package cn.edu.bjut.textmining.chapter2;

import java.text.BreakIterator;
import java.util.ArrayList;
import java.util.List;
import java.util.Locale;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

public class SentenceSplitDemo {
    private static final String DOT_PLACEHOLDER = "<DOT>";
    private static final Pattern COMMON_TITLE = Pattern.compile(
            "\\b(Mr|Mrs|Ms|Dr|Prof|Sr|Jr)\\.", Pattern.CASE_INSENSITIVE);

    public static void main(String[] args) {
        String text = "Mr. Smith likes Java. Text mining is useful!";
        System.out.println(splitSentences(text));
    }

    public static List<String> splitSentences(String text) {
        List<String> sentences = new ArrayList<String>();
        if (text == null || text.trim().length() == 0) {
            return sentences;
        }

        String protectedText = protectCommonTitles(text);
        BreakIterator iterator = BreakIterator.getSentenceInstance(Locale.US);
        iterator.setText(protectedText);
        int start = iterator.first();
        int end = iterator.next();
        while (end != BreakIterator.DONE) {
            String sentence = protectedText.substring(start, end)
                    .replace(DOT_PLACEHOLDER, ".")
                    .trim();
            if (sentence.length() > 0) {
                sentences.add(sentence);
            }
            start = end;
            end = iterator.next();
        }
        return sentences;
    }

    private static String protectCommonTitles(String text) {
        Matcher matcher = COMMON_TITLE.matcher(text);
        StringBuffer buffer = new StringBuffer();
        while (matcher.find()) {
            matcher.appendReplacement(
                    buffer,
                    Matcher.quoteReplacement(matcher.group(1) + DOT_PLACEHOLDER));
        }
        matcher.appendTail(buffer);
        return buffer.toString();
    }
}