aboutsummaryrefslogtreecommitdiff
path: root/ext/mixed
diff options
context:
space:
mode:
authortoasted-nutbread <toasted-nutbread@users.noreply.github.com>2020-05-06 19:37:36 -0400
committerGitHub <noreply@github.com>2020-05-06 19:37:36 -0400
commitf361139d744e58a6c33841cee227d13d1970bb98 (patch)
tree187d229001e2ef5dc8df40563ed3c554488f42e3 /ext/mixed
parent4eeb879fb88402503f6ee849fa71e29ab5047b7c (diff)
Japanese util refactor (#510)
* Convert mixed japanese.js to utility class * Copy functions from bg/js/japanese.js into mixed/js/japanese.js * Remove bg/js/japanese.js * Make wanakana dependency optional * Update tests
Diffstat (limited to 'ext/mixed')
-rw-r--r--ext/mixed/js/japanese.js507
1 files changed, 452 insertions, 55 deletions
diff --git a/ext/mixed/js/japanese.js b/ext/mixed/js/japanese.js
index ced486dd..801dec84 100644
--- a/ext/mixed/js/japanese.js
+++ b/ext/mixed/js/japanese.js
@@ -16,6 +16,11 @@
*/
const jp = (() => {
+ const ITERATION_MARK_CODE_POINT = 0x3005;
+ const HIRAGANA_SMALL_TSU_CODE_POINT = 0x3063;
+ const KATAKANA_SMALL_TSU_CODE_POINT = 0x30c3;
+ const KANA_PROLONGED_SOUND_MARK_CODE_POINT = 0x30fc;
+
const HIRAGANA_RANGE = [0x3040, 0x309f];
const KATAKANA_RANGE = [0x30a0, 0x30ff];
const KANA_RANGES = [HIRAGANA_RANGE, KATAKANA_RANGE];
@@ -65,20 +70,65 @@ const jp = (() => {
const SMALL_KANA_SET = new Set(Array.from('ぁぃぅぇぉゃゅょゎァィゥェォャュョヮ'));
+ const HALFWIDTH_KATAKANA_MAPPING = new Map([
+ ['ヲ', 'ヲヺ-'],
+ ['ァ', 'ァ--'],
+ ['ィ', 'ィ--'],
+ ['ゥ', 'ゥ--'],
+ ['ェ', 'ェ--'],
+ ['ォ', 'ォ--'],
+ ['ャ', 'ャ--'],
+ ['ュ', 'ュ--'],
+ ['ョ', 'ョ--'],
+ ['ッ', 'ッ--'],
+ ['ー', 'ー--'],
+ ['ア', 'ア--'],
+ ['イ', 'イ--'],
+ ['ウ', 'ウヴ-'],
+ ['エ', 'エ--'],
+ ['オ', 'オ--'],
+ ['カ', 'カガ-'],
+ ['キ', 'キギ-'],
+ ['ク', 'クグ-'],
+ ['ケ', 'ケゲ-'],
+ ['コ', 'コゴ-'],
+ ['サ', 'サザ-'],
+ ['シ', 'シジ-'],
+ ['ス', 'スズ-'],
+ ['セ', 'セゼ-'],
+ ['ソ', 'ソゾ-'],
+ ['タ', 'タダ-'],
+ ['チ', 'チヂ-'],
+ ['ツ', 'ツヅ-'],
+ ['テ', 'テデ-'],
+ ['ト', 'トド-'],
+ ['ナ', 'ナ--'],
+ ['ニ', 'ニ--'],
+ ['ヌ', 'ヌ--'],
+ ['ネ', 'ネ--'],
+ ['ノ', 'ノ--'],
+ ['ハ', 'ハバパ'],
+ ['ヒ', 'ヒビピ'],
+ ['フ', 'フブプ'],
+ ['ヘ', 'ヘベペ'],
+ ['ホ', 'ホボポ'],
+ ['マ', 'マ--'],
+ ['ミ', 'ミ--'],
+ ['ム', 'ム--'],
+ ['メ', 'メ--'],
+ ['モ', 'モ--'],
+ ['ヤ', 'ヤ--'],
+ ['ユ', 'ユ--'],
+ ['ヨ', 'ヨ--'],
+ ['ラ', 'ラ--'],
+ ['リ', 'リ--'],
+ ['ル', 'ル--'],
+ ['レ', 'レ--'],
+ ['ロ', 'ロ--'],
+ ['ワ', 'ワ--'],
+ ['ン', 'ン--']
+ ]);
- // Character code testing functions
-
- function isCodePointKanji(codePoint) {
- return isCodePointInRanges(codePoint, CJK_UNIFIED_IDEOGRAPHS_RANGES);
- }
-
- function isCodePointKana(codePoint) {
- return isCodePointInRanges(codePoint, KANA_RANGES);
- }
-
- function isCodePointJapanese(codePoint) {
- return isCodePointInRanges(codePoint, JAPANESE_RANGES);
- }
function isCodePointInRanges(codePoint, ranges) {
for (const [min, max] of ranges) {
@@ -89,63 +139,410 @@ const jp = (() => {
return false;
}
+ function getWanakana() {
+ try {
+ if (typeof wanakana !== 'undefined') {
+ // eslint-disable-next-line no-undef
+ return wanakana;
+ }
+ } catch (e) {
+ // NOP
+ }
+ return null;
+ }
+
- // String testing functions
+ class JapaneseUtil {
+ constructor(wanakana=null) {
+ this._wanakana = wanakana;
+ }
- function isStringEntirelyKana(str) {
- if (str.length === 0) { return false; }
- for (const c of str) {
- if (!isCodePointKana(c.codePointAt(0))) {
- return false;
+ // Character code testing functions
+
+ isCodePointKanji(codePoint) {
+ return isCodePointInRanges(codePoint, CJK_UNIFIED_IDEOGRAPHS_RANGES);
+ }
+
+ isCodePointKana(codePoint) {
+ return isCodePointInRanges(codePoint, KANA_RANGES);
+ }
+
+ isCodePointJapanese(codePoint) {
+ return isCodePointInRanges(codePoint, JAPANESE_RANGES);
+ }
+
+ // String testing functions
+
+ isStringEntirelyKana(str) {
+ if (str.length === 0) { return false; }
+ for (const c of str) {
+ if (!isCodePointInRanges(c.codePointAt(0), KANA_RANGES)) {
+ return false;
+ }
}
+ return true;
}
- return true;
- }
- function isStringPartiallyJapanese(str) {
- if (str.length === 0) { return false; }
- for (const c of str) {
- if (isCodePointJapanese(c.codePointAt(0))) {
- return true;
+ isStringPartiallyJapanese(str) {
+ if (str.length === 0) { return false; }
+ for (const c of str) {
+ if (isCodePointInRanges(c.codePointAt(0), JAPANESE_RANGES)) {
+ return true;
+ }
}
+ return false;
}
- return false;
- }
+ // Mora functions
- // Mora functions
+ isMoraPitchHigh(moraIndex, pitchAccentPosition) {
+ switch (pitchAccentPosition) {
+ case 0: return (moraIndex > 0);
+ case 1: return (moraIndex < 1);
+ default: return (moraIndex > 0 && moraIndex < pitchAccentPosition);
+ }
+ }
- function isMoraPitchHigh(moraIndex, pitchAccentPosition) {
- switch (pitchAccentPosition) {
- case 0: return (moraIndex > 0);
- case 1: return (moraIndex < 1);
- default: return (moraIndex > 0 && moraIndex < pitchAccentPosition);
+ getKanaMorae(text) {
+ const morae = [];
+ let i;
+ for (const c of text) {
+ if (SMALL_KANA_SET.has(c) && (i = morae.length) > 0) {
+ morae[i - 1] += c;
+ } else {
+ morae.push(c);
+ }
+ }
+ return morae;
}
- }
- function getKanaMorae(text) {
- const morae = [];
- let i;
- for (const c of text) {
- if (SMALL_KANA_SET.has(c) && (i = morae.length) > 0) {
- morae[i - 1] += c;
- } else {
- morae.push(c);
+ // Conversion functions
+
+ convertKatakanaToHiragana(text) {
+ const wanakana = this._getWanakana();
+ let result = '';
+ for (const c of text) {
+ if (wanakana.isKatakana(c)) {
+ result += wanakana.toHiragana(c);
+ } else {
+ result += c;
+ }
}
+
+ return result;
}
- return morae;
- }
+ convertHiraganaToKatakana(text) {
+ const wanakana = this._getWanakana();
+ let result = '';
+ for (const c of text) {
+ if (wanakana.isHiragana(c)) {
+ result += wanakana.toKatakana(c);
+ } else {
+ result += c;
+ }
+ }
+
+ return result;
+ }
+
+ convertToRomaji(text) {
+ const wanakana = this._getWanakana();
+ return wanakana.toRomaji(text);
+ }
+
+ convertReading(expression, reading, readingMode) {
+ switch (readingMode) {
+ case 'hiragana':
+ return this.convertKatakanaToHiragana(reading);
+ case 'katakana':
+ return this.convertHiraganaToKatakana(reading);
+ case 'romaji':
+ if (reading) {
+ return this.convertToRomaji(reading);
+ } else {
+ if (this.isStringEntirelyKana(expression)) {
+ return this.convertToRomaji(expression);
+ }
+ }
+ return reading;
+ case 'none':
+ return '';
+ default:
+ return reading;
+ }
+ }
+
+ convertNumericToFullWidth(text) {
+ let result = '';
+ for (const char of text) {
+ let c = char.codePointAt(0);
+ if (c >= 0x30 && c <= 0x39) { // ['0', '9']
+ c += 0xff10 - 0x30; // 0xff10 = '0' full width
+ result += String.fromCodePoint(c);
+ } else {
+ result += char;
+ }
+ }
+ return result;
+ }
+
+ convertHalfWidthKanaToFullWidth(text, sourceMap=null) {
+ let result = '';
+
+ // This function is safe to use charCodeAt instead of codePointAt, since all
+ // the relevant characters are represented with a single UTF-16 character code.
+ for (let i = 0, ii = text.length; i < ii; ++i) {
+ const c = text[i];
+ const mapping = HALFWIDTH_KATAKANA_MAPPING.get(c);
+ if (typeof mapping !== 'string') {
+ result += c;
+ continue;
+ }
+
+ let index = 0;
+ switch (text.charCodeAt(i + 1)) {
+ case 0xff9e: // dakuten
+ index = 1;
+ break;
+ case 0xff9f: // handakuten
+ index = 2;
+ break;
+ }
+
+ let c2 = mapping[index];
+ if (index > 0) {
+ if (c2 === '-') { // invalid
+ index = 0;
+ c2 = mapping[0];
+ } else {
+ ++i;
+ }
+ }
+
+ if (sourceMap !== null && index > 0) {
+ sourceMap.combine(result.length, 1);
+ }
+ result += c2;
+ }
+
+ return result;
+ }
+
+ convertAlphabeticToKana(text, sourceMap=null) {
+ let part = '';
+ let result = '';
+
+ for (const char of text) {
+ // Note: 0x61 is the character code for 'a'
+ let c = char.codePointAt(0);
+ if (c >= 0x41 && c <= 0x5a) { // ['A', 'Z']
+ c += (0x61 - 0x41);
+ } else if (c >= 0x61 && c <= 0x7a) { // ['a', 'z']
+ // NOP; c += (0x61 - 0x61);
+ } else if (c >= 0xff21 && c <= 0xff3a) { // ['A', 'Z'] fullwidth
+ c += (0x61 - 0xff21);
+ } else if (c >= 0xff41 && c <= 0xff5a) { // ['a', 'z'] fullwidth
+ c += (0x61 - 0xff41);
+ } else if (c === 0x2d || c === 0xff0d) { // '-' or fullwidth dash
+ c = 0x2d; // '-'
+ } else {
+ if (part.length > 0) {
+ result += this._convertAlphabeticPartToKana(part, sourceMap, result.length);
+ part = '';
+ }
+ result += char;
+ continue;
+ }
+ part += String.fromCodePoint(c);
+ }
+
+ if (part.length > 0) {
+ result += this._convertAlphabeticPartToKana(part, sourceMap, result.length);
+ }
+ return result;
+ }
+
+ // Furigana distribution
+
+ distributeFurigana(expression, reading) {
+ const fallback = [{furigana: reading, text: expression}];
+ if (!reading) {
+ return fallback;
+ }
+
+ let isAmbiguous = false;
+ const segmentize = (reading2, groups) => {
+ if (groups.length === 0 || isAmbiguous) {
+ return [];
+ }
+
+ const group = groups[0];
+ if (group.mode === 'kana') {
+ if (this.convertKatakanaToHiragana(reading2).startsWith(this.convertKatakanaToHiragana(group.text))) {
+ const readingLeft = reading2.substring(group.text.length);
+ const segs = segmentize(readingLeft, groups.splice(1));
+ if (segs) {
+ return [{text: group.text, furigana: ''}].concat(segs);
+ }
+ }
+ } else {
+ let foundSegments = null;
+ for (let i = reading2.length; i >= group.text.length; --i) {
+ const readingUsed = reading2.substring(0, i);
+ const readingLeft = reading2.substring(i);
+ const segs = segmentize(readingLeft, groups.slice(1));
+ if (segs) {
+ if (foundSegments !== null) {
+ // more than one way to segmentize the tail, mark as ambiguous
+ isAmbiguous = true;
+ return null;
+ }
+ foundSegments = [{text: group.text, furigana: readingUsed}].concat(segs);
+ }
+ // there is only one way to segmentize the last non-kana group
+ if (groups.length === 1) {
+ break;
+ }
+ }
+ return foundSegments;
+ }
+ };
+
+ const groups = [];
+ let modePrev = null;
+ for (const c of expression) {
+ const codePoint = c.codePointAt(0);
+ const modeCurr = this.isCodePointKanji(codePoint) || codePoint === ITERATION_MARK_CODE_POINT ? 'kanji' : 'kana';
+ if (modeCurr === modePrev) {
+ groups[groups.length - 1].text += c;
+ } else {
+ groups.push({mode: modeCurr, text: c});
+ modePrev = modeCurr;
+ }
+ }
+
+ const segments = segmentize(reading, groups);
+ if (segments && !isAmbiguous) {
+ return segments;
+ }
+ return fallback;
+ }
+
+ distributeFuriganaInflected(expression, reading, source) {
+ const output = [];
+
+ let stemLength = 0;
+ const shortest = Math.min(source.length, expression.length);
+ const sourceHiragana = this.convertKatakanaToHiragana(source);
+ const expressionHiragana = this.convertKatakanaToHiragana(expression);
+ while (stemLength < shortest && sourceHiragana[stemLength] === expressionHiragana[stemLength]) {
+ ++stemLength;
+ }
+ const offset = source.length - stemLength;
+
+ const stemExpression = source.substring(0, source.length - offset);
+ const stemReading = reading.substring(
+ 0,
+ offset === 0 ? reading.length : reading.length - expression.length + stemLength
+ );
+ for (const segment of this.distributeFurigana(stemExpression, stemReading)) {
+ output.push(segment);
+ }
+
+ if (stemLength !== source.length) {
+ output.push({text: source.substring(stemLength), furigana: ''});
+ }
+
+ return output;
+ }
+
+ // Miscellaneous
+
+ collapseEmphaticSequences(text, fullCollapse, sourceMap=null) {
+ let result = '';
+ let collapseCodePoint = -1;
+ const hasSourceMap = (sourceMap !== null);
+ for (const char of text) {
+ const c = char.codePointAt(0);
+ if (
+ c === HIRAGANA_SMALL_TSU_CODE_POINT ||
+ c === KATAKANA_SMALL_TSU_CODE_POINT ||
+ c === KANA_PROLONGED_SOUND_MARK_CODE_POINT
+ ) {
+ if (collapseCodePoint !== c) {
+ collapseCodePoint = c;
+ if (!fullCollapse) {
+ result += char;
+ continue;
+ }
+ }
+ } else {
+ collapseCodePoint = -1;
+ result += char;
+ continue;
+ }
+
+ if (hasSourceMap) {
+ sourceMap.combine(Math.max(0, result.length - 1), 1);
+ }
+ }
+ return result;
+ }
+
+ // Private
+
+ _getWanakana() {
+ const wanakana = this._wanakana;
+ if (wanakana === null) { throw new Error('Functions which use WanaKana are not supported in this context'); }
+ return wanakana;
+ }
+
+ _convertAlphabeticPartToKana(text, sourceMap, sourceMapStart) {
+ const wanakana = this._getWanakana();
+ const result = wanakana.toHiragana(text);
+
+ // Generate source mapping
+ if (sourceMap !== null) {
+ let i = 0;
+ let resultPos = 0;
+ const ii = text.length;
+ while (i < ii) {
+ // Find smallest matching substring
+ let iNext = i + 1;
+ let resultPosNext = result.length;
+ while (iNext < ii) {
+ const t = wanakana.toHiragana(text.substring(0, iNext));
+ if (t === result.substring(0, t.length)) {
+ resultPosNext = t.length;
+ break;
+ }
+ ++iNext;
+ }
+
+ // Merge characters
+ const removals = iNext - i - 1;
+ if (removals > 0) {
+ sourceMap.combine(sourceMapStart, removals);
+ }
+ ++sourceMapStart;
+
+ // Empty elements
+ const additions = resultPosNext - resultPos - 1;
+ for (let j = 0; j < additions; ++j) {
+ sourceMap.insert(sourceMapStart, 0);
+ ++sourceMapStart;
+ }
+
+ i = iNext;
+ resultPos = resultPosNext;
+ }
+ }
+
+ return result;
+ }
+ }
- // Exports
- return {
- isCodePointKanji,
- isCodePointKana,
- isCodePointJapanese,
- isStringEntirelyKana,
- isStringPartiallyJapanese,
- isMoraPitchHigh,
- getKanaMorae
- };
+ return new JapaneseUtil(getWanakana());
})();