You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
yacy_search_server/source/de/anomic/server/serverFileUtils.java

513 lines
20 KiB

// serverFileUtils.java
// -------------------------------------------
// (C) by Michael Peter Christen; mc@anomic.de
// first published on http://www.anomic.de
// Frankfurt, Germany, 2004
// last major change: 05.08.2004
//
// This program is free software; you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation; either version 2 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
//
// Using this software in any meaning (reading, learning, copying, compiling,
// running) means that you agree that the Author(s) is (are) not responsible
// for cost, loss of data or any harm that may be caused directly or indirectly
// by usage of this softare or this documentation. The usage of this software
// is on your own risk. The installation and usage (starting/running) of this
// software may allow other people or application to access your computer and
// any attached devices and is highly dependent on the configuration of the
// software which must be done by the user of the software; the author(s) is
// (are) also not responsible for proper configuration and usage of the
// software, even if provoked by documentation provided together with
// the software.
//
// Any changes to this file according to the GPL as documented in the file
// gpl.txt aside this file in the shipment you received can be done to the
// lines that follows this copyright notice here, but changes must not be
// done inside the copyright notive above. A re-distribution must contain
// the intact and unchanged copyright notice.
// Contributions and changes to the program code must be marked as such.
package de.anomic.server;
import java.io.BufferedOutputStream;
import java.io.BufferedReader;
import java.io.ByteArrayInputStream;
import java.io.ByteArrayOutputStream;
import java.io.File;
import java.io.FileInputStream;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.io.OutputStream;
import java.io.PrintWriter;
import java.io.Reader;
import java.io.Writer;
import java.util.Comparator;
import java.util.HashSet;
import java.util.Iterator;
import java.util.Map;
import java.util.Set;
import java.util.StringTokenizer;
import java.util.TreeSet;
import java.util.zip.GZIPInputStream;
import java.util.zip.GZIPOutputStream;
import java.util.zip.ZipEntry;
import java.util.zip.ZipOutputStream;
import de.anomic.kelondro.kelondroRow;
import de.anomic.kelondro.kelondroRowSet;
import de.anomic.tools.nxTools;
public final class serverFileUtils {
private static final int DEFAULT_BUFFER_SIZE = 512;
public static long copy(InputStream source, OutputStream dest) throws IOException {
return copy(source,dest, -1);
}
/**
* Copies an InputStream to an OutputStream.
* @param source InputStream
* @param dest OutputStream
* @param count the total amount of bytes to copy
* @return Total number of bytes copied.
*
* @see #copy(InputStream source, File dest)
* @see #copyRange(File source, OutputStream dest, int start)
* @see #copy(File source, OutputStream dest)
* @see #copy(File source, File dest)
*/
public static long copy(InputStream source, OutputStream dest, long count) throws IOException {
byte[] buffer = new byte[DEFAULT_BUFFER_SIZE];
int chunkSize = (int) ((count > 0) ? Math.min(count, DEFAULT_BUFFER_SIZE) : DEFAULT_BUFFER_SIZE);
int c; long total = 0;
while ((c = source.read(buffer, 0, chunkSize)) > 0) {
dest.write(buffer, 0, c);
dest.flush();
total += c;
if (count > 0) {
chunkSize = (int) Math.min(count-total, DEFAULT_BUFFER_SIZE);
if (chunkSize == 0) break;
}
}
dest.flush();
return total;
}
public static int copy(File source, String inputCharset, Writer dest) throws IOException {
InputStream fis = null;
try {
fis = new FileInputStream(source);
return copy(fis, dest, inputCharset);
} finally {
if (fis != null) try { fis.close(); } catch (Exception e) {}
}
}
public static int copy(InputStream source, Writer dest, String inputCharset) throws IOException {
InputStreamReader reader = new InputStreamReader(source,inputCharset);
return copy(reader,dest);
}
public static int copy(String source, Writer dest) throws IOException {
dest.write(source);
dest.flush();
return source.length();
}
public static int copy(Reader source, Writer dest) throws IOException {
char[] buffer = new char[4096];
int count = 0;
int n = 0;
try {
while (-1 != (n = source.read(buffer))) {
dest.write(buffer, 0, n);
count += n;
}
} catch (Exception e) {
// an "sun.io.MalformedInputException: Missing byte-order mark" - exception may occur here
throw new IOException(e.getMessage());
}
return count;
}
public static void copy(InputStream source, File dest) throws IOException {
copy(source,dest,-1);
}
/**
* Copies an InputStream to a File.
* @param source InputStream
* @param dest File
* @param the amount of bytes to copy
* @see #copy(InputStream source, OutputStream dest)
* @see #copyRange(File source, OutputStream dest, int start)
* @see #copy(File source, OutputStream dest)
* @see #copy(File source, File dest)
*/
public static void copy(InputStream source, File dest, long count) throws IOException {
FileOutputStream fos = null;
try {
fos = new FileOutputStream(dest);
copy(source, fos, count);
} finally {
if (fos != null) try {fos.close();} catch (Exception e) {}
}
}
/**
* Copies a part of a File to an OutputStream.
* @param source File
* @param dest OutputStream
* @param start Number of bytes to skip from the beginning of the File
* @see #copy(InputStream source, OutputStream dest)
* @see #copy(InputStream source, File dest)
* @see #copy(File source, OutputStream dest)
* @see #copy(File source, File dest)
*/
public static void copyRange(File source, OutputStream dest, int start) throws IOException {
InputStream fis = null;
try {
fis = new FileInputStream(source);
long skipped = fis.skip(start);
if (skipped != start) throw new IllegalStateException("Unable to skip '" + start + "' bytes. Only '" + skipped + "' bytes skipped.");
copy(fis, dest, -1);
} finally {
if (fis != null) try { fis.close(); } catch (Exception e) {}
}
}
/**
* Copies a File to an OutputStream.
* @param source File
* @param dest OutputStream
* @see #copy(InputStream source, OutputStream dest)
* @see #copy(InputStream source, File dest)
* @see #copyRange(File source, OutputStream dest, int start)
* @see #copy(File source, File dest)
*/
public static void copy(File source, OutputStream dest) throws IOException {
InputStream fis = null;
try {
fis = new FileInputStream(source);
copy(fis, dest, -1);
} finally {
if (fis != null) try { fis.close(); } catch (Exception e) {}
}
}
public static void copy(File source, File dest) throws IOException {
copy(source,dest,-1);
}
/**
* Copies a File to a File.
* @param source File
* @param dest File
* @param count the amount of bytes to copy
* @see #copy(InputStream source, OutputStream dest)
* @see #copy(InputStream source, File dest)
* @see #copyRange(File source, OutputStream dest, int start)
* @see #copy(File source, OutputStream dest)
*/
public static void copy(File source, File dest, long count) throws IOException {
FileInputStream fis = null;
FileOutputStream fos = null;
try {
fis = new FileInputStream(source);
fos = new FileOutputStream(dest);
copy(fis, fos, count);
} finally {
if (fis != null) try {fis.close();} catch (Exception e) {}
if (fos != null) try {fos.close();} catch (Exception e) {}
}
}
public static byte[] read(InputStream source) throws IOException {
return read(source,-1);
}
public static byte[] read(InputStream source, long count) throws IOException {
ByteArrayOutputStream baos = (count > 0)
? new ByteArrayOutputStream((int)count)
: new ByteArrayOutputStream();
copy(source, baos, count);
baos.close();
// convert Stream into array
return baos.toByteArray();
}
public static byte[] read(File source) throws IOException {
byte[] buffer = new byte[(int) source.length()];
InputStream fis = null;
try {
fis = new FileInputStream(source);
int p = 0, c;
while ((c = fis.read(buffer, p, buffer.length - p)) > 0) p += c;
} finally {
if (fis != null) try { fis.close(); } catch (Exception e) {}
}
return buffer;
}
public static byte[] readAndZip(File source) throws IOException {
ByteArrayOutputStream byteOut = null;
GZIPOutputStream zipOut = null;
try {
byteOut = new ByteArrayOutputStream((int)(source.length()/2));
zipOut = new GZIPOutputStream(byteOut);
copy(source, zipOut);
zipOut.close();
return byteOut.toByteArray();
} finally {
if (zipOut != null) try { zipOut.close(); } catch (Exception e) {}
if (byteOut != null) try { byteOut.close(); } catch (Exception e) {}
}
}
public static void writeAndGZip(byte[] source, File dest) throws IOException {
FileOutputStream fos = null;
try {
fos = new FileOutputStream(dest);
writeAndGZip(source, fos);
} finally {
if (fos != null) try {fos.close();} catch (Exception e) {}
}
}
public static void writeAndGZip(byte[] source, OutputStream dest) throws IOException {
GZIPOutputStream zipOut = null;
try {
zipOut = new GZIPOutputStream(dest);
write(source, zipOut);
zipOut.close();
} finally {
if (zipOut != null) try { zipOut.close(); } catch (Exception e) {}
}
}
public static void write(String source, Writer dest) throws IOException {
dest.write(source);
dest.flush();
}
public static void write(byte[] source, OutputStream dest) throws IOException {
dest.write(source, 0, source.length);
dest.flush();
}
public static void write(byte[] source, File dest) throws IOException {
copy(new ByteArrayInputStream(source), dest);
}
/**
* This function determines if a byte array is gzip compressed and uncompress it
* @param source properly gzip compressed byte array
* @return uncompressed byte array
* @throws IOException
*/
public static byte[] uncompressGZipArray(byte[] source) throws IOException {
if (source == null) return null;
// support of gzipped data (requested by roland)
if ((source.length > 1) && (((source[1] << 8) | source[0]) == GZIPInputStream.GZIP_MAGIC)) {
System.out.println("DEBUG: uncompressGZipArray - uncompressing source");
try {
ByteArrayInputStream byteInput = new ByteArrayInputStream(source);
ByteArrayOutputStream byteOutput = new ByteArrayOutputStream();
GZIPInputStream zippedContent = new GZIPInputStream(byteInput);
byte[] data = new byte[1024];
int read = 0;
// reading gzip file and store it uncompressed
while((read = zippedContent.read(data, 0, 1024)) != -1) {
byteOutput.write(data, 0, read);
}
zippedContent.close();
byteOutput.close();
source = byteOutput.toByteArray();
} catch (Exception e) {
if (!e.getMessage().equals("Not in GZIP format")) {
throw new IOException(e.getMessage());
}
}
}
return source;
}
public static HashSet<String> loadList(File file) {
HashSet<String> set = new HashSet<String>();
BufferedReader br = null;
try {
br = new BufferedReader(new InputStreamReader(new FileInputStream(file)));
String line;
while ((line = br.readLine()) != null) {
line = line.trim();
if ((line.length() > 0) && (!(line.startsWith("#")))) set.add(line.trim().toLowerCase());
}
br.close();
} catch (IOException e) {
} finally {
if (br != null) try { br.close(); } catch (Exception e) {}
}
return set;
}
public static Map<String, String> loadHashMap(File f) {
// load props
try {
byte[] b = read(f);
return nxTools.table(nxTools.strings(b));
} catch (IOException e2) {
System.err.println("ERROR: " + f.toString() + " not found in settings path");
return null;
}
}
public static void saveMap(File file, Map<String, String> props, String comment) throws IOException {
PrintWriter pw = null;
File tf = new File(file.toString() + "." + (System.currentTimeMillis() % 1000));
pw = new PrintWriter(new BufferedOutputStream(new FileOutputStream(tf)));
pw.println("# " + comment);
Iterator<Map.Entry<String, String>> i = props.entrySet().iterator();
String key, value;
Map.Entry<String, String> entry;
while (i.hasNext()) {
entry = i.next();
key = entry.getKey();
value = entry.getValue().replaceAll("\n", "\\\\n");
pw.println(key + "=" + value);
}
pw.println("# EOF");
pw.close();
file.delete();
tf.renameTo(file);
}
public static Set<String> loadSet(File file, int chunksize, boolean tree) throws IOException {
Set<String> set = (tree) ? (Set<String>) new TreeSet<String>() : (Set<String>) new HashSet<String>();
byte[] b = read(file);
for (int i = 0; (i + chunksize) <= b.length; i++) {
set.add(new String(b, i, chunksize));
}
return set;
}
public static Set<String> loadSet(File file, String sep, boolean tree) throws IOException {
Set<String> set = (tree) ? (Set<String>) new TreeSet<String>() : (Set<String>) new HashSet<String>();
byte[] b = read(file);
StringTokenizer st = new StringTokenizer(new String(b, "UTF-8"), sep);
while (st.hasMoreTokens()) {
set.add(st.nextToken());
}
return set;
}
public static void saveSet(File file, String format, Set<String> set, String sep) throws IOException {
File tf = new File(file.toString() + ".tmp" + (System.currentTimeMillis() % 1000));
OutputStream os = null;
if ((format == null) || (format.equals("plain"))) {
os = new BufferedOutputStream(new FileOutputStream(tf));
} else if (format.equals("gzip")) {
os = new GZIPOutputStream(new FileOutputStream(tf));
} else if (format.equals("zip")) {
ZipOutputStream zos = new ZipOutputStream(new FileOutputStream(file));
String name = file.getName();
if (name.endsWith(".zip")) name = name.substring(0, name.length() - 4);
zos.putNextEntry(new ZipEntry(name + ".txt"));
os = zos;
}
for (Iterator<String> i = set.iterator(); i.hasNext(); ) {
os.write((i.next().toString()).getBytes());
if (sep != null) os.write(sep.getBytes());
}
os.close();
file.delete();
tf.renameTo(file);
}
public static void saveSet(File file, String format, kelondroRowSet set, String sep) throws IOException {
File tf = new File(file.toString() + ".tmp" + (System.currentTimeMillis() % 1000));
OutputStream os = null;
if ((format == null) || (format.equals("plain"))) {
os = new BufferedOutputStream(new FileOutputStream(tf));
} else if (format.equals("gzip")) {
os = new GZIPOutputStream(new FileOutputStream(tf));
} else if (format.equals("zip")) {
ZipOutputStream zos = new ZipOutputStream(new FileOutputStream(file));
String name = file.getName();
if (name.endsWith(".zip")) name = name.substring(0, name.length() - 4);
zos.putNextEntry(new ZipEntry(name + ".txt"));
os = zos;
}
Iterator<kelondroRow.Entry> i = set.rows();
String key;
if (i.hasNext()) {
key = new String(i.next().getColBytes(0));
os.write(key.getBytes());
}
while (i.hasNext()) {
key = new String(((kelondroRow.Entry) i.next()).getColBytes(0));
if (sep != null) os.write(sep.getBytes());
os.write(key.getBytes());
}
os.close();
file.delete();
tf.renameTo(file);
}
/**
* Moves all files from a directory to another.
* @param from_dir Directory which contents will be moved.
* @param to_dir Directory to move into. It must exist already.
*/
public static void moveAll(File from_dir, File to_dir) {
if (!(from_dir.isDirectory())) return;
if (!(to_dir.isDirectory())) return;
String[] list = from_dir.list();
for (int i = 0; i < list.length; i++) (new File(from_dir, list[i])).renameTo(new File(to_dir, list[i]));
}
public static class dirlistComparator implements Comparator<File> {
public int compare(File file1, File file2) {
if (file1.isDirectory() && !file2.isDirectory()) {
return -1;
} else if (!file1.isDirectory() && file2.isDirectory()) {
return 1;
} else {
return file1.getName().compareToIgnoreCase(file2.getName());
}
}
}
public static void main(String[] args) {
try {
writeAndGZip("ein zwei drei, Zauberei".getBytes(), new File("zauberei.txt.gz"));
} catch (IOException e) {
e.printStackTrace();
}
}
}