aboutsummaryrefslogtreecommitdiff
path: root/ext
diff options
context:
space:
mode:
Diffstat (limited to 'ext')
-rw-r--r--ext/bg/data/options-schema.json38
-rw-r--r--ext/bg/js/handlebars.js2
-rw-r--r--ext/bg/js/options.js8
-rw-r--r--ext/bg/js/search.js2
-rw-r--r--ext/bg/js/settings/main.js12
-rw-r--r--ext/bg/js/translator.js138
-rw-r--r--ext/bg/settings.html75
-rw-r--r--ext/mixed/js/japanese.js295
8 files changed, 529 insertions, 41 deletions
diff --git a/ext/bg/data/options-schema.json b/ext/bg/data/options-schema.json
index d4bd3c21..a20a0619 100644
--- a/ext/bg/data/options-schema.json
+++ b/ext/bg/data/options-schema.json
@@ -65,6 +65,7 @@
"general",
"audio",
"scanning",
+ "translation",
"dictionaries",
"parsing",
"anki"
@@ -350,6 +351,43 @@
}
}
},
+ "translation": {
+ "type": "object",
+ "required": [
+ "convertHalfWidthCharacters",
+ "convertNumericCharacters",
+ "convertAlphabeticCharacters",
+ "convertHiraganaToKatakana",
+ "convertKatakanaToHiragana"
+ ],
+ "properties": {
+ "convertHalfWidthCharacters": {
+ "type": "string",
+ "enum": ["false", "true", "variant"],
+ "default": "false"
+ },
+ "convertNumericCharacters": {
+ "type": "string",
+ "enum": ["false", "true", "variant"],
+ "default": "false"
+ },
+ "convertAlphabeticCharacters": {
+ "type": "string",
+ "enum": ["false", "true", "variant"],
+ "default": "false"
+ },
+ "convertHiraganaToKatakana": {
+ "type": "string",
+ "enum": ["false", "true", "variant"],
+ "default": "false"
+ },
+ "convertKatakanaToHiragana": {
+ "type": "string",
+ "enum": ["false", "true", "variant"],
+ "default": "variant"
+ }
+ }
+ },
"dictionaries": {
"type": "object",
"additionalProperties": {
diff --git a/ext/bg/js/handlebars.js b/ext/bg/js/handlebars.js
index 6d1581be..62f89ee4 100644
--- a/ext/bg/js/handlebars.js
+++ b/ext/bg/js/handlebars.js
@@ -61,7 +61,7 @@ function handlebarsFuriganaPlain(options) {
function handlebarsKanjiLinks(options) {
let result = '';
for (const c of options.fn(this)) {
- if (jpIsKanji(c)) {
+ if (jpIsCharCodeKanji(c.charCodeAt(0))) {
result += `<a href="#" class="kanji-link">${c}</a>`;
} else {
result += c;
diff --git a/ext/bg/js/options.js b/ext/bg/js/options.js
index c2da76b1..120b34af 100644
--- a/ext/bg/js/options.js
+++ b/ext/bg/js/options.js
@@ -319,6 +319,14 @@ function profileOptionsCreateDefaults() {
enableOnSearchPage: true
},
+ translation: {
+ convertHalfWidthCharacters: 'false',
+ convertNumericCharacters: 'false',
+ convertAlphabeticCharacters: 'false',
+ convertHiraganaToKatakana: 'false',
+ convertKatakanaToHiragana: 'variant'
+ },
+
dictionaries: {},
parsing: {
diff --git a/ext/bg/js/search.js b/ext/bg/js/search.js
index 673f066b..f5c641a8 100644
--- a/ext/bg/js/search.js
+++ b/ext/bg/js/search.js
@@ -265,7 +265,7 @@ class DisplaySearch extends Display {
text !== this.clipboardPreviousText
) {
this.clipboardPreviousText = text;
- if (jpIsJapaneseText(text)) {
+ if (jpIsStringPartiallyJapanese(text)) {
this.setQuery(this.isWanakanaEnabled() ? window.wanakana.toKana(text) : text);
window.history.pushState(null, '', `${window.location.pathname}?query=${encodeURIComponent(text)}`);
this.onSearchQueryUpdated(this.query.value, true);
diff --git a/ext/bg/js/settings/main.js b/ext/bg/js/settings/main.js
index bed57f7d..b2ac82f9 100644
--- a/ext/bg/js/settings/main.js
+++ b/ext/bg/js/settings/main.js
@@ -72,6 +72,12 @@ async function formRead(options) {
options.scanning.modifier = $('#scan-modifier-key').val();
options.scanning.popupNestingMaxDepth = parseInt($('#popup-nesting-max-depth').val(), 10);
+ options.translation.convertHalfWidthCharacters = $('#translation-convert-half-width-characters').val();
+ options.translation.convertNumericCharacters = $('#translation-convert-numeric-characters').val();
+ options.translation.convertAlphabeticCharacters = $('#translation-convert-alphabetic-characters').val();
+ options.translation.convertHiraganaToKatakana = $('#translation-convert-hiragana-to-katakana').val();
+ options.translation.convertKatakanaToHiragana = $('#translation-convert-katakana-to-hiragana').val();
+
options.parsing.enableScanningParser = $('#parsing-scan-enable').prop('checked');
options.parsing.enableMecabParser = $('#parsing-mecab-enable').prop('checked');
options.parsing.termSpacing = $('#parsing-term-spacing').prop('checked');
@@ -141,6 +147,12 @@ async function formWrite(options) {
$('#scan-modifier-key').val(options.scanning.modifier);
$('#popup-nesting-max-depth').val(options.scanning.popupNestingMaxDepth);
+ $('#translation-convert-half-width-characters').val(options.translation.convertHalfWidthCharacters);
+ $('#translation-convert-numeric-characters').val(options.translation.convertNumericCharacters);
+ $('#translation-convert-alphabetic-characters').val(options.translation.convertAlphabeticCharacters);
+ $('#translation-convert-hiragana-to-katakana').val(options.translation.convertHiraganaToKatakana);
+ $('#translation-convert-katakana-to-hiragana').val(options.translation.convertKatakanaToHiragana);
+
$('#parsing-scan-enable').prop('checked', options.parsing.enableScanningParser);
$('#parsing-mecab-enable').prop('checked', options.parsing.enableMecabParser);
$('#parsing-reading-mode').val(options.parsing.readingMode);
diff --git a/ext/bg/js/translator.js b/ext/bg/js/translator.js
index b6e9604d..0f89111f 100644
--- a/ext/bg/js/translator.js
+++ b/ext/bg/js/translator.js
@@ -151,7 +151,7 @@ class Translator {
async findTermsGrouped(text, details, options) {
const dictionaries = dictEnabledSet(options);
const titles = Object.keys(dictionaries);
- const [definitions, length] = await this.findTermsInternal(text, dictionaries, options.scanning.alphanumeric, details);
+ const [definitions, length] = await this.findTermsInternal(text, dictionaries, details, options);
const definitionsGrouped = dictTermsGroup(definitions, dictionaries);
await this.buildTermFrequencies(definitionsGrouped, titles);
@@ -169,7 +169,7 @@ class Translator {
const dictionaries = dictEnabledSet(options);
const secondarySearchTitles = Object.keys(options.dictionaries).filter((dict) => options.dictionaries[dict].allowSecondarySearches);
const titles = Object.keys(dictionaries);
- const [definitions, length] = await this.findTermsInternal(text, dictionaries, options.scanning.alphanumeric, details);
+ const [definitions, length] = await this.findTermsInternal(text, dictionaries, details, options);
const {sequencedDefinitions, defaultDefinitions} = await this.getSequencedDefinitions(definitions, options.general.mainDictionary);
const definitionsMerged = [];
const mergedByTermIndices = new Set();
@@ -206,26 +206,24 @@ class Translator {
async findTermsSplit(text, details, options) {
const dictionaries = dictEnabledSet(options);
const titles = Object.keys(dictionaries);
- const [definitions, length] = await this.findTermsInternal(text, dictionaries, options.scanning.alphanumeric, details);
+ const [definitions, length] = await this.findTermsInternal(text, dictionaries, details, options);
await this.buildTermFrequencies(definitions, titles);
return [definitions, length];
}
- async findTermsInternal(text, dictionaries, alphanumeric, details) {
- if (!alphanumeric && text.length > 0) {
- const c = text[0];
- if (!jpIsKana(c) && !jpIsKanji(c)) {
- return [[], 0];
- }
+ async findTermsInternal(text, dictionaries, details, options) {
+ text = Translator.getSearchableText(text, options);
+ if (text.length === 0) {
+ return [[], 0];
}
const titles = Object.keys(dictionaries);
const deinflections = (
details.wildcard ?
await this.findTermWildcard(text, titles, details.wildcard) :
- await this.findTermDeinflections(text, titles)
+ await this.findTermDeinflections(text, titles, options)
);
let definitions = [];
@@ -240,6 +238,7 @@ class Translator {
definitions.push({
source: deinflection.source,
+ rawSource: deinflection.rawSource,
reasons: deinflection.reasons,
score: definition.score,
id: definition.id,
@@ -260,7 +259,7 @@ class Translator {
let length = 0;
for (const definition of definitions) {
- length = Math.max(length, definition.source.length);
+ length = Math.max(length, definition.rawSource.length);
}
return [definitions, length];
@@ -274,6 +273,7 @@ class Translator {
return [{
source: text,
+ rawSource: text,
term: text,
rules: 0,
definitions,
@@ -281,9 +281,8 @@ class Translator {
}];
}
- async findTermDeinflections(text, titles) {
- const text2 = jpKatakanaToHiragana(text);
- const deinflections = (text === text2 ? this.getDeinflections(text) : this.getDeinflections2(text, text2));
+ async findTermDeinflections(text, titles, options) {
+ const deinflections = this.getAllDeinflections(text, options);
if (deinflections.length === 0) {
return [];
@@ -321,30 +320,77 @@ class Translator {
return deinflections.filter((e) => e.definitions.length > 0);
}
- getDeinflections(text) {
+ getAllDeinflections(text, options) {
+ const translationOptions = options.translation;
+ const textOptionVariantArray = [
+ Translator.getTextOptionEntryVariants(translationOptions.convertHalfWidthCharacters),
+ Translator.getTextOptionEntryVariants(translationOptions.convertNumericCharacters),
+ Translator.getTextOptionEntryVariants(translationOptions.convertAlphabeticCharacters),
+ Translator.getTextOptionEntryVariants(translationOptions.convertHiraganaToKatakana),
+ Translator.getTextOptionEntryVariants(translationOptions.convertKatakanaToHiragana)
+ ];
+
const deinflections = [];
+ const used = new Set();
+ for (const [halfWidth, numeric, alphabetic, katakana, hiragana] of Translator.getArrayVariants(textOptionVariantArray)) {
+ let text2 = text;
+ let sourceMapping = null;
+ if (halfWidth) {
+ if (sourceMapping === null) { sourceMapping = Translator.createTextSourceMapping(text2); }
+ text2 = jpConvertHalfWidthKanaToFullWidth(text2, sourceMapping);
+ }
+ if (numeric) {
+ text2 = jpConvertNumericTofullWidth(text2);
+ }
+ if (alphabetic) {
+ if (sourceMapping === null) { sourceMapping = Translator.createTextSourceMapping(text2); }
+ text2 = jpConvertAlphabeticToKana(text2, sourceMapping);
+ }
+ if (katakana) {
+ text2 = jpHiraganaToKatakana(text2);
+ }
+ if (hiragana) {
+ text2 = jpKatakanaToHiragana(text2);
+ }
- for (let i = text.length; i > 0; --i) {
- const textSubstring = text.substring(0, i);
- deinflections.push(...this.deinflector.deinflect(textSubstring));
+ for (let i = text2.length; i > 0; --i) {
+ const text2Substring = text2.substring(0, i);
+ if (used.has(text2Substring)) { break; }
+ used.add(text2Substring);
+ for (const deinflection of this.deinflector.deinflect(text2Substring)) {
+ deinflection.rawSource = Translator.getDeinflectionRawSource(text, i, sourceMapping);
+ deinflections.push(deinflection);
+ }
+ }
}
-
return deinflections;
}
- getDeinflections2(text1, text2) {
- const deinflections = [];
+ static getTextOptionEntryVariants(value) {
+ switch (value) {
+ case 'true': return [true];
+ case 'variant': return [false, true];
+ default: return [false];
+ }
+ }
- for (let i = text1.length; i > 0; --i) {
- const text1Substring = text1.substring(0, i);
- const text2Substring = text2.substring(0, i);
- deinflections.push(...this.deinflector.deinflect(text1Substring));
- if (text1Substring !== text2Substring) {
- deinflections.push(...this.deinflector.deinflect(text2Substring));
- }
+ static getDeinflectionRawSource(source, length, sourceMapping) {
+ if (sourceMapping === null) {
+ return source.substring(0, length);
}
- return deinflections;
+ let result = '';
+ let index = 0;
+ for (let i = 0; i < length; ++i) {
+ const c = sourceMapping[i];
+ result += source.substring(index, index + c);
+ index += c;
+ }
+ return result;
+ }
+
+ static createTextSourceMapping(text) {
+ return new Array(text.length).fill(1);
}
async findKanji(text, options) {
@@ -527,4 +573,38 @@ class Translator {
const pos = name.indexOf(':');
return (pos >= 0 ? name.substring(0, pos) : name);
}
+
+ static *getArrayVariants(arrayVariants) {
+ const ii = arrayVariants.length;
+
+ let total = 1;
+ for (let i = 0; i < ii; ++i) {
+ total *= arrayVariants[i].length;
+ }
+
+ for (let a = 0; a < total; ++a) {
+ const variant = [];
+ let index = a;
+ for (let i = 0; i < ii; ++i) {
+ const entryVariants = arrayVariants[i];
+ variant.push(entryVariants[index % entryVariants.length]);
+ index = Math.floor(index / entryVariants.length);
+ }
+ yield variant;
+ }
+ }
+
+ static getSearchableText(text, options) {
+ if (!options.scanning.alphanumeric) {
+ const ii = text.length;
+ for (let i = 0; i < ii; ++i) {
+ if (!jpIsCharCodeJapanese(text.charCodeAt(i))) {
+ text = text.substring(0, i);
+ break;
+ }
+ }
+ }
+
+ return text;
+ }
}
diff --git a/ext/bg/settings.html b/ext/bg/settings.html
index f73f79c8..3e06d4b5 100644
--- a/ext/bg/settings.html
+++ b/ext/bg/settings.html
@@ -397,6 +397,81 @@
</div>
</div>
+ <div>
+ <h3>Translation Options</h3>
+
+ <p class="help-block">
+ The following options can be used during the translation process to provide alternate versions of the input text to search for.
+ This can be helpful when the input text doesn't exactly match the term or expression found in the database.
+ </p>
+
+ <p class="help-block">
+ The conversion options below are listed in the order that the conversions are applied to the input text.
+ Each conversion has three possible values:
+ </p>
+
+ <ul class="help-block">
+ <li>
+ <strong>Disabled</strong><br>
+ This conversion will never be applied to the input text.
+ </li>
+ <li>
+ <strong>Enabled</strong><br>
+ This conversion will always be applied to the input text.
+ </li>
+ <li>
+ <strong>Use both variants</strong><br>
+ The translator will check the database for two variations: the raw input text and the converted input text.
+ When multiple options use variants, the translator will search for combinations of the converted text.
+ </li>
+ </ul>
+
+ <div class="form-group">
+ <label for="translation-convert-half-width-characters">Convert half width characters to full width <span class="label-light">(ﾖﾐﾁｬﾝ &rarr; ヨミチャン)</span></label>
+ <select class="form-control" id="translation-convert-half-width-characters">
+ <option value="false">Disabled</option>
+ <option value="true">Enabled</option>
+ <option value="variant">Use both variants</option>
+ </select>
+ </div>
+
+ <div class="form-group">
+ <label for="translation-convert-numeric-characters">Convert numeric characters to full width <span class="label-light">(1234 &rarr; １２３４)</span></label>
+ <select class="form-control" id="translation-convert-numeric-characters">
+ <option value="false">Disabled</option>
+ <option value="true">Enabled</option>
+ <option value="variant">Use both variants</option>
+ </select>
+ </div>
+
+ <div class="form-group">
+ <label for="translation-convert-alphabetic-characters">Convert alphabetic characters to hiragana <span class="label-light">(yomichan &rarr; よみちゃん)</span></label>
+ <select class="form-control" id="translation-convert-alphabetic-characters">
+ <option value="false">Disabled</option>
+ <option value="true">Enabled</option>
+ <option value="variant">Use both variants</option>
+ </select>
+ </div>
+
+ <div class="form-group">
+ <label for="translation-convert-hiragana-to-katakana">Convert hiragana to katakana <span class="label-light">(よみちゃん &rarr; ヨミチャン)</span></label>
+ <select class="form-control" id="translation-convert-hiragana-to-katakana">
+ <option value="false">Disabled</option>
+ <option value="true">Enabled</option>
+ <option value="variant">Use both variants</option>
+ </select>
+ </div>
+
+ <div class="form-group">
+ <label for="translation-convert-katakana-to-hiragana">Convert katakana to hiragana <span class="label-light">(ヨミチャン &rarr; よみちゃん)</span></label>
+ <select class="form-control" id="translation-convert-katakana-to-hiragana">
+ <option value="false">Disabled</option>
+ <option value="true">Enabled</option>
+ <option value="variant">Use both variants</option>
+ </select>
+ </div>
+ </div>
+
<div id="popup-content-scanning">
<h3>Popup Content Scanning Options</h3>
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;
+}