org.ansj.dic.impl.File2Stream Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of ansj_seg Show documentation
best java chinese word seg !
The newest version!
package org.ansj.dic.impl;

import org.ansj.dic.PathToStream;
import org.ansj.exception.LibraryException;
import org.nlpcn.commons.lang.util.IOUtil;
import org.nlpcn.commons.lang.util.logging.Log;
import org.nlpcn.commons.lang.util.logging.LogFactory;

import java.io.*;
import java.nio.charset.StandardCharsets;
import java.util.Vector;

/**
 * 将文件转换为流 file://c:/dic.txt
 * 
 * @author ansj
 *
 */
public class File2Stream extends PathToStream {

	private static final Log LOG = LogFactory.getLog(File2Stream.class);

	@Override
	public InputStream toStream(String path) {
		LOG.info("path to stream " + path);

		if (path.startsWith("file://")) {
			path = path.substring(7);
		}

		File file = new File(path);

		if (file.exists() && file.canRead()) {

			try {
				if (file.isDirectory()) {
					return multiple(path);
				} else {
					return new FileInputStream(file);
				}
			} catch (Exception e) {
				throw new LibraryException(e);
			}
		}
		throw new LibraryException(" path :" + path + " file:" + file.getAbsolutePath() + " not found or can not to read");

	}

	private InputStream multiple(String path) throws FileNotFoundException {
		File[] libs = new File[0];

		File file = new File(path);

		if (file.exists() && file.canRead()) {
			if (file.isFile()) {
				libs = new File[1];
				libs[0] = file;
			} else if (file.isDirectory()) {

				File[] files = file.listFiles(new FileFilter() {
					@Override
					public boolean accept(File file) {
						return file.canRead() && !file.isHidden() && !file.isDirectory();
					}
				});

				if (files != null && files.length > 0) {
					libs = files;
				}
			}
		}

		if (libs.length == 0) {
			throw new LibraryException("not find any file in path : " + path);
		}

		if (libs.length == 1) {
			return new FileInputStream(libs[0]);
		}

		Vector vector = new Vector<>(libs.length);

		for (int i = 0; i < libs.length; i++) {
			vector.add(new FileInputStream(libs[i]));

            if (i < libs.length - 1) {
                vector.add(new ByteArrayInputStream(IOUtil.LINE.getBytes(StandardCharsets.UTF_8)));
            }
		}

		return new SequenceInputStream(vector.elements());
	}

}