dhp-graph-dump/dump/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/skgif/oozie_app/workflow.xml

<workflow-app name="dump_graph" xmlns="uri:oozie:workflow:0.5">
    <parameters>
        <property>
            <name>sourcePath</name>
            <description>the source path</description>
        </property>

        <property>
            <name>outputPath</name>
            <description>the output 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>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="dump_venue"/>

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

    <action name="select_eosc_results">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Selecting subset of results</name>
            <class>eu.dnetlib.dhp.oa.graph.dump.filterentities.SelectEOSCEntities</class>
            <jar>dump-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=4
                --executor-memory=4G
                --driver-memory=${sparkDriverMemory}
                --conf spark.executor.memoryOverhead=5G
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=15000
            </spark-opts>
            <arg>--sourcePath</arg><arg>${sourcePath}</arg>
            <arg>--filterPath</arg><arg>${filterPath}/eoscIds/</arg>
        </spark>
        <ok to="filter"/>
        <error to="Kill"/>
    </action>

    <action name="filter">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Selecting subset of results</name>
            <class>eu.dnetlib.dhp.oa.graph.dump.filterentities.FilterEntities</class>
            <jar>dump-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=4
                --executor-memory=4G
                --driver-memory=${sparkDriverMemory}
                --conf spark.executor.memoryOverhead=5G
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                --conf spark.sql.shuffle.partitions=15000
            </spark-opts>
            <arg>--sourcePath</arg><arg>${sourcePath}</arg>
            <arg>--workingDir</arg><arg>${filterPath}/graph/</arg>
            <arg>--filterPath</arg><arg>${filterPath}/eoscIds/</arg>
        </spark>
        <ok to="select_relevant_graph_subset"/>
        <error to="Kill"/>
    </action>
    <action name="select_relevant_graph_subset">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Selecting relevant linked relations and entities</name>
            <class>eu.dnetlib.dhp.oa.graph.dump.filterentities.SelectConnectedEntities</class>
            <jar>dump-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=4
                --executor-memory=4G
                --driver-memory=${sparkDriverMemory}
                --conf spark.executor.memoryOverhead=5G
                --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=15000
            </spark-opts>
            <arg>--sourcePath</arg><arg>${sourcePath}</arg>
            <arg>--workingDir</arg><arg>${filterPath}/graph/</arg>
<!--            <arg>&#45;&#45;workingDir</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/graph/</arg>-->
            <arg>--filterPath</arg><arg>${filterPath}/eoscIds/</arg>
        </spark>
        <ok to="select_relation"/>

        <error to="Kill"/>
    </action>

    <action name="select_relation">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>select relations </name>
            <class>eu.dnetlib.dhp.oa.graph.dump.skgif.SelectRelation</class>
            <jar>dump-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=4
                --executor-memory=4G
                --driver-memory=${sparkDriverMemory}
                --conf spark.executor.memoryOverhead=5G
                --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=15000
            </spark-opts>
<!--            <arg>&#45;&#45;sourcePath</arg><arg>${workingDir}/graph/</arg>-->
            <arg>--sourcePath</arg><arg>${sourcePath}</arg>
            <arg>--relationPath</arg><arg>${sourcePath}/relation</arg>
            <!--            <arg>&#45;&#45;workingDir</arg><arg>${workingDir}/</arg>-->
            <arg>--workingDir</arg><arg>${filterPath}/</arg>
<!--            <arg>&#45;&#45;workingDir</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/</arg>-->
<!--            <arg>&#45;&#45;sourcePath</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/graph/</arg>-->
        </spark>

        <ok to="emit_from_result"/>
        <error to="Kill"/>
    </action>
    <action name="emit_from_result">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Emit from Result</name>
            <class>eu.dnetlib.dhp.oa.graph.dump.skgif.EmitFromEntities</class>
            <jar>dump-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory=${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}
            </spark-opts>
            <arg>--sourcePath</arg><arg>${filterPath}/graph/</arg>
            <arg>--workingDir</arg><arg>${filterPath}/</arg>
            <arg>--outputPath</arg><arg>${outputPath}</arg>
<!--            <arg>&#45;&#45;workingDir</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/</arg>-->
<!--            <arg>&#45;&#45;sourcePath</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/graph/</arg>-->
        </spark>
        <ok to="dump_research_product"/>
        <error to="Kill"/>
    </action>

<!--    <action name="dump_result">-->
<!--        <spark xmlns="uri:oozie:spark-action:0.2">-->
<!--            <master>yarn</master>-->
<!--            <mode>cluster</mode>-->
<!--            <name>Dump table results </name>-->
<!--            <class>eu.dnetlib.dhp.oa.graph.dump.skgif.DumpResult</class>-->
<!--            <jar>dump-${projectVersion}.jar</jar>-->
<!--            <spark-opts>-->
<!--                &#45;&#45;executor-cores=4-->
<!--                &#45;&#45;executor-memory=4G-->
<!--                &#45;&#45;driver-memory=${sparkDriverMemory}-->
<!--                &#45;&#45;conf spark.executor.memoryOverhead=5G-->
<!--                &#45;&#45;conf spark.extraListeners=${spark2ExtraListeners}-->
<!--                &#45;&#45;conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}-->
<!--                &#45;&#45;conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}-->
<!--                &#45;&#45;conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}-->
<!--                &#45;&#45;conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}-->
<!--                &#45;&#45;conf spark.sql.shuffle.partitions=15000-->
<!--            </spark-opts>-->
<!--&lt;!&ndash;            <arg>&#45;&#45;sourcePath</arg><arg>${workingDir}/graph/</arg>&ndash;&gt;-->
<!--            <arg>&#45;&#45;sourcePath</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/graph/</arg>-->
<!--            <arg>&#45;&#45;outputPath</arg><arg>${outputPath}</arg>-->
<!--&lt;!&ndash;            <arg>&#45;&#45;workingDir</arg><arg>${workingDir}/</arg>&ndash;&gt;-->
<!--            <arg>&#45;&#45;workingDir</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/</arg>-->
<!--        </spark>-->
<!--        <ok to="dump_datasource"/>-->
<!--        <error to="Kill"/>-->
<!--    </action>-->
    <action name="dump_datasource">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Dump table datasource </name>
            <class>eu.dnetlib.dhp.oa.graph.dump.skgif.DumpDatasource</class>
            <jar>dump-${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=${sparkSqlWarehouseDir}
            </spark-opts>
                        <arg>--sourcePath</arg><arg>${filterPath}/graph/</arg>
<!--            <arg>&#45;&#45;sourcePath</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/graph/</arg>-->
            <arg>--outputPath</arg><arg>${outputPath}</arg>
                        <arg>--workingDir</arg><arg>${filterPath}/</arg>
<!--            <arg>&#45;&#45;workingDir</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/</arg>-->
        </spark>
        <ok to="dump_venue"/>
        <error to="Kill"/>
    </action>
    <action name="dump_venue">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Dump table venues </name>
            <class>eu.dnetlib.dhp.oa.graph.dump.skgif.DumpVenue</class>
            <jar>dump-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory=${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}
            </spark-opts>
                        <arg>--sourcePath</arg><arg>${filterPath}/graph/</arg>
<!--            <arg>&#45;&#45;sourcePath</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/graph/</arg>-->
            <arg>--outputPath</arg><arg>${outputPath}</arg>
                        <arg>--workingDir</arg><arg>${filterPath}/</arg>
<!--            <arg>&#45;&#45;workingDir</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/</arg>-->
        </spark>
        <ok to="dump_organization"/>
        <error to="Kill"/>
    </action>

    <action name="dump_organization">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Dump table organization </name>
            <class>eu.dnetlib.dhp.oa.graph.dump.skgif.DumpOrganization</class>
            <jar>dump-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory=${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}
            </spark-opts>
                        <arg>--sourcePath</arg><arg>${filterPath}/graph/</arg>
<!--            <arg>&#45;&#45;sourcePath</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/graph/</arg>-->
            <arg>--outputPath</arg><arg>${outputPath}</arg>
                        <arg>--workingDir</arg><arg>${filterPath}/</arg>
<!--            <arg>&#45;&#45;workingDir</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/</arg>-->
        </spark>
        <ok to="dump_grant"/>
        <error to="Kill"/>
    </action>
    <action name="dump_grant">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Dump table project </name>
            <class>eu.dnetlib.dhp.oa.graph.dump.skgif.DumpGrant</class>
            <jar>dump-${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}

            </spark-opts>
                        <arg>--sourcePath</arg><arg>${filterPath}/graph/</arg>
<!--            <arg>&#45;&#45;sourcePath</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/graph/</arg>-->
            <arg>--outputPath</arg><arg>${outputPath}</arg>
                        <arg>--workingDir</arg><arg>${filterPath}/</arg>
<!--            <arg>&#45;&#45;workingDir</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/</arg>-->
        </spark>
        <ok to="End"/>
        <error to="Kill"/>
    </action>

    <action name="dump_research_product">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Dump table results </name>
            <class>eu.dnetlib.dhp.oa.graph.dump.skgif.DumpResearchProduct</class>
            <jar>dump-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=4
                --executor-memory=4G
                --driver-memory=${sparkDriverMemory}
                --conf spark.executor.memoryOverhead=5G
                --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=15000
            </spark-opts>
                        <arg>--sourcePath</arg><arg>${filterPath}/graph/</arg>
<!--            <arg>&#45;&#45;sourcePath</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/graph/</arg>-->
            <arg>--outputPath</arg><arg>${outputPath}</arg>
                        <arg>--workingDir</arg><arg>${filterPath}/</arg>
<!--            <arg>&#45;&#45;workingDir</arg><arg>/user/miriam.baglioni/oa/graph/dump/temp/working_dir/</arg>-->
        </spark>
        <ok to="dump_datasource"/>
        <error to="Kill"/>
    </action>
    <end name="End"/>

</workflow-app>