hanzi-tools
Version:
Converts from Chinese characters to pinyin, between simplified and traditional, and does word segmentation.
213 lines (202 loc) • 8.36 kB
JavaScript
let segment = require("./segment"),
pinyin = require("pinyin"),
{ pinyinDict } = require("./pinyinDict"),
{ normalizeEnglish } = require("./punctuation"),
tag = require("./tag"),
{ isCharacterText } = require("./util");
const tones = [
['ā','ē','ī','ō','ū','ǖ','Ā','Ē','Ī','Ō','Ū','Ǖ'],
['á','é','í','ó','ú','ǘ','Á','É','Í','Ó','Ú','Ǘ'],
['ǎ','ě','ǐ','ǒ','ǔ','ǚ','Ǎ','Ě','Ǐ','Ǒ','Ǔ','Ǚ'],
['à','è','ì','ò','ù','ǜ','À','È','Ì','Ò','Ù','Ǜ'],
['a','e','i','o','u','ü','A','E','I','O','U','Ü']
]; //Use tones[tone - 1] to get all possible characters with that tone
function pinyinify(text, isDetailed) {
let segments = segment(text);
let pinyinSegments = [];
segments.forEach((text, i, segments) => {
if (text.length === 1) pinyinSegments.push(pinyinifyChar(text, segments, i));
else pinyinSegments.push(pinyinDict[text]);
});
let out = [];
pinyinSegments.forEach((text, i) => {
if (shouldPutSpaceBetween(segments[i], segments[i + 1])) {
out.push(text + " ");
} else {
out.push(text);
}
});
out = out.join("").trim();
if (isDetailed) {
segments = segments;
pinyinSegments = pinyinSegments.map(normalizeEnglish);
return {
segments,
pinyinSegments,
pinyinSegmentsSyllables: pinyinSegments.map(segment => segment.split("\u200B")),
pinyin: normalizeEnglish(out)
};
}
return normalizeEnglish(out);
}
function pinyinifyChar(text, cuts, cutIndex) {
let disambiguatedChar = decideAmbiguousChar(text, cuts, cutIndex);
if (disambiguatedChar) {
return disambiguatedChar;
}
let word = pinyinDict[text];
if (word) {
return word;
}
let arr = pinyin(text, {
heteronym: true,
segment: true
});
let syllables = arr.map((x) => x[0]);
return syllables.join("\u200B");
}
function decideAmbiguousChar(char, cuts, cutIndex) {
let previousText = cuts.slice(Math.max(0, cutIndex - 10), cutIndex);
let afterText = cuts.slice(cutIndex + 1, cutIndex + 10);
let nextTags, prevTags;
switch (char) {
case "觉":
case "覺":
if (previousText.join("").includes("睡")) return "jiào";
return "jué";
case "长":
case "長":
prevTags = tag(previousText.join(""));
nextTags = tag(afterText.join(""));
let nextTag = nextTags && nextTags.length && nextTags[0].tag;
if (nextTag === "uz")
return "zhǎng";
let prevTag = prevTags && prevTags.length && prevTags[prevTags.length - 1].tag;
if (prevTag === "n") return "zhǎng";
if (prevTag !== "d" && nextTag === "ul") {
return "zhǎng";
}
// zhǎng has higher frequency due to compond words,
// but cháng is more common as an individual character.
return "cháng";
case "得":
nextTags = tag(afterText.join(""));
prevTags = tag(previousText.join(""));
if (nextTags && nextTags.length) {
let afterTag = nextTags[0].tag;
let prevTag = prevTags.length && prevTags[prevTags.length - 1].tag
if (prevTag === "v") {
break;
}
if (prevTag === "a" || prevTag === "b" || prevTag === "nr") {
break;
}
if (afterTag === "ul") {
return "dé";
}
if (prevTag === "d" || prevTag === "r") {
return "děi";
}
if (nextTags[0].word === "还" || nextTags[0].word === "還") {
if (nextTags[1] && nextTags[1].tag[0] === "r" || nextTags[1].tag[0] === "n") {
return "děi";
}
}
if (afterTag[0] === "t" || afterTag[0] === "v" || afterTag[0] === "p" || afterTag[0] === "l" || afterTag[0] === "n") {
return "děi";
}
}
break;
case "还":
case "還":
if (previousText.join("").includes("把")) return "huán";
nextTags = tag(afterText.join(""));
if (nextTags && nextTags.length) {
let afterTag = nextTags[0].tag;
if (afterText[0][0] === "有") break;
if (afterTag[0] === "r" || afterTag[0] === "n") {
return "huán";
}
}
break;
case "行":
prevTags = tag(previousText.join(""));
nextTags = tag(afterText.join(""));
if (prevTags.length && prevTags[prevTags.length - 1].tag === "m") {
return "háng";
}
break;
case "只":
let prev = tag(previousText.join("")).slice(-1)[0];
let after = tag(afterText.join(""))[0];
if (prev && prev.tag === "m") return "zhī";
if (after && after.tag === "n") return "zhī";
return "zhǐ";
case "系":
nextTags = tag(afterText.join(""));
if (nextTags && nextTags.length) {
let afterTag = nextTags[0].tag;
if (afterTag === "f" || afterTag[0] === "u") return "jì";
}
return "xì";
case "地":
prevTags = tag(previousText.join(""));
nextTags = tag(afterText.join(""));
if (prevTags.length && prevTags[prevTags.length - 1].tag === "r") {
return "dì";
}
break;
case "弹":
nextTags = tag(afterText.join(""));
if (afterText.includes("吉他")) return "tán";
if (nextTags && nextTags.length) {
let afterTag = nextTags[0].tag;
if (afterTag[0] === "n") return "tán";
}
break;
case "重":
nextTags = tag(afterText.join(""));
if (nextTags && nextTags.length) {
let afterTag = nextTags[0].tag;
if (afterTag[0] === "v") return "chóng";
}
break;
case "不":
if (afterText.length > 0) {
const nextTone = getTone(afterText[0].charAt(0));
if (nextTone === 4) return "bú";
}
break;
case "一":
if (afterText.length > 0) {
const nextTone = getTone(afterText[0].charAt(0));
if (nextTone === 1 || nextTone == 2 || nextTone == 3) return "yì";
else if (nextTone == 4) return "yí";
}
break;
}
}
function shouldPutSpaceBetween(word1, word2) {
if (!word2) return false;
if (word2 === " ") return false;
if (isCharacterText(word1) && isCharacterText(word2)) return true;
if (isCharacterText(word1) && /[ `"'“‘\(\[（【0-9]/.test(word2)) return true;
if (/[`"'“‘\(\[（【]/.test(word1)) return false;
let punctuationPattern = /[0-9\.\?\!\)\]\}！？，。：；’”）%~\@\#\^\&\*]/;
let numberPattern = /[0-9]/;
if (numberPattern.test(word1) && numberPattern.test(word2)) return false;
if (punctuationPattern.test(word1) && !punctuationPattern.test(word2)) return true;
if (punctuationPattern.test(word1) && numberPattern.test(word2)) return true;
if (isCharacterText(word1) && punctuationPattern.test(word2)) return false;
if (isCharacterText(word1) !== isCharacterText(word2)) return true;
return false;
}
function getTone(char) {
if(typeof pinyinDict[char] === 'undefined') return 0;
const pinyinChars = [...pinyinDict[char]]; //Getting and destructuring the pinyin string
for (i = 0; i < 4; i++) { //Going through the four tones and checking if there is a match
if (tones[i].some(c => pinyinChars.includes(c))) return i + 1;
}
return 5;
}
module.exports = pinyinify;