Morphological Analysis with Swift

FoundationSample.swift

// ref: https://developer.apple.com/documentation/foundation/nslinguistictagger
// ref: https://developer.apple.com/documentation/foundation/nslinguistictagger/1410036-enumeratetags
// ref: https://dev.classmethod.jp/articles/ios10-morphological-analysis-from-speechrecognizer/

import Foundation

private func analyzeText(_ text: String, scheme: NSLinguisticTagScheme) {
    let tagger = NSLinguisticTagger(tagSchemes: NSLinguisticTagger.availableTagSchemes(forLanguage: "ja"), options: 0)
    tagger.string = text
    tagger.enumerateTags(
        in: NSRange(location: 0, length: text.count),
        scheme: scheme,
        options: [.omitWhitespace]
    ) { tag, tokenRange, sentenceRange, stop in
        let subString = (text as NSString).substring(with: tokenRange)
        if let tag = tag {
            print("\(subString): \(tag.rawValue)")
        }
    }
}

analyzeText("私はウホーイです。あなたは誰ですか？", scheme: .tokenType) // 形態素解析
analyzeText("私はウホーイです。あなたは誰ですか？", scheme: .lexicalClass) // 品詞分解 // !!!: 日本語非対応

FoundationSampleResult.txt

私: Word
は: Word
ウ: Word
ホーイ: Word
です: Word
。: Punctuation
あなた: Word
は: Word
誰: Word
です: Word
か: Word
？: Punctuation

MorphologicalAnalysisSample.swift

import NaturalLanguage

private func analyzeText(_ text: String, scheme: NLTagScheme) -> [String] {
    var results: [String] = []
    
    let tagger = NLTagger(tagSchemes: NLTagger.availableTagSchemes(for: .word, language: .japanese))
    tagger.string = text
    let options: NLTagger.Options = [.omitPunctuation, .omitWhitespace]
    tagger.enumerateTags(in: text.startIndex..<text.endIndex, unit: .word, scheme: scheme, options: options) { tag, tokenRange in
        results.append("\(text[tokenRange])")
        return true
    }
    
    return results
}

print(analyzeText("私はウホーイです。あなたは誰ですか？", scheme: .tokenType))

MorphologicalAnalysisSampleResult.txt

["私", "は", "ウ", "ホーイ", "です", "あなた", "は", "誰", "です", "か"]

NaturalLanguageSample.swift

// ref: https://developer.apple.com/documentation/naturallanguage
// ref: https://developer.apple.com/documentation/naturallanguage/identifying_parts_of_speech
// ref: https://developer.apple.com/forums/thread/669890
// ref: https://dev.classmethod.jp/articles/ios12-natural-language-tokenizing/
// ref: https://dev.classmethod.jp/articles/ios12-natural-language-identifying-parts-of-speech/

import NaturalLanguage

private func analyzeText(_ text: String, scheme: NLTagScheme) {
    let tagger = NLTagger(tagSchemes: NLTagger.availableTagSchemes(for: .word, language: .japanese))
    tagger.string = text
    let options: NLTagger.Options = [/* .omitPunctuation, */.omitWhitespace]
    tagger.enumerateTags(in: text.startIndex..<text.endIndex, unit: .word, scheme: scheme, options: options) { tag, tokenRange in
        if let tag = tag {
            print("\(text[tokenRange]): \(tag.rawValue)")
        }
        return true
    }
}

analyzeText("私はウホーイです。あなたは誰ですか？", scheme: .tokenType) // 形態素解析
analyzeText("私はウホーイです。あなたは誰ですか？", scheme: .lexicalClass) // 品詞分解 // !!!: 日本語非対応

NaturalLanguageSampleResult.txt

私: Word
は: Word
ウ: Word
ホーイ: Word
です: Word
。: Punctuation
あなた: Word
は: Word
誰: Word
です: Word
か: Word
？: Punctuation