2021-07-23 16:38:32 +02:00
|
|
|
<workflow-app name="Create Scholix final Graph" xmlns="uri:oozie:workflow:0.5">
|
2020-02-19 10:07:08 +01:00
|
|
|
<parameters>
|
2020-03-27 17:07:34 +01:00
|
|
|
<property>
|
2021-06-30 17:27:55 +02:00
|
|
|
<name>sourcePath</name>
|
2020-02-19 10:07:08 +01:00
|
|
|
<description>the working dir base path</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
2021-06-30 17:27:55 +02:00
|
|
|
<name>targetPath</name>
|
2021-07-23 16:38:32 +02:00
|
|
|
<description>the final graph path</description>
|
2020-02-19 10:07:08 +01:00
|
|
|
</property>
|
|
|
|
</parameters>
|
|
|
|
|
2021-07-23 16:38:32 +02:00
|
|
|
<start to="ImportDatasetEntities"/>
|
2020-02-19 10:07:08 +01:00
|
|
|
|
|
|
|
<kill name="Kill">
|
|
|
|
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
|
|
|
|
</kill>
|
|
|
|
|
2021-07-23 16:38:32 +02:00
|
|
|
<action name="ImportDatasetEntities">
|
2021-06-30 17:27:55 +02:00
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
2021-07-23 16:38:32 +02:00
|
|
|
<name>Import JSONRDD to Dataset kryo</name>
|
|
|
|
<class>eu.dnetlib.dhp.sx.graph.SparkConvertRDDtoDataset</class>
|
2020-03-27 17:07:34 +01:00
|
|
|
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
2020-07-30 09:25:56 +02:00
|
|
|
<spark-opts>
|
2021-06-30 17:27:55 +02:00
|
|
|
--executor-memory=${sparkExecutorMemory}
|
2020-07-30 09:25:56 +02:00
|
|
|
--executor-cores=${sparkExecutorCores}
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
2021-06-30 17:27:55 +02:00
|
|
|
--conf spark.extraListeners=${spark2ExtraListeners}
|
|
|
|
--conf spark.sql.shuffle.partitions=3000
|
|
|
|
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
|
|
|
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
|
|
|
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
2020-07-30 09:25:56 +02:00
|
|
|
</spark-opts>
|
2021-06-30 17:27:55 +02:00
|
|
|
<arg>--master</arg><arg>yarn</arg>
|
2021-07-23 16:38:32 +02:00
|
|
|
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
|
2021-07-25 11:15:37 +02:00
|
|
|
<arg>--targetPath</arg><arg>${targetPath}</arg>
|
2020-03-27 17:07:34 +01:00
|
|
|
</spark>
|
2021-07-01 22:13:45 +02:00
|
|
|
<ok to="CreateSummaries"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
|
|
|
|
<action name="CreateSummaries">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Convert Entities to summaries</name>
|
|
|
|
<class>eu.dnetlib.dhp.sx.graph.SparkCreateSummaryObject</class>
|
|
|
|
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
|
|
|
<spark-opts>
|
|
|
|
--executor-memory=${sparkExecutorMemory}
|
|
|
|
--executor-cores=${sparkExecutorCores}
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
|
|
|
--conf spark.extraListeners=${spark2ExtraListeners}
|
2021-11-02 15:58:15 +01:00
|
|
|
--conf spark.sql.shuffle.partitions=20000
|
2021-07-01 22:13:45 +02:00
|
|
|
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
|
|
|
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
|
|
|
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
|
|
|
</spark-opts>
|
|
|
|
<arg>--master</arg><arg>yarn</arg>
|
2021-07-23 16:38:32 +02:00
|
|
|
<arg>--sourcePath</arg><arg>${targetPath}/entities</arg>
|
2021-07-01 22:13:45 +02:00
|
|
|
<arg>--targetPath</arg><arg>${targetPath}/provision/summaries</arg>
|
|
|
|
</spark>
|
2021-07-06 17:18:06 +02:00
|
|
|
<ok to="CreateScholix"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<action name="CreateScholix">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Generate Scholix Dataset</name>
|
|
|
|
<class>eu.dnetlib.dhp.sx.graph.SparkCreateScholix</class>
|
|
|
|
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
|
|
|
<spark-opts>
|
|
|
|
--executor-memory=${sparkExecutorMemory}
|
|
|
|
--executor-cores=${sparkExecutorCores}
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
|
|
|
--conf spark.extraListeners=${spark2ExtraListeners}
|
2021-12-14 09:56:55 +01:00
|
|
|
--conf spark.sql.shuffle.partitions=30000
|
2021-07-06 17:18:06 +02:00
|
|
|
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
|
|
|
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
|
|
|
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
|
|
|
</spark-opts>
|
|
|
|
<arg>--master</arg><arg>yarn</arg>
|
|
|
|
<arg>--summaryPath</arg><arg>${targetPath}/provision/summaries</arg>
|
|
|
|
<arg>--targetPath</arg><arg>${targetPath}/provision/scholix</arg>
|
2021-07-25 11:15:37 +02:00
|
|
|
<arg>--relationPath</arg><arg>${targetPath}/relation</arg>
|
2021-07-06 17:18:06 +02:00
|
|
|
|
|
|
|
</spark>
|
2020-02-19 10:07:08 +01:00
|
|
|
<ok to="End"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
<end name="End"/>
|
|
|
|
</workflow-app>
|