diff options
author | siikamiika <siikamiika@users.noreply.github.com> | 2017-10-01 04:17:02 +0300 |
---|---|---|
committer | siikamiika <siikamiika@users.noreply.github.com> | 2017-10-01 04:17:02 +0300 |
commit | 803ccc1661384ee484298f3495585d3923495dd4 (patch) | |
tree | 77b4311d85ff57d5f8c2fcadecb9a2e883d90025 /ext/bg/js/translator.js | |
parent | b24c70523479c16eca848f5aafaa887549689ac6 (diff) |
merged: first draft, only supports JMdict
Diffstat (limited to 'ext/bg/js/translator.js')
-rw-r--r-- | ext/bg/js/translator.js | 69 |
1 files changed, 66 insertions, 3 deletions
diff --git a/ext/bg/js/translator.js b/ext/bg/js/translator.js index 8fa2b60b..db287d63 100644 --- a/ext/bg/js/translator.js +++ b/ext/bg/js/translator.js @@ -49,15 +49,78 @@ class Translator { } async findTermsMerged(text, dictionaries, alphanumeric) { - const titles = Object.keys(dictionaries); + // const titles = Object.keys(dictionaries); const {length, definitions} = await this.findTerms(text, dictionaries, alphanumeric); - const definitionsMerged = dictTermsGroup(definitions, dictionaries); + // const definitionsMerged = dictTermsMerge(definitions, dictionaries, this.database); // for (const definition of definitionsMerged) { // await this.buildTermFrequencies(definition, titles); // } - return {length, definitions: definitionsMerged}; + const sequences = {}; + const stray = []; + for (const definition of definitions) { + if (typeof definition.sequence !== 'undefined') { + if (!sequences[definition.sequence]) { + sequences[definition.sequence] = { + reasons: definition.reasons, + score: Number.MIN_SAFE_INTEGER, + expression: new Set(), + reading: new Set(), + source: definition.source, + definitions: [] + }; + } + const seq = sequences[definition.sequence]; + seq.score = Math.max(seq.score, definition.score); + } else { + stray.push(definition); + } + } + + const definitionsMerged = dictTermsGroup(stray, dictionaries); + for (const sequence in sequences) { + const entry = await this.database.findEntry(Number(sequence)); + + const result = sequences[sequence]; + const glossaries = new Map(); + for (const definition of entry) { + + const gloss = definition.glossary.join('||'); + if (!glossaries.get(gloss)) { + const tags = await this.expandTags(definition.tags, definition.dictionary); + tags.push(dictTagBuildSource(definition.dictionary)); + glossaries.set(gloss, { + expressions: new Set(), + readings: new Set(), + tags: dictTagsSort(tags), // TODO: use correct tags + source: result.source, + reasons: [], + score: definition.score, + id: definition.id, + dictionary: definition.dictionary + }); + } + glossaries.get(gloss).expressions.add(definition.expression); + glossaries.get(gloss).readings.add(definition.reading); + + result.expression.add(definition.expression); + result.reading.add(definition.reading); + } + + for (const gloss of glossaries.keys()) { + const definition = glossaries.get(gloss); + definition.glossary = gloss.split('||'); + result.definitions.push(definition); + } + //dictTermsSort(groupDefs, dictionaries) + + result.expression = Array.from(result.expression).join(', '); + result.reading = Array.from(result.reading).join(', '); + definitionsMerged.push(result); + } + + return {length, definitions: dictTermsSort(definitionsMerged)}; } async findTermsSplit(text, dictionaries, alphanumeric) { |