2020-03-24 17:40:58 +01:00
|
|
|
<workflow-app name="Duplicate Scan" xmlns="uri:oozie:workflow:0.5">
|
2020-03-18 14:27:49 +01:00
|
|
|
<parameters>
|
|
|
|
<property>
|
2020-03-19 15:01:07 +01:00
|
|
|
<name>graphBasePath</name>
|
2020-03-18 14:27:49 +01:00
|
|
|
<description>the raw graph base path</description>
|
|
|
|
</property>
|
2020-03-18 17:41:56 +01:00
|
|
|
<property>
|
|
|
|
<name>isLookUpUrl</name>
|
|
|
|
<description>the address of the lookUp service</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>actionSetId</name>
|
|
|
|
<description>id of the actionSet</description>
|
|
|
|
</property>
|
2020-03-19 15:01:07 +01:00
|
|
|
<property>
|
|
|
|
<name>workingPath</name>
|
|
|
|
<description>path for the working directory</description>
|
|
|
|
</property>
|
2020-03-18 14:27:49 +01:00
|
|
|
<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>
|
|
|
|
</parameters>
|
|
|
|
|
2020-03-20 16:48:14 +01:00
|
|
|
<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>
|
|
|
|
</configuration>
|
|
|
|
</global>
|
|
|
|
|
2020-03-24 17:40:58 +01:00
|
|
|
<start to="CreateSimRel"/>
|
2020-03-18 14:27:49 +01:00
|
|
|
|
|
|
|
<kill name="Kill">
|
|
|
|
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
|
|
|
|
</kill>
|
|
|
|
|
2020-03-24 17:40:58 +01:00
|
|
|
<action name="CreateSimRel">
|
2020-03-18 14:27:49 +01:00
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
2020-03-20 19:17:24 +01:00
|
|
|
<prepare>
|
|
|
|
<delete path="${workingPath}/${actionSetId}/*_simrel"/>
|
|
|
|
</prepare>
|
2020-03-20 16:48:14 +01:00
|
|
|
<master>yarn</master>
|
2020-03-18 14:27:49 +01:00
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Create Similarity Relations</name>
|
2020-03-20 16:48:14 +01:00
|
|
|
<class>eu.dnetlib.dhp.dedup.SparkCreateSimRels</class>
|
2020-03-18 14:27:49 +01:00
|
|
|
<jar>dhp-dedup-${projectVersion}.jar</jar>
|
2020-03-20 16:48:14 +01:00
|
|
|
<spark-opts>
|
|
|
|
--executor-memory ${sparkExecutorMemory}
|
|
|
|
--executor-cores ${sparkExecutorCores}
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
|
|
|
--conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener"
|
|
|
|
--conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener"
|
2020-03-18 14:27:49 +01:00
|
|
|
</spark-opts>
|
2020-03-20 16:48:14 +01:00
|
|
|
<arg>-mt</arg><arg>yarn</arg>
|
2020-03-20 13:01:56 +01:00
|
|
|
<arg>--i</arg><arg>${graphBasePath}</arg>
|
2020-03-18 17:41:56 +01:00
|
|
|
<arg>--la</arg><arg>${isLookUpUrl}</arg>
|
|
|
|
<arg>--asi</arg><arg>${actionSetId}</arg>
|
2020-03-20 13:01:56 +01:00
|
|
|
<arg>--w</arg><arg>${workingPath}</arg>
|
2020-03-18 14:27:49 +01:00
|
|
|
</spark>
|
2020-03-24 17:40:58 +01:00
|
|
|
<ok to="CreateMergeRel"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<action name="CreateMergeRel">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<prepare>
|
|
|
|
<delete path='${workingPath}/${actionSetId}/*_mergerel'/>
|
|
|
|
</prepare>
|
2020-03-25 11:15:35 +01:00
|
|
|
<master>yarn</master>
|
2020-03-24 17:40:58 +01:00
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Create Merge Relations</name>
|
|
|
|
<class>eu.dnetlib.dhp.dedup.SparkCreateConnectedComponent</class>
|
|
|
|
<jar>dhp-dedup-${projectVersion}.jar</jar>
|
|
|
|
<spark-opts>
|
|
|
|
--executor-memory ${sparkExecutorMemory}
|
|
|
|
--executor-cores ${sparkExecutorCores}
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
|
|
|
--conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener"
|
|
|
|
--conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener"
|
|
|
|
--conf spark.sql.warehouse.dir="/user/hive/warehouse"
|
|
|
|
</spark-opts>
|
2020-03-25 14:16:06 +01:00
|
|
|
<arg>-mt</arg><arg>yarn</arg>
|
2020-03-24 17:40:58 +01:00
|
|
|
<arg>--i</arg><arg>${graphBasePath}</arg>
|
|
|
|
<arg>--w</arg><arg>${workingPath}</arg>
|
|
|
|
<arg>--la</arg><arg>${isLookUpUrl}</arg>
|
|
|
|
<arg>--asi</arg><arg>${actionSetId}</arg>
|
|
|
|
</spark>
|
|
|
|
<ok to="CreateDedupRecord"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<action name="CreateDedupRecord">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<prepare>
|
|
|
|
<delete path='${workingPath}/${actionSetId}/*_deduprecord'/>
|
|
|
|
</prepare>
|
2020-03-25 11:15:35 +01:00
|
|
|
<master>yarn</master>
|
2020-03-24 17:40:58 +01:00
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>Create Dedup Record</name>
|
|
|
|
<class>eu.dnetlib.dhp.dedup.SparkCreateDedupRecord</class>
|
|
|
|
<jar>dhp-dedup-${projectVersion}.jar</jar>
|
|
|
|
<spark-opts>
|
|
|
|
--executor-memory ${sparkExecutorMemory}
|
|
|
|
--executor-cores ${sparkExecutorCores}
|
|
|
|
--driver-memory=${sparkDriverMemory}
|
|
|
|
--conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener"
|
|
|
|
--conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener"
|
|
|
|
--conf spark.sql.warehouse.dir="/user/hive/warehouse"
|
|
|
|
</spark-opts>
|
2020-03-25 14:16:06 +01:00
|
|
|
<arg>-mt</arg><arg>yarn</arg>
|
2020-03-24 17:40:58 +01:00
|
|
|
<arg>--i</arg><arg>${graphBasePath}</arg>
|
|
|
|
<arg>--w</arg><arg>${workingPath}</arg>
|
|
|
|
<arg>--la</arg><arg>${isLookUpUrl}</arg>
|
|
|
|
<arg>--asi</arg><arg>${actionSetId}</arg>
|
|
|
|
</spark>
|
2020-03-18 14:27:49 +01:00
|
|
|
<ok to="End"/>
|
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
<end name="End"/>
|
|
|
|
</workflow-app>
|