factored out entity grouping implementation, extended to consider results from delegated authorities rather than identical records from other sources
parent
42e8f76778
commit
44a937f4ed
@ -0,0 +1,26 @@
|
||||
[
|
||||
{
|
||||
"paramName": "issm",
|
||||
"paramLongName": "isSparkSessionManaged",
|
||||
"paramDescription": "when true will stop SparkSession after job execution",
|
||||
"paramRequired": false
|
||||
},
|
||||
{
|
||||
"paramName": "i",
|
||||
"paramLongName": "inputPath",
|
||||
"paramDescription": "the source path",
|
||||
"paramRequired": true
|
||||
},
|
||||
{
|
||||
"paramName": "o",
|
||||
"paramLongName": "outputPath",
|
||||
"paramDescription": "path of the output graph",
|
||||
"paramRequired": true
|
||||
},
|
||||
{
|
||||
"paramName": "c",
|
||||
"paramLongName": "graphTableClassName",
|
||||
"paramDescription": "the graph entity class name",
|
||||
"paramRequired": true
|
||||
}
|
||||
]
|
||||
@ -0,0 +1,20 @@
|
||||
[
|
||||
{
|
||||
"paramName": "issm",
|
||||
"paramLongName": "isSparkSessionManaged",
|
||||
"paramDescription": "when true will stop SparkSession after job execution",
|
||||
"paramRequired": false
|
||||
},
|
||||
{
|
||||
"paramName": "gin",
|
||||
"paramLongName": "graphInputPath",
|
||||
"paramDescription": "the graph root path",
|
||||
"paramRequired": true
|
||||
},
|
||||
{
|
||||
"paramName": "out",
|
||||
"paramLongName": "outputPath",
|
||||
"paramDescription": "the output merged graph root path",
|
||||
"paramRequired": true
|
||||
}
|
||||
]
|
||||
@ -0,0 +1,18 @@
|
||||
<configuration>
|
||||
<property>
|
||||
<name>jobTracker</name>
|
||||
<value>yarnRM</value>
|
||||
</property>
|
||||
<property>
|
||||
<name>nameNode</name>
|
||||
<value>hdfs://nameservice1</value>
|
||||
</property>
|
||||
<property>
|
||||
<name>oozie.use.system.libpath</name>
|
||||
<value>true</value>
|
||||
</property>
|
||||
<property>
|
||||
<name>oozie.action.sharelib.for.spark</name>
|
||||
<value>spark2</value>
|
||||
</property>
|
||||
</configuration>
|
||||
@ -0,0 +1,289 @@
|
||||
<workflow-app name="Group Graph Entities" xmlns="uri:oozie:workflow:0.5">
|
||||
<parameters>
|
||||
<property>
|
||||
<name>graphBasePath</name>
|
||||
<description>the input graph base path</description>
|
||||
</property>
|
||||
<property>
|
||||
<name>workingPath</name>
|
||||
<description>path of the working directory</description>
|
||||
</property>
|
||||
<property>
|
||||
<name>graphOutputPath</name>
|
||||
<description>path of the output graph</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>${oozieActionShareLibForSpark2}</value>
|
||||
</property>
|
||||
</configuration>
|
||||
</global>
|
||||
|
||||
<start to="group_entities"/>
|
||||
|
||||
<kill name="Kill">
|
||||
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
|
||||
</kill>
|
||||
|
||||
<action name="group_entities">
|
||||
<spark xmlns="uri:oozie:spark-action:0.2">
|
||||
<master>yarn</master>
|
||||
<mode>cluster</mode>
|
||||
<name>group graph entities</name>
|
||||
<class>eu.dnetlib.dhp.oa.merge.GroupEntitiesSparkJob</class>
|
||||
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
||||
<spark-opts>
|
||||
--executor-cores=${sparkExecutorCores}
|
||||
--executor-memory=${sparkExecutorMemory}
|
||||
--driver-memory=${sparkDriverMemory}
|
||||
--conf spark.extraListeners=${spark2ExtraListeners}
|
||||
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
||||
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
||||
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
||||
--conf spark.sql.shuffle.partitions=15000
|
||||
</spark-opts>
|
||||
<arg>--graphInputPath</arg><arg>${graphBasePath}</arg>
|
||||
<arg>--outputPath</arg><arg>${workingPath}/grouped_entities</arg>
|
||||
</spark>
|
||||
<ok to="fork_dispatch_entities"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
<fork name="fork_dispatch_entities">
|
||||
<path start="dispatch_datasource"/>
|
||||
<path start="dispatch_project"/>
|
||||
<path start="dispatch_organization"/>
|
||||
<path start="dispatch_publication"/>
|
||||
<path start="dispatch_dataset"/>
|
||||
<path start="dispatch_software"/>
|
||||
<path start="dispatch_otherresearchproduct"/>
|
||||
</fork>
|
||||
|
||||
<action name="dispatch_datasource">
|
||||
<spark xmlns="uri:oozie:spark-action:0.2">
|
||||
<master>yarn</master>
|
||||
<mode>cluster</mode>
|
||||
<name>Dispatch publications</name>
|
||||
<class>eu.dnetlib.dhp.oa.merge.DispatchEntitiesSparkJob</class>
|
||||
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
||||
<spark-opts>
|
||||
--executor-cores=${sparkExecutorCores}
|
||||
--executor-memory=${sparkExecutorMemory}
|
||||
--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.shuffle.partitions=7680
|
||||
</spark-opts>
|
||||
<arg>--inputPath</arg><arg>${workingPath}/grouped_entities</arg>
|
||||
<arg>--outputPath</arg><arg>${graphOutputPath}/datasource</arg>
|
||||
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Datasource</arg>
|
||||
</spark>
|
||||
<ok to="wait_dispatch"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
<action name="dispatch_project">
|
||||
<spark xmlns="uri:oozie:spark-action:0.2">
|
||||
<master>yarn</master>
|
||||
<mode>cluster</mode>
|
||||
<name>Dispatch project</name>
|
||||
<class>eu.dnetlib.dhp.oa.merge.DispatchEntitiesSparkJob</class>
|
||||
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
||||
<spark-opts>
|
||||
--executor-cores=${sparkExecutorCores}
|
||||
--executor-memory=${sparkExecutorMemory}
|
||||
--driver-memory=${sparkDriverMemory}
|
||||
--conf spark.extraListeners=${spark2ExtraListeners}
|
||||
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
||||
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
||||
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
||||
--conf spark.sql.shuffle.partitions=7680
|
||||
</spark-opts>
|
||||
<arg>--inputPath</arg><arg>${workingPath}/grouped_entities</arg>
|
||||
<arg>--outputPath</arg><arg>${graphOutputPath}/project</arg>
|
||||
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Project</arg>
|
||||
</spark>
|
||||
<ok to="wait_dispatch"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
<action name="dispatch_organization">
|
||||
<spark xmlns="uri:oozie:spark-action:0.2">
|
||||
<master>yarn</master>
|
||||
<mode>cluster</mode>
|
||||
<name>Dispatch organization</name>
|
||||
<class>eu.dnetlib.dhp.oa.merge.DispatchEntitiesSparkJob</class>
|
||||
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
||||
<spark-opts>
|
||||
--executor-cores=${sparkExecutorCores}
|
||||
--executor-memory=${sparkExecutorMemory}
|
||||
--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.shuffle.partitions=7680
|
||||
</spark-opts>
|
||||
<arg>--inputPath</arg><arg>${workingPath}/grouped_entities</arg>
|
||||
<arg>--outputPath</arg><arg>${graphOutputPath}/organization</arg>
|
||||
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Organization</arg>
|
||||
</spark>
|
||||
<ok to="wait_dispatch"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
<action name="dispatch_publication">
|
||||
<spark xmlns="uri:oozie:spark-action:0.2">
|
||||
<master>yarn</master>
|
||||
<mode>cluster</mode>
|
||||
<name>Dispatch publication</name>
|
||||
<class>eu.dnetlib.dhp.oa.merge.DispatchEntitiesSparkJob</class>
|
||||
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
||||
<spark-opts>
|
||||
--executor-cores=${sparkExecutorCores}
|
||||
--executor-memory=${sparkExecutorMemory}
|
||||
--driver-memory=${sparkDriverMemory}
|
||||
--conf spark.extraListeners=${spark2ExtraListeners}
|
||||
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
||||
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
||||
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
||||
--conf spark.sql.shuffle.partitions=7680
|
||||
</spark-opts>
|
||||
<arg>--inputPath</arg><arg>${workingPath}/grouped_entities</arg>
|
||||
<arg>--outputPath</arg><arg>${graphOutputPath}/publication</arg>
|
||||
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
|
||||
</spark>
|
||||
<ok to="wait_dispatch"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
<action name="dispatch_dataset">
|
||||
<spark xmlns="uri:oozie:spark-action:0.2">
|
||||
<master>yarn</master>
|
||||
<mode>cluster</mode>
|
||||
<name>Dispatch dataset</name>
|
||||
<class>eu.dnetlib.dhp.oa.merge.DispatchEntitiesSparkJob</class>
|
||||
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
||||
<spark-opts>
|
||||
--executor-cores=${sparkExecutorCores}
|
||||
--executor-memory=${sparkExecutorMemory}
|
||||
--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.shuffle.partitions=7680
|
||||
</spark-opts>
|
||||
<arg>--inputPath</arg><arg>${workingPath}/grouped_entities</arg>
|
||||
<arg>--outputPath</arg><arg>${graphOutputPath}/dataset</arg>
|
||||
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
|
||||
</spark>
|
||||
<ok to="wait_dispatch"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
<action name="dispatch_software">
|
||||
<spark xmlns="uri:oozie:spark-action:0.2">
|
||||
<master>yarn</master>
|
||||
<mode>cluster</mode>
|
||||
<name>Dispatch software</name>
|
||||
<class>eu.dnetlib.dhp.oa.merge.DispatchEntitiesSparkJob</class>
|
||||
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
||||
<spark-opts>
|
||||
--executor-cores=${sparkExecutorCores}
|
||||
--executor-memory=${sparkExecutorMemory}
|
||||
--driver-memory=${sparkDriverMemory}
|
||||
--conf spark.extraListeners=${spark2ExtraListeners}
|
||||
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
||||
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
||||
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
||||
--conf spark.sql.shuffle.partitions=7680
|
||||
</spark-opts>
|
||||
<arg>--inputPath</arg><arg>${workingPath}/grouped_entities</arg>
|
||||
<arg>--outputPath</arg><arg>${graphOutputPath}/software</arg>
|
||||
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
|
||||
</spark>
|
||||
<ok to="wait_dispatch"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
<action name="dispatch_otherresearchproduct">
|
||||
<spark xmlns="uri:oozie:spark-action:0.2">
|
||||
<master>yarn</master>
|
||||
<mode>cluster</mode>
|
||||
<name>Dispatch otherresearchproduct</name>
|
||||
<class>eu.dnetlib.dhp.oa.merge.DispatchEntitiesSparkJob</class>
|
||||
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
|
||||
<spark-opts>
|
||||
--executor-cores=${sparkExecutorCores}
|
||||
--executor-memory=${sparkExecutorMemory}
|
||||
--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.shuffle.partitions=7680
|
||||
</spark-opts>
|
||||
<arg>--inputPath</arg><arg>${workingPath}/grouped_entities</arg>
|
||||
<arg>--outputPath</arg><arg>${graphOutputPath}/otherresearchproduct</arg>
|
||||
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
|
||||
</spark>
|
||||
<ok to="wait_dispatch"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
<join name="wait_dispatch" to="End"/>
|
||||
|
||||
<end name="End"/>
|
||||
</workflow-app>
|
||||
Loading…
Reference in New Issue