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

<workflow-app name="merge graphs" xmlns="uri:oozie:workflow:0.5">

    <parameters>
        <property>
            <name>betaInputGraph</name>
            <description>the beta graph name (or path)</description>
        </property>
        <property>
            <name>prodInputGraph</name>
            <description>the production graph name (or path)</description>
        </property>
        <property>
            <name>outputGraph</name>
            <description>the merged graph name (or path)</description>
        </property>
        <property>
            <name>priority</name>
            <description>decides from which infrastructure the content must win in case of ID clash</description>
        </property>
        <property>
            <name>inputGraphFormat</name>
            <value>HIVE</value>
            <description>the input graph data format</description>
        </property>
        <property>
            <name>outputGraphFormat</name>
            <value>HIVE</value>
            <description>the output graph data format</description>
        </property>

        <property>
            <name>hiveMetastoreUris</name>
            <description>hive server metastore URIs</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>
        <property>
            <name>sparkSqlWarehouseDir</name>
            <value>spark 2.* db directory location</value>
        </property>
    </parameters>

	<start to="should_drop_db"/>

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

    <decision name="should_drop_db">
        <switch>
            <case to="fork_merge_graph">${wf:conf('outputGraphFormat') eq 'JSON'}</case>
            <case to="reset_DB">${wf:conf('outputGraphFormat') eq 'HIVE'}</case>
            <default to="reset_DB"/>
        </switch>
    </decision>

    <action name="reset_DB">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>reset_DB</name>
            <class>eu.dnetlib.dhp.common.ResetHiveDbApplication</class>
            <jar>dhp-graph-mapper-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory ${sparkExecutorMemory}
                --executor-cores ${sparkExecutorCores}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=7680
            </spark-opts>
            <arg>--dbName</arg><arg>${outputGraph}</arg>
            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
        </spark>
        <ok to="fork_merge_graph"/>
        <error to="Kill"/>
    </action>

    <fork name="fork_merge_graph">
        <path start="merge_publication"/>
        <path start="merge_dataset"/>
        <path start="merge_otherresearchproduct"/>
        <path start="merge_software"/>
        <path start="merge_datasource"/>
        <path start="merge_organization"/>
        <path start="merge_project"/>
        <path start="merge_relation"/>
    </fork>

    <action name="merge_publication">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Merge publications</name>
            <class>eu.dnetlib.dhp.oa.graph.merge.MergeGraphSparkJob</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}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=7680
            </spark-opts>
            <arg>--betaInputGraph</arg><arg>${betaInputGraph}</arg>
            <arg>--prodInputGraph</arg><arg>${prodInputGraph}</arg>
            <arg>--outputGraph</arg><arg>${outputGraph}</arg>
            <arg>--inputGraphFormat</arg><arg>${inputGraphFormat}</arg>
            <arg>--outputGraphFormat</arg><arg>${outputGraphFormat}</arg>
            <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
            <arg>--priority</arg><arg>${priority}</arg>
            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
        </spark>
        <ok to="wait_merge"/>
        <error to="Kill"/>
    </action>

    <action name="merge_dataset">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Merge datasets</name>
            <class>eu.dnetlib.dhp.oa.graph.merge.MergeGraphSparkJob</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}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=7680
            </spark-opts>
            <arg>--betaInputGraph</arg><arg>${betaInputGraph}</arg>
            <arg>--prodInputGraph</arg><arg>${prodInputGraph}</arg>
            <arg>--outputGraph</arg><arg>${outputGraph}</arg>
            <arg>--inputGraphFormat</arg><arg>${inputGraphFormat}</arg>
            <arg>--outputGraphFormat</arg><arg>${outputGraphFormat}</arg>
            <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
            <arg>--priority</arg><arg>${priority}</arg>
            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
        </spark>
        <ok to="wait_merge"/>
        <error to="Kill"/>
    </action>

    <action name="merge_otherresearchproduct">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Merge otherresearchproducts</name>
            <class>eu.dnetlib.dhp.oa.graph.merge.MergeGraphSparkJob</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}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=7680
            </spark-opts>
            <arg>--betaInputGraph</arg><arg>${betaInputGraph}</arg>
            <arg>--prodInputGraph</arg><arg>${prodInputGraph}</arg>
            <arg>--outputGraph</arg><arg>${outputGraph}</arg>
            <arg>--inputGraphFormat</arg><arg>${inputGraphFormat}</arg>
            <arg>--outputGraphFormat</arg><arg>${outputGraphFormat}</arg>
            <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
            <arg>--priority</arg><arg>${priority}</arg>
            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
        </spark>
        <ok to="wait_merge"/>
        <error to="Kill"/>
    </action>

    <action name="merge_software">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Merge softwares</name>
            <class>eu.dnetlib.dhp.oa.graph.merge.MergeGraphSparkJob</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}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=7680
            </spark-opts>
            <arg>--betaInputGraph</arg><arg>${betaInputGraph}</arg>
            <arg>--prodInputGraph</arg><arg>${prodInputGraph}</arg>
            <arg>--outputGraph</arg><arg>${outputGraph}</arg>
            <arg>--inputGraphFormat</arg><arg>${inputGraphFormat}</arg>
            <arg>--outputGraphFormat</arg><arg>${outputGraphFormat}</arg>
            <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
            <arg>--priority</arg><arg>${priority}</arg>
            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
        </spark>
        <ok to="wait_merge"/>
        <error to="Kill"/>
    </action>

    <action name="merge_datasource">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Merge datasources</name>
            <class>eu.dnetlib.dhp.oa.graph.merge.MergeGraphSparkJob</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}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=7680
            </spark-opts>
            <arg>--betaInputGraph</arg><arg>${betaInputGraph}</arg>
            <arg>--prodInputGraph</arg><arg>${prodInputGraph}</arg>
            <arg>--outputGraph</arg><arg>${outputGraph}</arg>
            <arg>--inputGraphFormat</arg><arg>${inputGraphFormat}</arg>
            <arg>--outputGraphFormat</arg><arg>${outputGraphFormat}</arg>
            <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Datasource</arg>
            <arg>--priority</arg><arg>${priority}</arg>
            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
        </spark>
        <ok to="wait_merge"/>
        <error to="Kill"/>
    </action>

    <action name="merge_organization">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Merge organizations</name>
            <class>eu.dnetlib.dhp.oa.graph.merge.MergeGraphSparkJob</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}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=7680
            </spark-opts>
            <arg>--betaInputGraph</arg><arg>${betaInputGraph}</arg>
            <arg>--prodInputGraph</arg><arg>${prodInputGraph}</arg>
            <arg>--outputGraph</arg><arg>${outputGraph}</arg>
            <arg>--inputGraphFormat</arg><arg>${inputGraphFormat}</arg>
            <arg>--outputGraphFormat</arg><arg>${outputGraphFormat}</arg>
            <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Organization</arg>
            <arg>--priority</arg><arg>${priority}</arg>
            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
        </spark>
        <ok to="wait_merge"/>
        <error to="Kill"/>
    </action>

    <action name="merge_project">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Merge projects</name>
            <class>eu.dnetlib.dhp.oa.graph.merge.MergeGraphSparkJob</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}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=7680
            </spark-opts>
            <arg>--betaInputGraph</arg><arg>${betaInputGraph}</arg>
            <arg>--prodInputGraph</arg><arg>${prodInputGraph}</arg>
            <arg>--outputGraph</arg><arg>${outputGraph}</arg>
            <arg>--inputGraphFormat</arg><arg>${inputGraphFormat}</arg>
            <arg>--outputGraphFormat</arg><arg>${outputGraphFormat}</arg>
            <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Project</arg>
            <arg>--priority</arg><arg>${priority}</arg>
            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
        </spark>
        <ok to="wait_merge"/>
        <error to="Kill"/>
    </action>

    <action name="merge_relation">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Merge relations</name>
            <class>eu.dnetlib.dhp.oa.graph.merge.MergeGraphSparkJob</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}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=7680
            </spark-opts>
            <arg>--betaInputGraph</arg><arg>${betaInputGraph}</arg>
            <arg>--prodInputGraph</arg><arg>${prodInputGraph}</arg>
            <arg>--outputGraph</arg><arg>${outputGraph}</arg>
            <arg>--inputGraphFormat</arg><arg>${inputGraphFormat}</arg>
            <arg>--outputGraphFormat</arg><arg>${outputGraphFormat}</arg>
            <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Relation</arg>
            <arg>--priority</arg><arg>${priority}</arg>
            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
        </spark>
        <ok to="wait_merge"/>
        <error to="Kill"/>
    </action>

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

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