "combine": string(diacritics.breve.code) },
"\u01E7": { "char": "g", "combine": string(diacritics.caron.code) },
"\u0123": { "char": "g", "combine": string(diacritics.cedilla.code) },
"\u0121": { "char": "g", "combine": string(diacritics.dotabove.code) },
"\u1E25": { "char": "h", "combine": string(diacritics.dotbelow.code) },
"\xEE": { "char": "i", "combine": string(diacritics.circumflex.code) },
"\xED": { "char": "i", "combine": string(diacritics.acute.code) },
"\xEC": { "char": "i", "combine": string(diacritics.grave.code) },
"\u0129": { "char": "i", "combine": string(diacritics.tilde.code) },
"\u1E31": { "char": "k", "combine": string(diacritics.acute.code) },
"\u1E33": { "char": "k", "combine": string(diacritics.dotbelow.code) },
"\u0146": { "char": "n", "combine": string(diacritics.ogonek.code) },
"\u0144": { "char": "n", "combine": string(diacritics.acute.code) },
"\xF5": { "char": "o", "combine": string(diacritics.tilde.code) },
"\xF6": { "char": "o", "combine": string(diacritics.diaeresis.code) },
"\u014D": { "char": "o", "combine": string(diacritics.macron.code) },
"\xF4": { "char": "o", "combine": string(diacritics.circumflex.code) },
"\xF3": { "char": "o", "combine": string(diacritics.acute.code) },
"\xF2": { "char": "o", "combine": string(diacritics.grave.code) },
"\u0155": { "char": "r", "combine": string(diacritics.acute.code) },
"\u015F": { "char": "s", "combine": string(diacritics.cedilla.code) },
"\u0219": { "char": "s", "combine": string(diacritics.commabelow.code) },
"\u1E69": { "char": "s", "combine": string(diacritics.dotbelow.code) + string(diacritics.dotabove.code) },
"\u0161": { "char": "s", "combine": string(diacritics.caron.code) },
"\u015B": { "char": "s", "combine": string(diacritics.acute.code) },
"\xFC": { "char": "u", "combine": string(diacritics.diaeresis.code) },
"\xF9": { "char": "u", "combine": string(diacritics.grave.code) },
"\xFA": { "char": "u", "combine": string(diacritics.acute.code) },
"\xFB": { "char": "u", "combine": string(diacritics.circumflex.code) },
"\u016D": { "char": "u", "combine": string(diacritics.breve.code) },
"\u0215": { "char": "u", "combine": string(diacritics.gravedouble.code) },
"\u0217": { "char": "u", "combine": string(diacritics.breveinverted.code) },
"\u016F": { "char": "u", "combine": string(diacritics.ringabove.code) },
"\u016B": { "char": "u", "combine": string(diacritics.macron.code) },
"\u1E81": { "char": "w", "combine": string(diacritics.grave.code) },
"\u1E83": { "char": "w", "combine": string(diacritics.acute.code) },
"\xF8": { "char": "o", "combine": string(diacritics.solidussm.code) }
//mimicks that somehow fails
//'c̆': { 'char': 'c', 'combine': string(diacritics.breve.code) },
//'c̈': { 'char': 'c', 'combine': string(diacritics.diaeresis.code) },
//'ę́': { 'char': 'e', 'combine': string(diacritics.ogonek.code) + string(diacritics.acute.code) },
//'m̂': { 'char': 'm', 'combine': string(diacritics.circumflex.code) },
//'n̂': { 'char': 'n', 'combine': string(diacritics.circumflex.code) },
//'ñ': { 'char': 'n', 'combine': string(diacritics.tilde.code) },
//'q̄': { 'char': 'q', 'combine': string(diacritics.macron.code) },
//'r̃': { 'char': 'r', 'combine': string(diacritics.tilde.code) },
//'s̈': { 'char': 's', 'combine': string(diacritics.diaeresis.code) },
};
for (const char of Object.keys(special_chars)) {
special_chars[char.toUpperCase()] = { "char": char, "combine": false };
}
const chars = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
const numbers = "0123456789";
const type = function() {
if (variantOffsets[variant])
return variantOffsets[variant];
if (offsets[variant])
return variant;
return "m";
}();
const combine_with = function() {
let array = null;
if (Array.isArray(combinings))
array = combinings;
if (typeof combinings === "string")
array = combinings.split(",");
if (!array)
return false;
let result2 = "";
array.forEach(function(diacritic) {
diacritic = diacritic.trim().toLowerCase();
for (const d in diacritics) {
if (diacritic === d || diacritic === diacritics[d].short) {
result2 += string(diacritics[d].code);
}
}
});
return result2;
}();
if (typeof str === "string" && special[type] && (special[type][str] || special[type][str.toLowerCase()])) {
return special[type][str] ? string(special[type][str]) : string(special[type][str.toLowerCase()]);
}
if (["roman", "r"].includes(type)) {
if (typeof str === "number") {
const parts = {
M: 1e3,
CM: 900,
D: 500,
CD: 400,
C: 100,
XC: 90,
L: 50,
XL: 40,
X: 10,
IX: 9,
V: 5,
IV: 4,
I: 1
};
let roman = "";
let num = str;
for (let i2 in parts) {
while (num >= parts[i2]) {
if (special[type][i2]) {
roman += i2;
} else {
for (let d of i2)
roman += d;
}
num -= parts[i2];
}
}
str = roman;
}
let result2 = str;
const romans = [
"VIII",
"viii",
"III",
"iii",
"XII",
"xii",
"VII",
"vii",
"IX",
"ix",
"XI",
"xi",
"IV",
"iv",
"VI",
"vi",
"II",
"ii",
"I",
"i",
"D",
"d",
"M",
"m",
"L",
"l",
"V",
"v",
"C",
"c",
"X",
"x"
];
for (number of romans) {
if (result2.indexOf(number.toString()) > -1) {
result2 = result2.replaceAll(number, string(special[type][number]));
}
}
return result2;
}
let result = "";
for (let c of str) {
let index;
const combine_special = c in special_chars ? special_chars[c].combine : false;
c = combine_special ? special_chars[c].char : c.normalize("NFD").replace(/[\u0300-\u036f]/g, "");
if (special[type] && special[type][c])
c = string(special[type][c]);
if (type && (index = chars.indexOf(c)) > -1) {
result += string(index + offsets[type][0]);
} else if (type && (index = numbers.indexOf(c)) > -1) {
result += string(index + offsets[type][1]);
} else {
result += c;
}
if (combine_special)
result += combine_special;
if (combine_with)
result += combine_with;
}
return result;
}
if (typeof module === "object" && module && typeof module.exports === "object") {
module.exports = toUnicodeVariant2;
}
}
});
// index.ts
var toUnicodeVariant = __toESM(require_toUnicodeVariant());
var string_to_unicode_variant = toUnicodeVariant.default;
export {
string_to_unicode_variant
};
//# sourceMappingURL=index.mjs.map