1
0
Fork 0

[scholix] added step to create tar archive

This commit is contained in:
Claudio Atzori 2022-08-11 11:25:24 +02:00
parent 2aa16d0432
commit cb7c07c54e
3 changed files with 118 additions and 22 deletions

View File

@ -5,13 +5,71 @@ import java.io.BufferedInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.Serializable;
import java.util.Optional;
import org.apache.commons.compress.archivers.tar.TarArchiveEntry;
import org.apache.commons.compress.archivers.tar.TarArchiveOutputStream;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.*;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
public class MakeTarArchive implements Serializable {
private static final Logger log = LoggerFactory.getLogger(MakeTarArchive.class);
public static void main(String[] args) throws Exception {
String jsonConfiguration = IOUtils
.toString(
MakeTarArchive.class
.getResourceAsStream(
"/eu/dnetlib/dhp/common/input_maketar_parameters.json"));
final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
parser.parseArgument(args);
final String outputPath = parser.get("hdfsPath");
log.info("hdfsPath: {}", outputPath);
final String hdfsNameNode = parser.get("nameNode");
log.info("nameNode: {}", hdfsNameNode);
final String inputPath = parser.get("sourcePath");
log.info("input path : {}", inputPath);
final int gBperSplit = Optional
.ofNullable(parser.get("splitSize"))
.map(Integer::valueOf)
.orElse(10);
Configuration conf = new Configuration();
conf.set("fs.defaultFS", hdfsNameNode);
FileSystem fileSystem = FileSystem.get(conf);
makeTArArchive(fileSystem, inputPath, outputPath, gBperSplit);
}
public static void makeTArArchive(FileSystem fileSystem, String inputPath, String outputPath, int gBperSplit)
throws IOException {
RemoteIterator<LocatedFileStatus> dirIterator = fileSystem.listLocatedStatus(new Path(inputPath));
while (dirIterator.hasNext()) {
LocatedFileStatus fileStatus = dirIterator.next();
Path p = fileStatus.getPath();
String pathString = p.toString();
String entity = pathString.substring(pathString.lastIndexOf("/") + 1);
MakeTarArchive.tarMaxSize(fileSystem, pathString, outputPath + "/" + entity, entity, gBperSplit);
}
}
private static TarArchiveOutputStream getTar(FileSystem fileSystem, String outputPath) throws IOException {
Path hdfsWritePath = new Path(outputPath);
if (fileSystem.exists(hdfsWritePath)) {
@ -21,7 +79,7 @@ public class MakeTarArchive implements Serializable {
return new TarArchiveOutputStream(fileSystem.create(hdfsWritePath).getWrappedStream());
}
private static void write(FileSystem fileSystem, String inputPath, String outputPath, String dir_name)
private static void write(FileSystem fileSystem, String inputPath, String outputPath, String dirName)
throws IOException {
Path hdfsWritePath = new Path(outputPath);
@ -37,7 +95,7 @@ public class MakeTarArchive implements Serializable {
new Path(inputPath), true);
while (iterator.hasNext()) {
writeCurrentFile(fileSystem, dir_name, iterator, ar, 0);
writeCurrentFile(fileSystem, dirName, iterator, ar, 0);
}
}
@ -59,32 +117,30 @@ public class MakeTarArchive implements Serializable {
new Path(inputPath), true);
boolean next = fileStatusListIterator.hasNext();
while (next) {
TarArchiveOutputStream ar = getTar(fileSystem, outputPath + "_" + (partNum + 1) + ".tar");
try (TarArchiveOutputStream ar = getTar(fileSystem, outputPath + "_" + (partNum + 1) + ".tar")) {
long current_size = 0;
while (next && current_size < bytesPerSplit) {
current_size = writeCurrentFile(fileSystem, dir_name, fileStatusListIterator, ar, current_size);
long currentSize = 0;
while (next && currentSize < bytesPerSplit) {
currentSize = writeCurrentFile(fileSystem, dir_name, fileStatusListIterator, ar, currentSize);
next = fileStatusListIterator.hasNext();
}
partNum += 1;
ar.close();
}
}
}
}
}
}
private static long writeCurrentFile(FileSystem fileSystem, String dir_name,
private static long writeCurrentFile(FileSystem fileSystem, String dirName,
RemoteIterator<LocatedFileStatus> fileStatusListIterator,
TarArchiveOutputStream ar, long current_size) throws IOException {
TarArchiveOutputStream ar, long currentSize) throws IOException {
LocatedFileStatus fileStatus = fileStatusListIterator.next();
Path p = fileStatus.getPath();
String p_string = p.toString();
if (!p_string.endsWith("_SUCCESS")) {
String name = p_string.substring(p_string.lastIndexOf("/") + 1);
String pString = p.toString();
if (!pString.endsWith("_SUCCESS")) {
String name = pString.substring(pString.lastIndexOf("/") + 1);
if (name.startsWith("part-") & name.length() > 10) {
String tmp = name.substring(0, 10);
if (name.contains(".")) {
@ -92,9 +148,9 @@ public class MakeTarArchive implements Serializable {
}
name = tmp;
}
TarArchiveEntry entry = new TarArchiveEntry(dir_name + "/" + name);
TarArchiveEntry entry = new TarArchiveEntry(dirName + "/" + name);
entry.setSize(fileStatus.getLen());
current_size += fileStatus.getLen();
currentSize += fileStatus.getLen();
ar.putArchiveEntry(entry);
InputStream is = fileSystem.open(fileStatus.getPath());
@ -110,7 +166,7 @@ public class MakeTarArchive implements Serializable {
ar.closeArchiveEntry();
}
return current_size;
return currentSize;
}
}

View File

@ -0,0 +1,30 @@
[
{
"paramName":"s",
"paramLongName":"sourcePath",
"paramDescription": "the path of the sequencial file to read",
"paramRequired": true
},
{
"paramName": "hdp",
"paramLongName": "hdfsPath",
"paramDescription": "the path used to store the output archive",
"paramRequired": true
},
{
"paramName":"nn",
"paramLongName":"nameNode",
"paramDescription": "the name node",
"paramRequired": true
},
{
"paramName":"ss",
"paramLongName":"splitSize",
"paramDescription": "the maximum size of the archive",
"paramRequired": false
}
]

View File

@ -140,11 +140,21 @@
<arg>--objectType</arg><arg>scholix</arg>
<arg>--maxPidNumberFilter</arg><arg>maxNumberOfPid</arg>
</spark>
<ok to="make_tar"/>
<error to="Kill"/>
</action>
<action name="make_tar">
<java>
<main-class>eu.dnetlib.dhp.common.MakeTarArchive</main-class>
<arg>--nameNode</arg><arg>${nameNode}</arg>
<arg>--hdfsPath</arg><arg>${targetPath}/tar</arg>
<arg>--sourcePath</arg><arg>${targetPath}/json</arg>
</java>
<ok to="End"/>
<error to="Kill"/>
</action>
<end name="End"/>
</workflow-app>