2021-07-13 18:06:44 +02:00
|
|
|
<workflow-app name="sub_dump_community_products" xmlns="uri:oozie:workflow:0.5">
|
|
|
|
|
|
|
|
<parameters>
|
|
|
|
<property>
|
|
|
|
<name>sourcePath</name>
|
|
|
|
<description>the source path</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>outputPath</name>
|
|
|
|
<description>the output path</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>hiveDbName</name>
|
|
|
|
<description>the target hive database name</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>hiveJdbcUrl</name>
|
|
|
|
<description>hive server jdbc url</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>hiveMetastoreUris</name>
|
|
|
|
<description>hive server metastore URIs</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>sparkDriverMemory</name>
|
|
|
|
<description>memory for driver process</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>sparkExecutorMemory</name>
|
|
|
|
<description>memory for individual executor</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>sparkExecutorCores</name>
|
|
|
|
<description>number of cores used by single executor</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>oozieActionShareLibForSpark2</name>
|
|
|
|
<description>oozie action sharelib for spark 2.*</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>spark2ExtraListeners</name>
|
|
|
|
<value>com.cloudera.spark.lineage.NavigatorAppListener</value>
|
|
|
|
<description>spark 2.* extra listeners classname</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>spark2SqlQueryExecutionListeners</name>
|
|
|
|
<value>com.cloudera.spark.lineage.NavigatorQueryListener</value>
|
|
|
|
<description>spark 2.* sql query execution listeners classname</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>spark2YarnHistoryServerAddress</name>
|
|
|
|
<description>spark 2.* yarn history server address</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>spark2EventLogDir</name>
|
|
|
|
<description>spark 2.* event log dir location</description>
|
|
|
|
</property>
|
|
|
|
</parameters>
|
|
|
|
|
|
|
|
<global>
|
|
|
|
<job-tracker>${jobTracker}</job-tracker>
|
|
|
|
<name-node>${nameNode}</name-node>
|
|
|
|
<configuration>
|
|
|
|
<property>
|
|
|
|
<name>mapreduce.job.queuename</name>
|
|
|
|
<value>${queueName}</value>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>oozie.launcher.mapred.job.queue.name</name>
|
|
|
|
<value>${oozieLauncherQueueName}</value>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>oozie.action.sharelib.for.spark</name>
|
|
|
|
<value>${oozieActionShareLibForSpark2}</value>
|
|
|
|
</property>
|
|
|
|
|
|
|
|
</configuration>
|
|
|
|
</global>
|
|
|
|
|
2022-04-05 15:04:52 +02:00
|
|
|
<start to="fork_dump"/>
|
2021-07-13 18:06:44 +02:00
|
|
|
|
|
|
|
<kill name="Kill">
|
|
|
|
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
|
|
|
|
</kill>
|
|
|
|
|
2022-04-05 15:04:52 +02:00
|
|
|
<fork name="fork_dump">
|
|
|
|
<path start="dump_publication"/>
|
|
|
|
<path start="dump_dataset"/>
|
|
|
|
<path start="dump_orp"/>
|
|
|
|
<path start="dump_software"/>
|
|
|
|
</fork>
|
|
|
|
|
|
|
|
<action name="dump_publication">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Dump table publication for community/funder related products</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkDumpCommunityProducts</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>${selectedResults}/publication</arg>
|
|
|
|
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
|
|
|
|
<arg>--outputPath</arg><arg>${workingDir}/dump/publication</arg>
|
|
|
|
<arg>--communityMapPath</arg><arg>${communityMapPath}</arg>
|
|
|
|
<arg>--dumpType</arg><arg>${dumpType}</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="join_dump"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<action name="dump_dataset">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Dump table dataset for community/funder related products</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkDumpCommunityProducts</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>${selectedResults}/dataset</arg>
|
|
|
|
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
|
|
|
|
<arg>--outputPath</arg><arg>${workingDir}/dump/dataset</arg>
|
|
|
|
<arg>--communityMapPath</arg><arg>${communityMapPath}</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="join_dump"/>
|
|
|
|
<error to="Kill"/>
|
2021-07-13 18:06:44 +02:00
|
|
|
</action>
|
|
|
|
|
2022-04-05 15:04:52 +02:00
|
|
|
<action name="dump_orp">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Dump table ORP for community related products</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkDumpCommunityProducts</class>
|
|
|
|
<jar>dhp-graph-mapper-${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=${sparkSqlWarehouseDir}
|
|
|
|
</spark-opts>
|
|
|
|
<arg>--sourcePath</arg><arg>${selectedResults}/otherresearchproduct</arg>
|
|
|
|
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
|
|
|
|
<arg>--outputPath</arg><arg>${workingDir}/dump/otherresearchproduct</arg>
|
|
|
|
<arg>--communityMapPath</arg><arg>${communityMapPath}</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="join_dump"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<action name="dump_software">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Dump table software for community related products</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkDumpCommunityProducts</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>${selectedResults}/software</arg>
|
|
|
|
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
|
|
|
|
<arg>--outputPath</arg><arg>${workingDir}/dump/software</arg>
|
|
|
|
<arg>--communityMapPath</arg><arg>${communityMapPath}</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="join_dump"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<join name="join_dump" to="prepareResultProject"/>
|
|
|
|
|
|
|
|
<action name="prepareResultProject">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Prepare association result subset of project info</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkPrepareResultProject</class>
|
|
|
|
<jar>dhp-graph-mapper-${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=${sparkSqlWarehouseDir}
|
|
|
|
</spark-opts>
|
|
|
|
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
|
|
|
|
<arg>--outputPath</arg><arg>${workingDir}/preparedInfo</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="fork_extendWithProject"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<fork name="fork_extendWithProject">
|
|
|
|
<path start="extend_publication"/>
|
|
|
|
<path start="extend_dataset"/>
|
|
|
|
<path start="extend_orp"/>
|
|
|
|
<path start="extend_software"/>
|
|
|
|
</fork>
|
|
|
|
|
|
|
|
<action name="extend_publication">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Extend dumped publications with information about project</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkUpdateProjectInfo</class>
|
|
|
|
<jar>dhp-graph-mapper-${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}/dump/publication</arg>
|
|
|
|
<arg>--outputPath</arg><arg>${outputPath}/ext/publication</arg>
|
|
|
|
<arg>--preparedInfoPath</arg><arg>${workingDir}/preparedInfo</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="join_extend"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<action name="extend_dataset">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Extend dumped dataset with information about project</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkUpdateProjectInfo</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>${workingDir}/dump/dataset</arg>
|
|
|
|
<arg>--outputPath</arg><arg>${outputPath}/ext/dataset</arg>
|
|
|
|
<arg>--preparedInfoPath</arg><arg>${workingDir}/preparedInfo</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="join_extend"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<action name="extend_orp">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Extend dumped ORP with information about project</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkUpdateProjectInfo</class>
|
|
|
|
<jar>dhp-graph-mapper-${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}/dump/otherresearchproduct</arg>
|
|
|
|
<arg>--outputPath</arg><arg>${outputPath}/ext/orp</arg>
|
|
|
|
<arg>--preparedInfoPath</arg><arg>${workingDir}/preparedInfo</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="join_extend"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<action name="extend_software">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Extend dumped software with information about project</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkUpdateProjectInfo</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>${workingDir}/dump/software</arg>
|
|
|
|
<arg>--outputPath</arg><arg>${outputPath}/ext/software</arg>
|
|
|
|
<arg>--preparedInfoPath</arg><arg>${workingDir}/preparedInfo</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="join_extend"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
2021-07-13 18:06:44 +02:00
|
|
|
|
|
|
|
|
2022-04-05 15:04:52 +02:00
|
|
|
<join name="join_extend" to="splitForCommunities"/>
|
2021-07-13 18:06:44 +02:00
|
|
|
<action name="splitForCommunities">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Split dumped result for community</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkSplitForCommunity</class>
|
|
|
|
<jar>dhp-graph-mapper-${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>${outputPath}</arg>
|
|
|
|
<arg>--communityMapPath</arg><arg>${communityMapPath}</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="End"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<end name="End"/>
|
|
|
|
|
|
|
|
</workflow-app>
|