从单词到字符的 Ngram 生成

Question

如何从 JAVA 中的单词序列生成 ngram？

string = '麦当劳世界'

n=3

输出应该是

['McD'、'cDo'、'Don'、'ona'、'nal'、'ald'、'lds'、'dsw'、'swo'、'wor'、'orl'、'rld ']

用 Java 给出一些想法，如何实现它。

Answer 1

import java.util.ArrayList;
import java.util.Collections;
import java.util.List;

public class NgramGenerator {

    public static void main(String[] args) {
        System.out.println(generateNgrams("ABC", 3));
        System.out.println(generateNgrams("McDonalds world", 3));
        System.out.println(generateNgrams("Mc Donalds world", 3));
        System.out.println(generateNgrams("McDonalds ", 1));
        System.out.println(generateNgrams("M  ", 1));
        System.out.println(generateNgrams("M ", 1));
        System.out.println(generateNgrams(" ", 1));
        System.out.println(generateNgrams("", 1));
        System.out.println(generateNgrams(null, 1));
    }

    private static List<String> generateNgrams(String word, int ngram) {

        if (ngram == 0 || word == null) {
            return Collections.emptyList();
        }

        word = word.replace(" ", "");

        if (word.isEmpty()) {
            return Collections.emptyList();
        }

        if (ngram > word.length()) {
            throw new IllegalArgumentException("Invalid word or n-gram");
        }

        List<String> ngrams = new ArrayList<>();
        for (int i = 0; i <= word.length() - ngram; i++) {
            ngrams.add(word.substring(i, i + ngram).trim());
        }
        return ngrams;
    }
}