分两步：

第一步：实现hankcs.hanlp/corpus.io.IIOAdapter

1.public class HadoopFileIoAdapter implements IIOAdapter {

2.

3.@Override

4.public InputStream open(String path) throws IOException {

5.Configuration conf = new Configuration();

6.FileSystem fs = FileSystem.get(URI.create(path), conf);

7.return fs.open(new Path(path));

8.}

9.

10.@Override

11.public OutputStream create(String path) throws IOException {

12.Configuration conf = new Configuration();

13.FileSystem fs = FileSystem.get(URI.create(path), conf);

14.OutputStream out = fs.create(new Path(path));

15.return out;

16.}

17. }

第二步：修改配置文件。root为hdfs上的数据包，把IOAdapter改为咱们上面实现的类

ok，这样你就能在分布式集群上使用hanlp进行分词了。