dnet-hadoop/dhp-workflows/dhp-dedup/src/main/resources/eu/dnetlib/dhp/dedup/oozie_app/workflow.xml

<workflow-app name="Dedup Entities" xmlns="uri:oozie:workflow:0.5">
    <parameters>
        <property>
            <name>sourcePath</name>
            <description>the source path</description>
        </property>
        <property>
            <name>entity</name>
            <description>the entity that should be processed</description>
        </property>
        <property>
            <name>dedupConf</name>
            <description>the dedup Configuration</description>
        </property>
        <property>
            <name>targetPath</name>
            <description>the target 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>
    </parameters>

    <start to="updateDeletedByInferenceRelation"/>


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

    <action name="DeleteWorkingPath">
        <fs>
            <delete path='${targetPath}/${entity}'/>
            <mkdir path="${targetPath}"/>  
            <mkdir path="${targetPath}/${entity}"/>              
        </fs>
        <ok to="CreateSimRels"/>
       <error to="Kill"/>
    </action>

    <action name="CreateSimRels">
        <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>Create Similarity Relations</name>
            <class>eu.dnetlib.dedup.SparkCreateSimRels</class>
            <jar>dhp-dedup-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory ${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                ${sparkExtraOPT}
            </spark-opts>
            <arg>-mt</arg><arg>yarn-cluster</arg>
            <arg>--sourcePath</arg><arg>${sourcePath}</arg>
            <arg>--targetPath</arg><arg>${targetPath}</arg>
            <arg>--entity</arg><arg>${entity}</arg>
            <arg>--dedupConf</arg><arg>${dedupConf}</arg>
        </spark>
        <ok to="CreateConnectedComponents"/>
        <error to="Kill"/>
    </action>

    <action name="CreateConnectedComponents">
        <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>Create Connected Components</name>
            <class>eu.dnetlib.dedup.SparkCreateConnectedComponent</class>
            <jar>dhp-dedup-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory ${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                ${sparkExtraOPT}
            </spark-opts>
            <arg>-mt</arg><arg>yarn-cluster</arg>
            <arg>--sourcePath</arg><arg>${sourcePath}</arg>
            <arg>--targetPath</arg><arg>${targetPath}</arg>
            <arg>--entity</arg><arg>${entity}</arg>
            <arg>--dedupConf</arg><arg>${dedupConf}</arg>
        </spark>
        <ok to="CreateDedupRecord"/>
        <error to="Kill"/>
    </action>

    <action name="CreateDedupRecord">
        <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>Create Dedup Record</name>
            <class>eu.dnetlib.dedup.SparkCreateDedupRecord</class>
            <jar>dhp-dedup-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory ${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                ${sparkExtraOPT}
            </spark-opts>
            <arg>-mt</arg><arg>yarn-cluster</arg>
            <arg>--sourcePath</arg><arg>${sourcePath}</arg>
            <arg>--dedupPath</arg><arg>${targetPath}</arg>
            <arg>--entity</arg><arg>${entity}</arg>
            <arg>--dedupConf</arg><arg>${dedupConf}</arg>
        </spark>
        <ok to="fixRelation"/>
        <error to="Kill"/>
    </action>

    <action name="fixRelation">
        <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>Propagate Dedup Relations</name>
            <class>eu.dnetlib.dedup.SparkPropagateRelationsJob</class>
            <jar>dhp-dedup-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory ${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                ${sparkExtraOPT}
            </spark-opts>
            <arg>-mt</arg><arg>yarn-cluster</arg>
           <arg>--mergeRelPath</arg><arg>${targetPath}/${entity}/mergeRel</arg>
            <arg>--relationPath</arg><arg>${sourcePath}/relation</arg>
            <arg>--targetRelPath</arg><arg>${targetPath}/${entity}/relation_propagated</arg>
        </spark>
        <ok to="updateDeletedByInferenceEntity"/>
        <error to="Kill"/>
    </action>


    <action name="updateDeletedByInferenceEntity">
        <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>Update ${entity} and add DedupRecord</name>
            <class>eu.dnetlib.dedup.SparkUpdateEntityJob</class>
            <jar>dhp-dedup-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory ${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                ${sparkExtraOPT}
            </spark-opts>
            <arg>-mt</arg><arg>yarn-cluster</arg>
            <arg>--entityPath</arg><arg>${sourcePath}/${entity}</arg>
            <arg>--mergeRelPath</arg><arg>${targetPath}/${entity}/mergeRel</arg>
            <arg>--entity</arg><arg>${entity}</arg>
            <arg>--dedupRecordPath</arg><arg>${targetPath}/${entity}/dedup_records</arg>
            <arg>--targetPath</arg><arg>${targetPath}/${entity}/updated_record</arg>
        </spark>
        <ok to="updateDeletedByInferenceRelation"/>
        <error to="Kill"/>
    </action>

    <action name="updateDeletedByInferenceRelation">
        <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>Update ${entity} set deleted by Inference</name>
            <class>eu.dnetlib.dedup.SparkUpdateEntityJob</class>
            <jar>dhp-dedup-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory ${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                ${sparkExtraOPT}
            </spark-opts>
            <arg>-mt</arg><arg>yarn-cluster</arg>
            <arg>--entityPath</arg><arg>${targetPath}/${entity}/relation_propagated</arg>
            <arg>--mergeRelPath</arg><arg>${targetPath}/${entity}/mergeRel</arg>
            <arg>--entity</arg><arg>relation</arg>
            <arg>--dedupRecordPath</arg><arg>${targetPath}/${entity}/dedup_records</arg>
            <arg>--targetPath</arg><arg>${targetPath}/${entity}/updated_relation</arg>
        </spark>
        <ok to="replaceEntity"/>
        <error to="Kill"/>
    </action>


    <action name="replaceEntity">
        <fs>
            <delete path='${sourcePath}/${entity}'/>
            <delete path='${sourcePath}/relation'/>
            <move source="${targetPath}/${entity}/updated_relation" target="${sourcePath}/relation" />
            <move source="${targetPath}/${entity}/updated_record" target="${sourcePath}/${entity}" />
        </fs>
        <ok to="End"/>
        <error to="Kill"/>
    </action>


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