sabeel
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

[orcid-enrichment] change the value of parameters.

This commit is contained in:
Miriam Baglioni 2024-01-29 18:19:48 +01:00
parent f804c58bc7
commit a5995ab557
1 changed files with 14 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
dhp-workflows/dhp-enrichment/src/main/resources/eu/dnetlib/dhp/wf/subworkflows/orcidtoresultfromsemrel/oozie_app/workflow.xml
View File

@ -114,7 +114,7 @@
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--hive_metastore_uris</arg><arg>${hive_metastore_uris}</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
<arg>--outputPath</arg><arg>${workingDir}/orcid/preparedInfo/targetOrcidAssoc</arg>
<arg>--outputPath</arg><arg>${workingDir}/orcid/targetOrcidAssoc</arg>
<arg>--allowedsemrels</arg><arg>${allowedsemrels}</arg>
</spark>
<ok to="wait"/>
@ -142,7 +142,7 @@
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--hive_metastore_uris</arg><arg>${hive_metastore_uris}</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
<arg>--outputPath</arg><arg>${workingDir}/orcid/preparedInfo/targetOrcidAssoc</arg>
<arg>--outputPath</arg><arg>${workingDir}/orcid/targetOrcidAssoc</arg>
<arg>--allowedsemrels</arg><arg>${allowedsemrels}</arg>
</spark>
<ok to="wait"/>
@ -170,7 +170,7 @@
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--hive_metastore_uris</arg><arg>${hive_metastore_uris}</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
<arg>--outputPath</arg><arg>${workingDir}/orcid/preparedInfo/targetOrcidAssoc</arg>
<arg>--outputPath</arg><arg>${workingDir}/orcid/targetOrcidAssoc</arg>
<arg>--allowedsemrels</arg><arg>${allowedsemrels}</arg>
</spark>
<ok to="wait"/>
@ -198,7 +198,7 @@
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--hive_metastore_uris</arg><arg>${hive_metastore_uris}</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
<arg>--outputPath</arg><arg>${workingDir}/orcid/preparedInfo/targetOrcidAssoc</arg>
<arg>--outputPath</arg><arg>${workingDir}/orcid/targetOrcidAssoc</arg>
<arg>--allowedsemrels</arg><arg>${allowedsemrels}</arg>
</spark>
<ok to="wait"/>
@ -225,8 +225,8 @@
--conf spark.dynamicAllocation.enabled=true
--conf spark.dynamicAllocation.maxExecutors=${spark2MaxExecutors}
</spark-opts>
<arg>--sourcePath</arg><arg>${workingDir}/orcid/orcidprop</arg>
<arg>--outputPath</arg><arg>${workingDir}/orcid/orcidprop/mergedOrcidAssoc</arg>
<arg>--sourcePath</arg><arg>${workingDir}/orcid/targetOrcidAssoc</arg>
<arg>--outputPath</arg><arg>${workingDir}/orcid/mergedOrcidAssoc</arg>
</spark>
<ok to="fork-join-exec-propagation"/>
<error to="Kill"/>
@ -247,9 +247,10 @@
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.SparkOrcidToResultFromSemRelJob</class>
<jar>dhp-enrichment-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--executor-cores=4
--executor-memory=4G
--driver-memory=${sparkDriverMemory}
--conf spark.executor.memoryOverhead=5G
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
@ -259,9 +260,9 @@
--conf spark.speculation=false
--conf spark.hadoop.mapreduce.map.speculative=false
--conf spark.hadoop.mapreduce.reduce.speculative=false
--conf spark.sql.shuffle.partitions=3840
--conf spark.sql.shuffle.partitions=15000
</spark-opts>
<arg>--possibleUpdatesPath</arg><arg>${workingDir}/orcid/orcidprop/mergedOrcidAssoc</arg>
<arg>--possibleUpdatesPath</arg><arg>${workingDir}/orcid/mergedOrcidAssoc</arg>
<arg>--sourcePath</arg><arg>${sourcePath}/publication</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
<arg>--outputPath</arg><arg>${outputPath}/publication</arg>
@ -291,7 +292,7 @@
--conf spark.hadoop.mapreduce.map.speculative=false
--conf spark.hadoop.mapreduce.reduce.speculative=false
</spark-opts>
<arg>--possibleUpdatesPath</arg><arg>${workingDir}/orcid/orcidprop/mergedOrcidAssoc</arg>
<arg>--possibleUpdatesPath</arg><arg>${workingDir}/orcid/mergedOrcidAssoc</arg>
<arg>--sourcePath</arg><arg>${sourcePath}/dataset</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
<arg>--outputPath</arg><arg>${outputPath}/dataset</arg>
@ -321,7 +322,7 @@
--conf spark.hadoop.mapreduce.map.speculative=false
--conf spark.hadoop.mapreduce.reduce.speculative=false
</spark-opts>
<arg>--possibleUpdatesPath</arg><arg>${workingDir}/orcid/orcidprop/mergedOrcidAssoc</arg>
<arg>--possibleUpdatesPath</arg><arg>${workingDir}/orcid/mergedOrcidAssoc</arg>
<arg>--sourcePath</arg><arg>${sourcePath}/otherresearchproduct</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
<arg>--outputPath</arg><arg>${outputPath}/otherresearchproduct</arg>
@ -351,7 +352,7 @@
--conf spark.hadoop.mapreduce.map.speculative=false
--conf spark.hadoop.mapreduce.reduce.speculative=false
</spark-opts>
<arg>--possibleUpdatesPath</arg><arg>${workingDir}/orcid/orcidprop/mergedOrcidAssoc</arg>
<arg>--possibleUpdatesPath</arg><arg>${workingDir}/orcid/mergedOrcidAssoc</arg>
<arg>--sourcePath</arg><arg>${sourcePath}/software</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
<arg>--outputPath</arg><arg>${outputPath}/software</arg>