pinyin-split/lib/cjs/index.js

Split up any kind of Pinyin into an array of syllables.

"use strict";
var __importDefault = (this && this.__importDefault) || function (mod) {
    return (mod && mod.__esModule) ? mod : { "default": mod };
};
Object.defineProperty(exports, "__esModule", { value: true });
exports.split = void 0;
const syllables_1 = __importDefault(require("./syllables"));
const normalize = (text) => {
    text = text.normalize('NFD').replace(/\u0304|\u0301|\u030c|\u0300/g, '');
    return text.normalize('NFC').replace(/(\w|ü)[1-5]/gi, '$1').toLowerCase();
};
function split(text, everything = false, wrapInList = false) {
    const list = Array();
    let prevWordFound = false;
    let wordEnd = text.length;
    while (wordEnd > 0) {
        let count = wordEnd;
        let wordFound = false;
        while (count > 0) {
            const word = text.substring(wordEnd - count, wordEnd);
            if (syllables_1.default.includes(normalize(word))) {
                wordFound = true;
                list.push(wrapInList ? [word] : word);
                wordEnd -= (count - 1);
                break;
            }
            count--;
        }
        if (!wordFound && everything) {
            const prevIndex = list.length - 1;
            const prevEntry = list[prevIndex];
            if (wordEnd === text.length || typeof prevEntry === 'object' || prevWordFound) {
                list.push(text[wordEnd - 1]);
            }
            else if (typeof prevEntry === 'string') {
                list[prevIndex] = text[wordEnd - 1] + prevEntry;
            }
        }
        wordEnd--;
        prevWordFound = wordFound;
    }
    return list.reverse();
}
exports.split = split;
exports.default = split;