merge branch with master

pull/51/head
Miriam Baglioni 4 years ago
commit f8e9bda24c

@ -1,119 +1,117 @@
package eu.dnetlib.dhp.common;
import org.apache.commons.compress.archivers.tar.TarArchiveEntry;
import org.apache.commons.compress.archivers.tar.TarArchiveOutputStream;
import org.apache.hadoop.fs.*;
package eu.dnetlib.dhp.common;
import java.io.BufferedInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.Serializable;
public class MakeTarArchive implements Serializable {
private static TarArchiveOutputStream getTar(FileSystem fileSystem, String outputPath) throws IOException {
Path hdfsWritePath = new Path(outputPath);
FSDataOutputStream fsDataOutputStream = null;
if (fileSystem.exists(hdfsWritePath)) {
fileSystem.delete(hdfsWritePath, true);
}
fsDataOutputStream = fileSystem.create(hdfsWritePath);
import org.apache.commons.compress.archivers.tar.TarArchiveEntry;
import org.apache.commons.compress.archivers.tar.TarArchiveOutputStream;
import org.apache.hadoop.fs.*;
return new TarArchiveOutputStream(fsDataOutputStream.getWrappedStream());
}
public class MakeTarArchive implements Serializable {
private static void write(FileSystem fileSystem, String inputPath, String outputPath, String dir_name)
throws IOException {
private static TarArchiveOutputStream getTar(FileSystem fileSystem, String outputPath) throws IOException {
Path hdfsWritePath = new Path(outputPath);
FSDataOutputStream fsDataOutputStream = null;
if (fileSystem.exists(hdfsWritePath)) {
fileSystem.delete(hdfsWritePath, true);
Path hdfsWritePath = new Path(outputPath);
FSDataOutputStream fsDataOutputStream = null;
if (fileSystem.exists(hdfsWritePath)) {
fileSystem.delete(hdfsWritePath, true);
}
fsDataOutputStream = fileSystem.create(hdfsWritePath);
}
fsDataOutputStream = fileSystem.create(hdfsWritePath);
return new TarArchiveOutputStream(fsDataOutputStream.getWrappedStream());
}
TarArchiveOutputStream ar = new TarArchiveOutputStream(fsDataOutputStream.getWrappedStream());
private static void write(FileSystem fileSystem, String inputPath, String outputPath, String dir_name)
throws IOException {
RemoteIterator<LocatedFileStatus> fileStatusListIterator = fileSystem
.listFiles(
new Path(inputPath), true);
Path hdfsWritePath = new Path(outputPath);
FSDataOutputStream fsDataOutputStream = null;
if (fileSystem.exists(hdfsWritePath)) {
fileSystem.delete(hdfsWritePath, true);
while (fileStatusListIterator.hasNext()) {
writeCurrentFile(fileSystem, dir_name, fileStatusListIterator, ar, 0);
}
}
fsDataOutputStream = fileSystem.create(hdfsWritePath);
ar.close();
}
TarArchiveOutputStream ar = new TarArchiveOutputStream(fsDataOutputStream.getWrappedStream());
public static void tarMaxSize(FileSystem fileSystem, String inputPath, String outputPath, String dir_name,
int gBperSplit) throws IOException {
final long bytesPerSplit = 1024L * 1024L * 1024L * gBperSplit;
RemoteIterator<LocatedFileStatus> fileStatusListIterator = fileSystem
.listFiles(
new Path(inputPath), true);
long sourceSize = fileSystem.getContentSummary(new Path(inputPath)).getSpaceConsumed();
while (fileStatusListIterator.hasNext()) {
writeCurrentFile(fileSystem, dir_name, fileStatusListIterator, ar, 0);
}
if (sourceSize < bytesPerSplit) {
write(fileSystem, inputPath, outputPath + ".tar", dir_name);
} else {
int partNum = 0;
ar.close();
}
RemoteIterator<LocatedFileStatus> fileStatusListIterator = fileSystem
.listFiles(
new Path(inputPath), true);
boolean next = fileStatusListIterator.hasNext();
while (next) {
TarArchiveOutputStream ar = getTar(fileSystem, outputPath + "_" + (partNum + 1) + ".tar");
public static void tarMaxSize(FileSystem fileSystem, String inputPath, String outputPath, String dir_name,
int gBperSplit) throws IOException {
final long bytesPerSplit = 1024L * 1024L * 1024L * gBperSplit;
long current_size = 0;
while (next && current_size < bytesPerSplit) {
current_size = writeCurrentFile(fileSystem, dir_name, fileStatusListIterator, ar, current_size);
next = fileStatusListIterator.hasNext();
long sourceSize = fileSystem.getContentSummary(new Path(inputPath)).getSpaceConsumed();
}
if (sourceSize < bytesPerSplit) {
write(fileSystem, inputPath, outputPath + ".tar", dir_name);
} else {
int partNum = 0;
partNum += 1;
ar.close();
}
RemoteIterator<LocatedFileStatus> fileStatusListIterator = fileSystem
.listFiles(
new Path(inputPath), true);
boolean next = fileStatusListIterator.hasNext();
while (next) {
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);
next = fileStatusListIterator.hasNext();
}
}
private static long writeCurrentFile(FileSystem fileSystem, String dir_name,
RemoteIterator<LocatedFileStatus> fileStatusListIterator,
TarArchiveOutputStream ar, long current_size) throws IOException {
LocatedFileStatus fileStatus = fileStatusListIterator.next();
partNum += 1;
ar.close();
}
Path p = fileStatus.getPath();
String p_string = p.toString();
if (!p_string.endsWith("_SUCCESS")) {
String name = p_string.substring(p_string.lastIndexOf("/") + 1);
if (name.trim().equalsIgnoreCase("communities_infrastructures")) {
name = "communities_infrastructures.json";
}
TarArchiveEntry entry = new TarArchiveEntry(dir_name + "/" + name);
entry.setSize(fileStatus.getLen());
current_size += fileStatus.getLen();
ar.putArchiveEntry(entry);
}
InputStream is = fileSystem.open(fileStatus.getPath());
}
BufferedInputStream bis = new BufferedInputStream(is);
private static long writeCurrentFile(FileSystem fileSystem, String dir_name,
RemoteIterator<LocatedFileStatus> fileStatusListIterator,
TarArchiveOutputStream ar, long current_size) throws IOException {
LocatedFileStatus fileStatus = fileStatusListIterator.next();
int count;
byte data[] = new byte[1024];
while ((count = bis.read(data, 0, data.length)) != -1) {
ar.write(data, 0, count);
}
bis.close();
ar.closeArchiveEntry();
Path p = fileStatus.getPath();
String p_string = p.toString();
if (!p_string.endsWith("_SUCCESS")) {
String name = p_string.substring(p_string.lastIndexOf("/") + 1);
if (name.trim().equalsIgnoreCase("communities_infrastructures")) {
name = "communities_infrastructures.json";
}
TarArchiveEntry entry = new TarArchiveEntry(dir_name + "/" + name);
entry.setSize(fileStatus.getLen());
current_size += fileStatus.getLen();
ar.putArchiveEntry(entry);
}
return current_size;
}
InputStream is = fileSystem.open(fileStatus.getPath());
BufferedInputStream bis = new BufferedInputStream(is);
int count;
byte data[] = new byte[1024];
while ((count = bis.read(data, 0, data.length)) != -1) {
ar.write(data, 0, count);
}
bis.close();
ar.closeArchiveEntry();
}
return current_size;
}
}

@ -9,7 +9,6 @@ import org.junit.jupiter.api.BeforeAll;
import org.junit.jupiter.api.Disabled;
import org.junit.jupiter.api.Test;
@Disabled
public class HttpConnectorTest {

@ -109,20 +109,20 @@ public class CleaningFunctions {
}
if (Objects.nonNull(r.getPid())) {
r
.setPid(
r
.getPid()
.stream()
.filter(Objects::nonNull)
.filter(sp -> StringUtils.isNotBlank(StringUtils.trim(sp.getValue())))
.filter(sp -> NONE.equalsIgnoreCase(sp.getValue()))
.filter(sp -> Objects.nonNull(sp.getQualifier()))
.filter(sp -> StringUtils.isNotBlank(sp.getQualifier().getClassid()))
.map(sp -> {
sp.setValue(StringUtils.trim(sp.getValue()));
return sp;
})
.collect(Collectors.toList()));
.setPid(
r
.getPid()
.stream()
.filter(Objects::nonNull)
.filter(sp -> StringUtils.isNotBlank(StringUtils.trim(sp.getValue())))
.filter(sp -> NONE.equalsIgnoreCase(sp.getValue()))
.filter(sp -> Objects.nonNull(sp.getQualifier()))
.filter(sp -> StringUtils.isNotBlank(sp.getQualifier().getClassid()))
.map(sp -> {
sp.setValue(StringUtils.trim(sp.getValue()));
return sp;
})
.collect(Collectors.toList()));
}
if (Objects.isNull(r.getResourcetype()) || StringUtils.isBlank(r.getResourcetype().getClassid())) {
r

@ -1,18 +1,18 @@
<workflow-app name="dump_community_products" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>sourcePath</name>
<description>the source path</description>
</property>
<property>
<name>isLookUpUrl</name>
<description>the isLookup service endpoint</description>
</property>
<property>
<name>outputPath</name>
<description>the output path</description>
</property>
<property>
<name>sourcePath</name>
<description>the source path</description>
</property>
<property>
<name>isLookUpUrl</name>
<description>the isLookup service endpoint</description>
</property>
<property>
<name>outputPath</name>
<description>the output path</description>
</property>
<property>
<name>accessToken</name>
<description>the access token used for the deposition in Zenodo</description>
@ -320,6 +320,7 @@
<ok to="join_extend"/>
<error to="Kill"/>
</action>
<action name="extend_orp">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
@ -344,6 +345,7 @@
<ok to="join_extend"/>
<error to="Kill"/>
</action>
<action name="extend_software">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
@ -371,43 +373,42 @@
<join name="join_extend" to="splitForCommunities"/>
<action name="splitForCommunities">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Split dumped result for community</name>
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkSplitForCommunity</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
</spark-opts>
<arg>--sourcePath</arg><arg>${workingDir}/ext</arg>
<arg>--outputPath</arg><arg>${workingDir}/split</arg>
<arg>--communityMapPath</arg><arg>${workingDir}/communityMap</arg>
</spark>
<ok to="make_archive"/>
<error to="Kill"/>
</action>
<action name="splitForCommunities">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Split dumped result for community</name>
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkSplitForCommunity</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
</spark-opts>
<arg>--sourcePath</arg><arg>${workingDir}/ext</arg>
<arg>--outputPath</arg><arg>${workingDir}/split</arg>
<arg>--communityMapPath</arg><arg>${workingDir}/communityMap</arg>
</spark>
<ok to="make_archive"/>
<error to="Kill"/>
</action>
<action name="make_archive">
<java>
<main-class>eu.dnetlib.dhp.oa.graph.dump.MakeTar</main-class>
<arg>--hdfsPath</arg><arg>${outputPath}</arg>
<arg>--nameNode</arg><arg>${nameNode}</arg>
<arg>--sourcePath</arg><arg>${workingDir}/split</arg>
<arg>--sourcePath</arg><arg>${workingDir}/split</arg>
</java>
<ok to="send_zenodo"/>
<error to="Kill"/>
</action>
<action name="send_zenodo">
<java>
<main-class>eu.dnetlib.dhp.oa.graph.dump.SendToZenodoHDFS</main-class>
@ -424,8 +425,6 @@
<error to="Kill"/>
</action>
<end name="End"/>
</workflow-app>

@ -1,18 +1,18 @@
<workflow-app name="dump_whole_graph" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>sourcePath</name>
<description>the source path</description>
</property>
<property>
<name>isLookUpUrl</name>
<description>the isLookup service endpoint</description>
</property>
<property>
<name>outputPath</name>
<description>the output path</description>
</property>
<property>
<name>sourcePath</name>
<description>the source path</description>
</property>
<property>
<name>isLookUpUrl</name>
<description>the isLookup service endpoint</description>
</property>
<property>
<name>outputPath</name>
<description>the output path</description>
</property>
<property>
<name>resultAggregation</name>
<description>true if all the result type have to be dumped under result. false otherwise</description>
@ -357,10 +357,8 @@
<error to="Kill"/>
</action>
<join name="join_dump" to="fork_context"/>
<fork name="fork_context">
<path start="create_entities_fromcontext"/>
<path start="create_relation_fromcontext"/>
@ -389,7 +387,6 @@
<error to="Kill"/>
</action>
<action name="create_relation_fromorgs">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
@ -418,7 +415,6 @@
<join name="join_context" to="fork_extract_relations"/>
<fork name="fork_extract_relations">
<path start="rels_from_pubs"/>
<path start="rels_from_dats"/>
@ -530,7 +526,6 @@
<error to="Kill"/>
</action>
<join name="join_extract_relations" to="collect_and_save"/>
<action name="collect_and_save">
@ -568,8 +563,7 @@
<ok to="send_zenodo"/>
<error to="Kill"/>
</action>
<action name="send_zenodo">
<java>
<main-class>eu.dnetlib.dhp.oa.graph.dump.SendToZenodoHDFS</main-class>

@ -6,11 +6,36 @@ import eu.dnetlib.dhp.provision.scholix.summary.ScholixSummary
import eu.dnetlib.dhp.schema.oaf.Relation
import org.apache.commons.io.IOUtils
import org.apache.spark.SparkConf
import org.apache.spark.sql.expressions.Aggregator
import org.apache.spark.sql.{Dataset, Encoder, Encoders, SaveMode, SparkSession}
object SparkGenerateScholixIndex {
def getScholixAggregator(): Aggregator[(String, Scholix), Scholix, Scholix] = new Aggregator[(String, Scholix), Scholix, Scholix]{
override def zero: Scholix = new Scholix()
override def reduce(b: Scholix, a: (String, Scholix)): Scholix = {
b.mergeFrom(a._2)
b
}
override def merge(wx: Scholix, wy: Scholix): Scholix = {
wx.mergeFrom(wy)
wx
}
override def finish(reduction: Scholix): Scholix = reduction
override def bufferEncoder: Encoder[Scholix] =
Encoders.kryo(classOf[Scholix])
override def outputEncoder: Encoder[Scholix] =
Encoders.kryo(classOf[Scholix])
}
def main(args: Array[String]): Unit = {
val parser = new ArgumentApplicationParser(IOUtils.toString(SparkGenerateScholixIndex.getClass.getResourceAsStream("/eu/dnetlib/dhp/provision/input_generate_summary_parameters.json")))
parser.parseArgument(args)
@ -40,7 +65,7 @@ object SparkGenerateScholixIndex {
(relation.getTarget, Scholix.generateScholixWithSource(summary,relation))
}).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/scholix_source")
}).repartition(6000).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/scholix_source")
val sTarget:Dataset[(String,Scholix)] = spark.read.load(s"$workingDirPath/scholix_source").as[(String, Scholix)]
@ -53,9 +78,16 @@ object SparkGenerateScholixIndex {
scholix.generateIdentifier()
scholix.generatelinkPublisher()
scholix
}).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/scholix")
}).repartition(6000).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/scholix_r")
val finalScholix:Dataset[Scholix] = spark.read.load(s"$workingDirPath/scholix_r").as[Scholix]
finalScholix.map(d => (d.getIdentifier, d))(Encoders.tuple(Encoders.STRING, scholixEncoder))
.groupByKey(_._1)(Encoders.STRING)
.agg(getScholixAggregator().toColumn)
.map(p => p._2)
.write.mode(SaveMode.Overwrite).save(s"$workingDirPath/scholix")
}

@ -5,6 +5,8 @@ import java.io.Serializable;
import java.util.*;
import java.util.stream.Collectors;
import org.apache.commons.lang3.StringUtils;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.provision.scholix.summary.ScholixSummary;
@ -91,13 +93,91 @@ public class Scholix implements Serializable {
s.setSource(ScholixResource.fromSummary(scholixSummary));
s.setIdentifier(rel.getTarget());
// ScholixResource mockTarget = new ScholixResource();
// mockTarget.setDnetIdentifier(rel.getTarget());
// s.setTarget(mockTarget);
// s.generateIdentifier();
return s;
}
private List<ScholixEntityId> mergeScholixEntityId(final List<ScholixEntityId> a, final List<ScholixEntityId> b) {
final List<ScholixEntityId> m = new ArrayList<>(a);
if (b != null)
b.forEach(s -> {
int tt = (int) m.stream().filter(t -> t.getName().equalsIgnoreCase(s.getName())).count();
if (tt == 0) {
m.add(s);
}
});
return m;
}
private List<ScholixIdentifier> mergeScholixIdnetifier(final List<ScholixIdentifier> a,
final List<ScholixIdentifier> b) {
final List<ScholixIdentifier> m = new ArrayList<>(a);
if (b != null)
b.forEach(s -> {
int tt = (int) m.stream().filter(t -> t.getIdentifier().equalsIgnoreCase(s.getIdentifier())).count();
if (tt == 0) {
m.add(s);
}
});
return m;
}
private List<ScholixCollectedFrom> mergeScholixCollectedFrom(final List<ScholixCollectedFrom> a,
final List<ScholixCollectedFrom> b) {
final List<ScholixCollectedFrom> m = new ArrayList<>(a);
if (b != null)
b.forEach(s -> {
int tt = (int) m
.stream()
.filter(t -> t.getProvider().getName().equalsIgnoreCase(s.getProvider().getName()))
.count();
if (tt == 0) {
m.add(s);
}
});
return m;
}
private ScholixRelationship mergeRelationships(final ScholixRelationship a, final ScholixRelationship b) {
ScholixRelationship result = new ScholixRelationship();
result.setName(StringUtils.isEmpty(a.getName()) ? b.getName() : a.getName());
result.setInverse(StringUtils.isEmpty(a.getInverse()) ? b.getInverse() : a.getInverse());
result.setSchema(StringUtils.isEmpty(a.getSchema()) ? b.getSchema() : a.getSchema());
return result;
}
private ScholixResource mergeResource(final ScholixResource a, final ScholixResource b) {
final ScholixResource result = new ScholixResource();
result.setCollectedFrom(mergeScholixCollectedFrom(a.getCollectedFrom(), b.getCollectedFrom()));
result.setCreator(mergeScholixEntityId(a.getCreator(), b.getCreator()));
result
.setDnetIdentifier(
StringUtils.isBlank(a.getDnetIdentifier()) ? b.getDnetIdentifier() : a.getDnetIdentifier());
result.setIdentifier(mergeScholixIdnetifier(a.getIdentifier(), b.getIdentifier()));
result.setObjectType(StringUtils.isNotBlank(a.getObjectType()) ? a.getObjectType() : b.getObjectType());
result
.setObjectSubType(
StringUtils.isNotBlank(a.getObjectSubType()) ? a.getObjectSubType() : b.getObjectSubType());
result.setPublisher(mergeScholixEntityId(a.getPublisher(), b.getPublisher()));
result
.setPublicationDate(
StringUtils.isNotBlank(a.getPublicationDate()) ? a.getPublicationDate() : b.getPublicationDate());
result.setTitle(StringUtils.isNotBlank(a.getTitle()) ? a.getTitle() : b.getTitle());
return result;
}
public void mergeFrom(final Scholix other) {
linkprovider = mergeScholixEntityId(linkprovider, other.getLinkprovider());
publisher = mergeScholixEntityId(publisher, other.getPublisher());
if (StringUtils.isEmpty(publicationDate))
publicationDate = other.getPublicationDate();
relationship = mergeRelationships(relationship, other.getRelationship());
source = mergeResource(source, other.getSource());
target = mergeResource(target, other.getTarget());
generateIdentifier();
}
public void generatelinkPublisher() {
Set<String> publisher = new HashSet<>();
if (source.getPublisher() != null)

@ -108,7 +108,7 @@
<arg>-m</arg> <arg>yarn-cluster</arg>
<arg>--workingPath</arg><arg>${workingDirPath}</arg>
</spark>
<ok to="DropAndCreateIndex"/>
<ok to="End"/>
<error to="Kill"/>
</action>

Loading…
Cancel
Save