D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 12 Projects 2 Releases 2 Wiki Activity

[dedup] tweaking num partitions, minor changes

This commit is contained in:
Claudio Atzori 2023-05-17 10:16:22 +02:00
parent e309688711
commit 909729a2fc
3 changed files with 14 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/SparkCreateMergeRels.java
View File

@ -59,7 +59,7 @@ public class SparkCreateMergeRels extends AbstractSparkAction {
ArgumentApplicationParser parser = new ArgumentApplicationParser( ArgumentApplicationParser parser = new ArgumentApplicationParser(
IOUtils IOUtils
.toString( .toString(
SparkCreateSimRels.class SparkCreateMergeRels.class
.getResourceAsStream( .getResourceAsStream(
"/eu/dnetlib/dhp/oa/dedup/createCC_parameters.json"))); "/eu/dnetlib/dhp/oa/dedup/createCC_parameters.json")));
parser.parseArgument(args); parser.parseArgument(args);
@ -99,7 +99,7 @@ public class SparkCreateMergeRels extends AbstractSparkAction {
final String subEntity = dedupConf.getWf().getSubEntityValue(); final String subEntity = dedupConf.getWf().getSubEntityValue();
final Class<OafEntity> clazz = ModelSupport.entityTypes.get(EntityType.valueOf(subEntity)); final Class<OafEntity> clazz = ModelSupport.entityTypes.get(EntityType.valueOf(subEntity));
log.info("Creating mergerels for: '{}'", subEntity); log.info("Creating merge rels for: '{}'", subEntity);
final int maxIterations = dedupConf.getWf().getMaxIterations(); final int maxIterations = dedupConf.getWf().getMaxIterations();
log.info("Max iterations {}", maxIterations); log.info("Max iterations {}", maxIterations);

4
dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/SparkWhitelistSimRels.java
View File

@ -31,7 +31,7 @@ import scala.Tuple2;
public class SparkWhitelistSimRels extends AbstractSparkAction { public class SparkWhitelistSimRels extends AbstractSparkAction {
private static final Logger log = LoggerFactory.getLogger(SparkCreateSimRels.class); private static final Logger log = LoggerFactory.getLogger(SparkWhitelistSimRels.class);
private static final String WHITELIST_SEPARATOR = "####"; private static final String WHITELIST_SEPARATOR = "####";
@ -43,7 +43,7 @@ public class SparkWhitelistSimRels extends AbstractSparkAction {
ArgumentApplicationParser parser = new ArgumentApplicationParser( ArgumentApplicationParser parser = new ArgumentApplicationParser(
IOUtils IOUtils
.toString( .toString(
SparkCreateSimRels.class SparkWhitelistSimRels.class
.getResourceAsStream( .getResourceAsStream(
"/eu/dnetlib/dhp/oa/dedup/whitelistSimRels_parameters.json"))); "/eu/dnetlib/dhp/oa/dedup/whitelistSimRels_parameters.json")));
parser.parseArgument(args); parser.parseArgument(args);

20
dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml
View File

@ -126,13 +126,13 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840 --conf spark.sql.shuffle.partitions=15000
</spark-opts> </spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg> <arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg> <arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
<arg>--actionSetId</arg><arg>${actionSetId}</arg> <arg>--actionSetId</arg><arg>${actionSetId}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg> <arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--numPartitions</arg><arg>8000</arg> <arg>--numPartitions</arg><arg>15000</arg>
</spark> </spark>
<ok to="WhitelistSimRels"/> <ok to="WhitelistSimRels"/>
<error to="Kill"/> <error to="Kill"/>
@ -153,14 +153,14 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840 --conf spark.sql.shuffle.partitions=15000
</spark-opts> </spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg> <arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg> <arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
<arg>--actionSetId</arg><arg>${actionSetId}</arg> <arg>--actionSetId</arg><arg>${actionSetId}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg> <arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--whiteListPath</arg><arg>${whiteListPath}</arg> <arg>--whiteListPath</arg><arg>${whiteListPath}</arg>
<arg>--numPartitions</arg><arg>8000</arg> <arg>--numPartitions</arg><arg>15000</arg>
</spark> </spark>
<ok to="CreateMergeRel"/> <ok to="CreateMergeRel"/>
<error to="Kill"/> <error to="Kill"/>
@ -181,7 +181,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840 --conf spark.sql.shuffle.partitions=15000
</spark-opts> </spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg> <arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg> <arg>--workingPath</arg><arg>${workingPath}</arg>
@ -208,7 +208,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840 --conf spark.sql.shuffle.partitions=15000
</spark-opts> </spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg> <arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg> <arg>--workingPath</arg><arg>${workingPath}</arg>
@ -235,13 +235,13 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840 --conf spark.sql.shuffle.partitions=15000
</spark-opts> </spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg> <arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg> <arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg> <arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
<arg>--actionSetId</arg><arg>${actionSetIdOpenorgs}</arg> <arg>--actionSetId</arg><arg>${actionSetIdOpenorgs}</arg>
<arg>--numPartitions</arg><arg>8000</arg> <arg>--numPartitions</arg><arg>15000</arg>
</spark> </spark>
<ok to="CreateOrgsDedupRecord"/> <ok to="CreateOrgsDedupRecord"/>
<error to="Kill"/> <error to="Kill"/>
@ -288,7 +288,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840 --conf spark.sql.shuffle.partitions=10000
</spark-opts> </spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg> <arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg> <arg>--workingPath</arg><arg>${workingPath}</arg>
@ -314,7 +314,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840 --conf spark.sql.shuffle.partitions=10000
</spark-opts> </spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg> <arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg> <arg>--workingPath</arg><arg>${workingPath}</arg>