BrBETA_dnet-hadoop/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/openorgs/oozie_app/workflow.xml

279 lines
12 KiB
XML

<workflow-app name="Openorgs Dedup" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>graphBasePath</name>
<description>the raw graph base path</description>
</property>
<property>
<name>isLookUpUrl</name>
<description>the address of the lookUp service</description>
</property>
<property>
<name>actionSetId</name>
<description>id of the actionSet</description>
</property>
<property>
<name>apiUrl</name>
<description>OpenOrgs API to finalise the suggestions import procedure</description>
</property>
<property>
<name>dbUrl</name>
<description>jdbc URL of the OpenOrgs database</description>
</property>
<property>
<name>dbUser</name>
<description>username to access the OpenOrgs database</description>
</property>
<property>
<name>dbPwd</name>
<description>password to access the OpenOrgs database</description>
</property>
<property>
<name>dbConnections</name>
<value>10</value>
<description>number of connections to the postgres db</description>
</property>
<property>
<name>workingPath</name>
<description>path for the working directory</description>
</property>
<property>
<name>cutConnectedComponent</name>
<description>max number of elements in a connected component</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>${oozieActionShareLibForSpark2}</value>
</property>
</configuration>
</global>
<start to="resume_from"/>
<decision name="resume_from">
<switch>
<case to="CreateSimRels">${wf:conf('resumeFrom') eq 'CreateSimRels'}</case>
<case to="CreateMergeRels">${wf:conf('resumeFrom') eq 'CreateMergeRels'}</case>
<case to="PrepareOrgRels">${wf:conf('resumeFrom') eq 'PrepareOrgRels'}</case>
<case to="update_openorgs">${wf:conf('resumeFrom') eq 'update_openorgs'}</case>
<default to="resetOrgSimRels"/>
</switch>
</decision>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<action name="resetOrgSimRels">
<fs>
<delete path="${workingPath}/${actionSetId}/organization_simrel"/>
<delete path="${workingPath}/${actionSetId}/organization_mergerel"/>
</fs>
<ok to="CreateSimRels"/>
<error to="Kill"/>
</action>
<action name="CreateSimRels">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Create Similarity Relations</name>
<class>eu.dnetlib.dhp.oa.dedup.SparkCreateSimRels</class>
<jar>dhp-dedup-openaire-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${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}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840
</spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
<arg>--actionSetId</arg><arg>${actionSetId}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--numPartitions</arg><arg>1000</arg>
</spark>
<ok to="CopyOpenorgsSimRels"/>
<error to="Kill"/>
</action>
<!-- copy simrels relations coming from openorgs -->
<action name="CopyOpenorgsSimRels">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Copy OpenOrgs Sim Rels</name>
<class>eu.dnetlib.dhp.oa.dedup.SparkCopyOpenorgsSimRels</class>
<jar>dhp-dedup-openaire-${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.shuffle.partitions=3840
</spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--actionSetId</arg><arg>${actionSetId}</arg>
<arg>--numPartitions</arg><arg>1000</arg>
</spark>
<ok to="CreateMergeRels"/>
<error to="Kill"/>
</action>
<action name="CreateMergeRels">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Create Merge Relations</name>
<class>eu.dnetlib.dhp.oa.dedup.SparkCreateMergeRels</class>
<jar>dhp-dedup-openaire-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${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}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840
</spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
<arg>--actionSetId</arg><arg>${actionSetId}</arg>
<arg>--cutConnectedComponent</arg><arg>${cutConnectedComponent}</arg>
</spark>
<ok to="PrepareOrgRels"/>
<error to="Kill"/>
</action>
<action name="PrepareOrgRels">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Prepare Organization Relations</name>
<class>eu.dnetlib.dhp.oa.dedup.SparkPrepareOrgRels</class>
<jar>dhp-dedup-openaire-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${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}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840
</spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
<arg>--actionSetId</arg><arg>${actionSetId}</arg>
<arg>--dbUrl</arg><arg>${dbUrl}</arg>
<arg>--dbTable</arg><arg>${dbTable}</arg>
<arg>--dbUser</arg><arg>${dbUser}</arg>
<arg>--dbPwd</arg><arg>${dbPwd}</arg>
<arg>--numConnections</arg><arg>${dbConnections}</arg>
</spark>
<ok to="PrepareNewOrgs"/>
<error to="Kill"/>
</action>
<action name="PrepareNewOrgs">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Prepare New Organizations</name>
<class>eu.dnetlib.dhp.oa.dedup.SparkPrepareNewOrgs</class>
<jar>dhp-dedup-openaire-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${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}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840
</spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
<arg>--actionSetId</arg><arg>${actionSetId}</arg>
<arg>--dbUrl</arg><arg>${dbUrl}</arg>
<arg>--dbTable</arg><arg>${dbTable}</arg>
<arg>--dbUser</arg><arg>${dbUser}</arg>
<arg>--dbPwd</arg><arg>${dbPwd}</arg>
<arg>--numConnections</arg><arg>${dbConnections}</arg>
</spark>
<ok to="update_openorgs"/>
<error to="Kill"/>
</action>
<action name="update_openorgs">
<java>
<main-class>eu.dnetlib.dhp.oa.dedup.UpdateOpenorgsJob</main-class>
<arg>--apiUrl</arg><arg>${apiUrl}</arg>
<arg>--delay</arg><arg>5</arg>
</java>
<ok to="End"/>
<error to="Kill"/>
</action>
<end name="End"/>
</workflow-app>