dnet-hadoop/dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/resolution/oozie_app/workflow.xml

<workflow-app name="Resolve relation and entities" xmlns="uri:oozie:workflow:0.5">
    <parameters>
        <property>
            <name>graphBasePath</name>
            <description>the path of the graph</description>
        </property>
        <property>
            <name>unresolvedPath</name>
            <description>the path of the unresolved Entities</description>
        </property>
        <property>
            <name>targetPath</name>
            <description>the target path after resolution</description>
        </property>
        <property>
            <name>shouldResolveEntities</name>
            <value>true</value>
            <description>allows to activate/deactivate the resolution process over the entities</description>
        </property>
    </parameters>

    <start to="reset_outputpath"/>

    <kill name="Kill">
        <message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
    </kill>

    <action name="reset_outputpath">
        <fs>
            <delete path="${targetPath}"/>
            <mkdir path="${targetPath}"/>
        </fs>
        <ok to="ResolveRelations"/>
        <error to="Kill"/>
    </action>

    <action name="ResolveRelations">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Resolve Relations in raw graph</name>
            <class>eu.dnetlib.dhp.oa.graph.resolution.SparkResolveRelation</class>
            <jar>dhp-graph-mapper-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory=${sparkExecutorMemory}
                --executor-cores=${sparkExecutorCores}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.shuffle.partitions=15000
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
            </spark-opts>
            <arg>--master</arg><arg>yarn</arg>
            <arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
            <arg>--workingPath</arg><arg>${workingDir}</arg>
            <arg>--targetPath</arg><arg>${targetPath}</arg>
        </spark>
        <ok to="decision_resolveEntities"/>
        <error to="Kill"/>
    </action>

    <decision name="decision_resolveEntities">
        <switch>
            <case to="copy_result">${wf:conf('shouldResolveEntities') eq false}</case>
            <case to="ResolveEntities">${wf:conf('shouldResolveEntities') eq true}</case>
            <default to="ResolveEntities"/>
        </switch>
    </decision>

    <action name="ResolveEntities">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Resolve Entities in raw graph</name>
            <class>eu.dnetlib.dhp.oa.graph.resolution.SparkResolveEntities</class>
            <jar>dhp-graph-mapper-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory=${sparkExecutorMemory}
                --executor-cores=${sparkExecutorCores}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.shuffle.partitions=10000
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
            </spark-opts>
            <arg>--master</arg><arg>yarn</arg>
            <arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
            <arg>--unresolvedPath</arg><arg>${unresolvedPath}</arg>
            <arg>--workingPath</arg><arg>${workingDir}</arg>
            <arg>--targetPath</arg><arg>${targetPath}</arg>
        </spark>
        <ok to="copy_entities"/>
        <error to="Kill"/>
    </action>

    <fork name="copy_result">
        <path start="copy_publication"/>
        <path start="copy_dataset"/>
        <path start="copy_otherresearchproduct"/>
        <path start="copy_software"/>
    </fork>

    <action name="copy_publication">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <arg>${nameNode}/${graphBasePath}/publication</arg>
            <arg>${nameNode}/${targetPath}/publication</arg>
        </distcp>
        <ok to="copy_wait_result"/>
        <error to="Kill"/>
    </action>

    <action name="copy_dataset">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <arg>${nameNode}/${graphBasePath}/dataset</arg>
            <arg>${nameNode}/${targetPath}/dataset</arg>
        </distcp>
        <ok to="copy_wait_result"/>
        <error to="Kill"/>
    </action>

    <action name="copy_otherresearchproduct">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <arg>${nameNode}/${graphBasePath}/otherresearchproduct</arg>
            <arg>${nameNode}/${targetPath}/otherresearchproduct</arg>
        </distcp>
        <ok to="copy_wait_result"/>
        <error to="Kill"/>
    </action>

    <action name="copy_software">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <arg>${nameNode}/${graphBasePath}/software</arg>
            <arg>${nameNode}/${targetPath}/software</arg>
        </distcp>
        <ok to="copy_wait_result"/>
        <error to="Kill"/>
    </action>

    <join name="copy_wait_result" to="copy_entities"/>

    <fork name="copy_entities">
        <path start="copy_organization"/>
        <path start="copy_projects"/>
        <path start="copy_datasource"/>
    </fork>

    <action name="copy_organization">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <arg>${nameNode}/${graphBasePath}/organization</arg>
            <arg>${nameNode}/${targetPath}/organization</arg>
        </distcp>
        <ok to="copy_wait"/>
        <error to="Kill"/>
    </action>

    <action name="copy_projects">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <arg>${nameNode}/${graphBasePath}/project</arg>
            <arg>${nameNode}/${targetPath}/project</arg>
        </distcp>
        <ok to="copy_wait"/>
        <error to="Kill"/>
    </action>

    <action name="copy_datasource">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <arg>${nameNode}/${graphBasePath}/datasource</arg>
            <arg>${nameNode}/${targetPath}/datasource</arg>
        </distcp>
        <ok to="copy_wait"/>
        <error to="Kill"/>
    </action>

    <join name="copy_wait" to="End"/>

    <end name="End"/>
</workflow-app>