BrBETA_dnet-hadoop/dhp-workflows/dhp-graph-provision-scholexplorer/src/main/resources/eu/dnetlib/dhp/sx/provision/oozie_app/workflow.xml

<workflow-app name="Materialize and Index graph to ElasticSearch" xmlns="uri:oozie:workflow:0.5">
    <parameters>
        <property>
            <name>workingDirPath</name>
            <description>the source path</description>
        </property>
        <property>
            <name>graphPath</name>
            <description>the graph path</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>index</name>-->
<!--            <description>index name</description>-->
<!--        </property>-->
<!--        <property>-->
<!--            <name>idScholix</name>-->
<!--            <description>the identifier name of the scholix </description>-->
<!--        </property>-->
<!--        <property>-->
<!--            <name>idSummary</name>-->
<!--            <description>the identifier name of the summary</description>-->
<!--        </property>-->
    </parameters>

    <start to="DeleteTargetPath"/>

    <kill name="Kill">
        <message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
    </kill>
    <action name="DeleteTargetPath">
        <fs>
            <delete path='${workingDirPath}'/>
            <mkdir path='${workingDirPath}'/>
        </fs>
        <ok to="CalculateRelatedItem"/>
        <error to="Kill"/>
    </action>

    <action name="CalculateRelatedItem">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <job-tracker>${jobTracker}</job-tracker>
            <name-node>${nameNode}</name-node>
            <master>yarn-cluster</master>
            <mode>cluster</mode>
            <name>calculate for each ID the number of related Dataset, publication and Unknown</name>
            <class>eu.dnetlib.dhp.provision.SparkExtractRelationCount</class>
            <jar>dhp-graph-provision-scholexplorer-${projectVersion}.jar</jar>
            <spark-opts>--executor-memory ${sparkExecutorMemory} --driver-memory=${sparkDriverMemory} ${sparkExtraOPT}</spark-opts>
            <arg>-mt</arg> <arg>yarn-cluster</arg>
            <arg>--workingDirPath</arg><arg>${workingDirPath}</arg>
            <arg>--relationPath</arg><arg>${graphPath}/relation</arg>
        </spark>
        <ok to="generateSummary"/>
        <error to="Kill"/>
    </action>

    <action name="generateSummary">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <job-tracker>${jobTracker}</job-tracker>
            <name-node>${nameNode}</name-node>
            <master>yarn-cluster</master>
            <mode>cluster</mode>
            <name>generate Summary</name>
            <class>eu.dnetlib.dhp.provision.SparkGenerateSummary</class>
            <jar>dhp-graph-provision-scholexplorer-${projectVersion}.jar</jar>
            <spark-opts>--executor-memory ${sparkExecutorMemory} --driver-memory=${sparkDriverMemory} ${sparkExtraOPT}</spark-opts>
            <arg>-mt</arg> <arg>yarn-cluster</arg>
            <arg>--workingDirPath</arg><arg>${workingDirPath}</arg>
            <arg>--graphPath</arg><arg>${graphPath}</arg>
        </spark>
        <ok to="generateScholix"/>
        <error to="Kill"/>
    </action>

    <action name="generateScholix">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <job-tracker>${jobTracker}</job-tracker>
            <name-node>${nameNode}</name-node>
            <master>yarn-cluster</master>
            <mode>cluster</mode>
            <name>generate Scholix</name>
            <class>eu.dnetlib.dhp.provision.SparkGenerateScholix</class>
            <jar>dhp-graph-provision-scholexplorer-${projectVersion}.jar</jar>
            <spark-opts>--executor-memory 6G --driver-memory=${sparkDriverMemory} ${sparkExtraOPT}</spark-opts>
            <arg>-mt</arg> <arg>yarn-cluster</arg>
            <arg>--workingDirPath</arg><arg>${workingDirPath}</arg>
            <arg>--graphPath</arg><arg>${graphPath}</arg>
        </spark>
        <ok to="End"/>
        <error to="Kill"/>
    </action>

<!--    <action name="indexSummary">-->
<!--        <spark xmlns="uri:oozie:spark-action:0.2">-->
<!--            <job-tracker>${jobTracker}</job-tracker>-->
<!--            <name-node>${nameNode}</name-node>-->
<!--            <master>yarn-cluster</master>-->
<!--            <mode>cluster</mode>-->
<!--            <name>index Summary</name>-->
<!--            <class>eu.dnetlib.dhp.provision.SparkIndexCollectionOnES</class>-->
<!--            <jar>dhp-graph-provision-scholexplorer-${projectVersion}.jar</jar>-->
<!--            <spark-opts>&#45;&#45;executor-memory ${sparkExecutorMemory} &#45;&#45;driver-memory=${sparkDriverMemory} ${sparkExtraOPT} &#45;&#45;conf spark.dynamicAllocation.maxExecutors="32" </spark-opts>-->
<!--            <arg>-mt</arg> <arg>yarn-cluster</arg>-->
<!--            <arg>&#45;&#45;sourcePath</arg><arg>${workingDirPath}/summary</arg>-->
<!--            <arg>&#45;&#45;index</arg><arg>${index}_object</arg>-->
<!--            <arg>&#45;&#45;idPath</arg><arg>id</arg>-->
<!--            <arg>&#45;&#45;type</arg><arg>summary</arg>-->
<!--        </spark>-->
<!--        <ok to="indexScholix"/>-->
<!--        <error to="Kill"/>-->
<!--    </action>-->

<!--    <action name="indexScholix">-->
<!--        <spark xmlns="uri:oozie:spark-action:0.2">-->
<!--            <job-tracker>${jobTracker}</job-tracker>-->
<!--            <name-node>${nameNode}</name-node>-->
<!--            <master>yarn-cluster</master>-->
<!--            <mode>cluster</mode>-->
<!--            <name>index scholix</name>-->
<!--            <class>eu.dnetlib.dhp.provision.SparkIndexCollectionOnES</class>-->
<!--            <jar>dhp-graph-provision-scholexplorer-${projectVersion}.jar</jar>-->
<!--            <spark-opts>&#45;&#45;executor-memory ${sparkExecutorMemory}  &#45;&#45;driver-memory=${sparkDriverMemory} ${sparkExtraOPT} &#45;&#45;conf spark.dynamicAllocation.maxExecutors="8" </spark-opts>-->
<!--            <arg>-mt</arg> <arg>yarn-cluster</arg>-->
<!--            <arg>&#45;&#45;sourcePath</arg><arg>${workingDirPath}/scholix_json</arg>-->
<!--            <arg>&#45;&#45;index</arg><arg>${index}_scholix</arg>-->
<!--            <arg>&#45;&#45;idPath</arg><arg>identifier</arg>-->
<!--            <arg>&#45;&#45;type</arg><arg>scholix</arg>-->
<!--        </spark>-->
<!--        <ok to="End"/>-->
<!--        <error to="Kill"/>-->
<!--    </action>-->

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