2020-08-05 14:17:40 +02:00
|
|
|
|
|
|
|
package eu.dnetlib.dhp.oa.graph.dump;
|
|
|
|
|
2020-08-07 17:49:14 +02:00
|
|
|
import java.io.Serializable;
|
|
|
|
import java.util.Optional;
|
2020-08-19 10:57:36 +02:00
|
|
|
|
2020-08-05 14:17:40 +02:00
|
|
|
import org.apache.commons.io.IOUtils;
|
|
|
|
import org.apache.commons.logging.Log;
|
|
|
|
import org.apache.commons.logging.LogFactory;
|
|
|
|
import org.apache.hadoop.conf.Configuration;
|
|
|
|
import org.apache.hadoop.fs.*;
|
2020-08-19 10:57:36 +02:00
|
|
|
|
2020-08-07 17:49:14 +02:00
|
|
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
2020-08-10 11:38:41 +02:00
|
|
|
import eu.dnetlib.dhp.common.api.MissingConceptDoiException;
|
|
|
|
import eu.dnetlib.dhp.common.api.ZenodoAPIClient;
|
2020-08-07 17:49:14 +02:00
|
|
|
import eu.dnetlib.dhp.oa.graph.dump.community.CommunityMap;
|
2020-08-05 14:17:40 +02:00
|
|
|
|
|
|
|
public class SendToZenodoHDFS implements Serializable {
|
|
|
|
|
2020-09-25 16:34:32 +02:00
|
|
|
private final static String NEW = "new"; // to be used for a brand new deposition in zenodo
|
|
|
|
private final static String VERSION = "version"; // to be used to upload a new version of a published deposition
|
|
|
|
private final static String UPDATE = "update"; // to upload content to an open deposition not published
|
|
|
|
|
2020-08-05 14:17:40 +02:00
|
|
|
private static final Log log = LogFactory.getLog(SendToZenodoHDFS.class);
|
|
|
|
|
2020-08-07 17:49:14 +02:00
|
|
|
public static void main(final String[] args) throws Exception, MissingConceptDoiException {
|
2020-08-05 14:17:40 +02:00
|
|
|
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
|
|
|
|
IOUtils
|
|
|
|
.toString(
|
|
|
|
SendToZenodoHDFS.class
|
|
|
|
.getResourceAsStream(
|
|
|
|
"/eu/dnetlib/dhp/oa/graph/dump/upload_zenodo.json")));
|
|
|
|
|
|
|
|
parser.parseArgument(args);
|
|
|
|
|
|
|
|
final String hdfsPath = parser.get("hdfsPath");
|
2020-08-11 15:27:31 +02:00
|
|
|
final String hdfsNameNode = parser.get("nameNode");
|
2020-08-05 14:17:40 +02:00
|
|
|
final String access_token = parser.get("accessToken");
|
|
|
|
final String connection_url = parser.get("connectionUrl");
|
|
|
|
final String metadata = parser.get("metadata");
|
2020-09-25 16:34:32 +02:00
|
|
|
final String depositionType = parser.get("depositionType");
|
2020-08-10 11:38:41 +02:00
|
|
|
final String concept_rec_id = Optional
|
|
|
|
.ofNullable(parser.get("conceptRecordId"))
|
|
|
|
.orElse(null);
|
2020-11-03 16:57:01 +01:00
|
|
|
final Boolean publish = Optional
|
|
|
|
.ofNullable(parser.get("publish"))
|
|
|
|
.map(Boolean::valueOf)
|
|
|
|
.orElse(false);
|
2020-09-25 16:34:32 +02:00
|
|
|
|
|
|
|
final String depositionId = Optional.ofNullable(parser.get("depositionId")).orElse(null);
|
2020-08-10 11:38:41 +02:00
|
|
|
final String communityMapPath = parser.get("communityMapPath");
|
2020-08-05 14:17:40 +02:00
|
|
|
|
|
|
|
Configuration conf = new Configuration();
|
|
|
|
conf.set("fs.defaultFS", hdfsNameNode);
|
|
|
|
|
|
|
|
FileSystem fileSystem = FileSystem.get(conf);
|
|
|
|
|
2020-08-11 15:27:31 +02:00
|
|
|
CommunityMap communityMap = Utils.readCommunityMap(fileSystem, communityMapPath);
|
2020-08-10 11:38:41 +02:00
|
|
|
|
2020-08-05 14:17:40 +02:00
|
|
|
RemoteIterator<LocatedFileStatus> fileStatusListIterator = fileSystem
|
|
|
|
.listFiles(
|
|
|
|
new Path(hdfsPath), true);
|
2020-08-07 17:49:14 +02:00
|
|
|
ZenodoAPIClient zenodoApiClient = new ZenodoAPIClient(connection_url, access_token);
|
2020-09-25 16:34:32 +02:00
|
|
|
switch (depositionType) {
|
|
|
|
case NEW:
|
|
|
|
zenodoApiClient.newDeposition();
|
|
|
|
break;
|
|
|
|
case VERSION:
|
|
|
|
if (concept_rec_id == null) {
|
|
|
|
throw new MissingConceptDoiException("No concept record id has been provided");
|
|
|
|
}
|
|
|
|
zenodoApiClient.newVersion(concept_rec_id);
|
|
|
|
break;
|
|
|
|
case UPDATE:
|
|
|
|
if (depositionId == null) {
|
|
|
|
throw new MissingConceptDoiException("No deposition id has been provided");
|
|
|
|
}
|
|
|
|
zenodoApiClient.uploadOpenDeposition(depositionId);
|
|
|
|
break;
|
2020-08-07 17:49:14 +02:00
|
|
|
}
|
|
|
|
|
2020-08-05 14:17:40 +02:00
|
|
|
while (fileStatusListIterator.hasNext()) {
|
|
|
|
LocatedFileStatus fileStatus = fileStatusListIterator.next();
|
|
|
|
|
|
|
|
Path p = fileStatus.getPath();
|
|
|
|
String p_string = p.toString();
|
|
|
|
if (!p_string.endsWith("_SUCCESS")) {
|
2020-08-11 15:27:31 +02:00
|
|
|
// String tmp = p_string.substring(0, p_string.lastIndexOf("/"));
|
|
|
|
String name = p_string.substring(p_string.lastIndexOf("/") + 1);
|
|
|
|
log.info("Sending information for community: " + name);
|
|
|
|
if (communityMap.containsKey(name.substring(0, name.lastIndexOf(".")))) {
|
|
|
|
name = communityMap.get(name.substring(0, name.lastIndexOf("."))).replace(" ", "_") + ".tar";
|
|
|
|
}
|
2020-08-05 14:17:40 +02:00
|
|
|
|
2020-08-07 17:49:14 +02:00
|
|
|
FSDataInputStream inputStream = fileSystem.open(p);
|
2020-08-11 15:27:31 +02:00
|
|
|
zenodoApiClient.uploadIS(inputStream, name, fileStatus.getLen());
|
2020-08-05 14:17:40 +02:00
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
2020-11-05 16:10:17 +01:00
|
|
|
if (!metadata.equals("")) {
|
|
|
|
zenodoApiClient.sendMretadata(metadata);
|
|
|
|
}
|
2020-08-05 14:17:40 +02:00
|
|
|
|
2020-11-03 16:57:01 +01:00
|
|
|
if (publish)
|
|
|
|
zenodoApiClient.publish();
|
2020-08-05 14:17:40 +02:00
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|