Java 用于整个文本文件的OpenNLP句子检测API

以下是单个字符串的OpenNLP语句检测器API代码：

package opennlp;

import java.io.FileInputStream;
import java.io.FileNotFoundException;
import java.io.IOException;
import java.io.InputStream;

import opennlp.tools.sentdetect.SentenceDetectorME;
import opennlp.tools.sentdetect.SentenceModel;

public class SentenceDetector {

    public static void main(String[] args) throws FileNotFoundException {
        InputStream modelIn = new FileInputStream("en-sent.zip");
        SentenceModel model = null;
        try {
           model = new SentenceModel(modelIn);  
        }
        catch (IOException e) {
          e.printStackTrace();
        }
        finally {
          if (modelIn != null) {
            try {
              modelIn.close();
            }
            catch (IOException e) {
            }
          }
        }
        SentenceDetectorME sentenceDetector = new SentenceDetectorME(model);
           String sentences[] = sentenceDetector.sentDetect(" First sentence. Second sentence.");

           for(String str : sentences)
               System.out.println(str);
    }
}

---

现在我的问题是如何传递整个文本文件并执行句子检测，而不是单个字符串？

简单的方法：将整个文件作为字符串读取，并以通常的方式对其进行pas。以下方法将文件内容读取为字符串：

public String readFileToString(String pathToFile) throws Exception{
    StringBuilder strFile = new StringBuilder();
    BufferedReader reader = new BufferedReader(new FileReader(pathToFile));
    char[] buffer = new char[512];
    int num = 0;
    while((num = reader.read(buffer)) != -1){
        String current = String.valueOf(buffer, 0, num);
        strFile.append(current);
        buffer = new char[512];
    }
    reader.close();
    return strFile.toString();
}

---

将文件读入字符串，然后像以前一样继续。@andreThompson查看最后一行编辑+我希望你能得到答案。：）也许已经晚了，但对其他人可能有用；ApacheTika可用于从不同的文件类型中提取元数据和文本，尤其是您的案例中的文本文件。