diff options
Diffstat (limited to 'ext/mixed/js/japanese.js')
-rw-r--r-- | ext/mixed/js/japanese.js | 295 |
1 files changed, 285 insertions, 10 deletions
diff --git a/ext/mixed/js/japanese.js b/ext/mixed/js/japanese.js index c9bc7b87..0da822d7 100644 --- a/ext/mixed/js/japanese.js +++ b/ext/mixed/js/japanese.js @@ -17,24 +17,153 @@ */ -function jpIsKanji(c) { - const code = c.charCodeAt(0); - return code >= 0x4e00 && code < 0x9fb0 || code >= 0x3400 && code < 0x4dc0; +const JP_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--'] +]); + +const JP_HIRAGANA_RANGE = [0x3040, 0x309f]; +const JP_KATAKANA_RANGE = [0x30a0, 0x30ff]; +const JP_KANA_RANGES = [JP_HIRAGANA_RANGE, JP_KATAKANA_RANGE]; + +const JP_CJK_COMMON_RANGE = [0x4e00, 0x9fff]; +const JP_CJK_RARE_RANGE = [0x3400, 0x4dbf]; +const JP_CJK_RANGES = [JP_CJK_COMMON_RANGE, JP_CJK_RARE_RANGE]; + +const JP_ITERATION_MARK_CHAR_CODE = 0x3005; + +// Japanese character ranges, roughly ordered in order of expected frequency +const JP_JAPANESE_RANGES = [ + JP_HIRAGANA_RANGE, + JP_KATAKANA_RANGE, + + JP_CJK_COMMON_RANGE, + JP_CJK_RARE_RANGE, + + [0xff66, 0xff9f], // Halfwidth katakana + + [0x30fb, 0x30fc], // Katakana punctuation + [0xff61, 0xff65], // Kana punctuation + [0x3000, 0x303f], // CJK punctuation + + [0xff10, 0xff19], // Fullwidth numbers + [0xff21, 0xff3a], // Fullwidth upper case Latin letters + [0xff41, 0xff5a], // Fullwidth lower case Latin letters + + [0xff01, 0xff0f], // Fullwidth punctuation 1 + [0xff1a, 0xff1f], // Fullwidth punctuation 2 + [0xff3b, 0xff3f], // Fullwidth punctuation 3 + [0xff5b, 0xff60], // Fullwidth punctuation 4 + [0xffe0, 0xffee], // Currency markers +]; + + +// Helper functions + +function _jpIsCharCodeInRanges(charCode, ranges) { + for (const [min, max] of ranges) { + if (charCode >= min && charCode <= max) { + return true; + } + } + return false; } -function jpIsKana(c) { - return wanakana.isKana(c); + +// Character code testing functions + +function jpIsCharCodeKanji(charCode) { + return _jpIsCharCodeInRanges(charCode, JP_CJK_RANGES); } -function jpIsJapaneseText(text) { - for (const c of text) { - if (jpIsKanji(c) || jpIsKana(c)) { +function jpIsCharCodeKana(charCode) { + return _jpIsCharCodeInRanges(charCode, JP_KANA_RANGES); +} + +function jpIsCharCodeJapanese(charCode) { + return _jpIsCharCodeInRanges(charCode, JP_JAPANESE_RANGES); +} + + +// String testing functions + +function jpIsStringEntirelyKana(str) { + if (str.length === 0) { return false; } + for (let i = 0, ii = str.length; i < ii; ++i) { + if (!jpIsCharCodeKana(str.charCodeAt(i))) { + return false; + } + } + return true; +} + +function jpIsStringPartiallyJapanese(str) { + if (str.length === 0) { return false; } + for (let i = 0, ii = str.length; i < ii; ++i) { + if (jpIsCharCodeJapanese(str.charCodeAt(i))) { return true; } } return false; } + +// Conversion functions + function jpKatakanaToHiragana(text) { let result = ''; for (const c of text) { @@ -75,7 +204,7 @@ function jpConvertReading(expressionFragment, readingFragment, readingMode) { if (readingFragment) { return jpToRomaji(readingFragment); } else { - if (jpIsKana(expressionFragment)) { + if (jpIsStringEntirelyKana(expressionFragment)) { return jpToRomaji(expressionFragment); } } @@ -134,7 +263,8 @@ function jpDistributeFurigana(expression, reading) { const groups = []; let modePrev = null; for (const c of expression) { - const modeCurr = jpIsKanji(c) || c.charCodeAt(0) === 0x3005 /* noma */ ? 'kanji' : 'kana'; + const charCode = c.charCodeAt(0); + const modeCurr = jpIsCharCodeKanji(charCode) || charCode === JP_ITERATION_MARK_CHAR_CODE ? 'kanji' : 'kana'; if (modeCurr === modePrev) { groups[groups.length - 1].text += c; } else { @@ -177,3 +307,148 @@ function jpDistributeFuriganaInflected(expression, reading, source) { return output; } + +function jpConvertHalfWidthKanaToFullWidth(text, sourceMapping) { + let result = ''; + const ii = text.length; + const hasSourceMapping = Array.isArray(sourceMapping); + + for (let i = 0; i < ii; ++i) { + const c = text[i]; + const mapping = JP_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 (hasSourceMapping && index > 0) { + index = result.length; + const v = sourceMapping.splice(index + 1, 1)[0]; + sourceMapping[index] += v; + } + result += c2; + } + + return result; +} + +function jpConvertNumericTofullWidth(text) { + let result = ''; + for (let i = 0, ii = text.length; i < ii; ++i) { + let c = text.charCodeAt(i); + if (c >= 0x30 && c <= 0x39) { // ['0', '9'] + c += 0xff10 - 0x30; // 0xff10 = '0' full width + result += String.fromCharCode(c); + } else { + result += text[i]; + } + } + return result; +} + +function jpConvertAlphabeticToKana(text, sourceMapping) { + let part = ''; + let result = ''; + const ii = text.length; + + if (sourceMapping.length === ii) { + sourceMapping.length = ii; + sourceMapping.fill(1); + } + + for (let i = 0; i < ii; ++i) { + // Note: 0x61 is the character code for 'a' + let c = text.charCodeAt(i); + 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 += jpToHiragana(part, sourceMapping, result.length); + part = ''; + } + result += text[i]; + continue; + } + part += String.fromCharCode(c); + } + + if (part.length > 0) { + result += jpToHiragana(part, sourceMapping, result.length); + } + return result; +} + +function jpToHiragana(text, sourceMapping, sourceMappingStart) { + const result = wanakana.toHiragana(text); + + // Generate source mapping + if (Array.isArray(sourceMapping)) { + if (typeof sourceMappingStart !== 'number') { sourceMappingStart = 0; } + 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) { + let sum = 0; + const vs = sourceMapping.splice(sourceMappingStart + 1, removals); + for (const v of vs) { sum += v; } + sourceMapping[sourceMappingStart] += sum; + } + ++sourceMappingStart; + + // Empty elements + const additions = resultPosNext - resultPos - 1; + for (let j = 0; j < additions; ++j) { + sourceMapping.splice(sourceMappingStart, 0, 0); + ++sourceMappingStart; + } + + i = iNext; + resultPos = resultPosNext; + } + } + + return result; +} |