added proeprties for the preparation step and actual propagation. Added the new parametrized workflow

This commit is contained in:
Miriam Baglioni 2020-04-23 12:13:34 +02:00
parent 6f35f5ca42
commit 15656684b9
4 changed files with 253 additions and 39 deletions

View File

@ -1,21 +1,51 @@
[ [
{
"paramName":"mt",
"paramLongName":"master",
"paramDescription": "should be local or yarn",
"paramRequired": true
},
{ {
"paramName":"s", "paramName":"s",
"paramLongName":"sourcePath", "paramLongName":"sourcePath",
"paramDescription": "the path of the sequencial file to read", "paramDescription": "the path of the sequencial file to read",
"paramRequired": true "paramRequired": true
}, },
{ {
"paramName":"ocm", "paramName":"h",
"paramLongName":"organizationtoresultcommunitymap", "paramLongName":"hive_metastore_uris",
"paramDescription": "the map in json format assocaition the organization original id with the list of communities", "paramDescription": "the hive metastore uris",
"paramRequired": true
},
{
"paramName":"sg",
"paramLongName":"saveGraph",
"paramDescription": "true if the new version of the graph must be saved",
"paramRequired": false
},
{
"paramName":"test",
"paramLongName":"isTest",
"paramDescription": "true if it is executing a test",
"paramRequired": false
},
{
"paramName": "out",
"paramLongName": "outputPath",
"paramDescription": "the path used to store temporary output files",
"paramRequired": true
},
{
"paramName": "ssm",
"paramLongName": "isSparkSessionManaged",
"paramDescription": "true if the spark session is managed, false otherwise",
"paramRequired": false
},
{
"paramName":"tn",
"paramLongName":"resultTableName",
"paramDescription": "the name of the result table we are currently working on",
"paramRequired": true
},
{
"paramName": "p",
"paramLongName": "preparedInfoPath",
"paramDescription": "the path where prepared info have been stored",
"paramRequired": true "paramRequired": true
} }
] ]

View File

@ -28,6 +28,12 @@
"paramLongName":"isTest", "paramLongName":"isTest",
"paramDescription": "true if it is executing a test", "paramDescription": "true if it is executing a test",
"paramRequired": false "paramRequired": false
},
{
"paramName": "out",
"paramLongName": "outputPath",
"paramDescription": "the path used to store temporary output files",
"paramRequired": true
} }
] ]

View File

@ -15,4 +15,44 @@
<name>oozie.action.sharelib.for.spark</name> <name>oozie.action.sharelib.for.spark</name>
<value>spark2</value> <value>spark2</value>
</property> </property>
<property>
<name>hive_metastore_uris</name>
<value>thrift://iis-cdh5-test-m3.ocean.icm.edu.pl:9083</value>
</property>
<property>
<name>spark2YarnHistoryServerAddress</name>
<value>http://iis-cdh5-test-gw.ocean.icm.edu.pl:18089</value>
</property>
<property>
<name>spark2EventLogDir</name>
<value>/user/spark/spark2ApplicationHistory</value>
</property>
<property>
<name>spark2ExtraListeners</name>
<value>com.cloudera.spark.lineage.NavigatorAppListener</value>
</property>
<property>
<name>spark2SqlQueryExecutionListeners</name>
<value>com.cloudera.spark.lineage.NavigatorQueryListener</value>
</property>
<property>
<name>sparkExecutorNumber</name>
<value>4</value>
</property>
<property>
<name>sparkDriverMemory</name>
<value>15G</value>
</property>
<property>
<name>sparkExecutorMemory</name>
<value>6G</value>
</property>
<property>
<name>sparkExecutorCores</name>
<value>1</value>
</property>
<property>
<name>spark2MaxExecutors</name>
<value>50</value>
</property>
</configuration> </configuration>

View File

@ -9,47 +9,185 @@
<description>organization community map</description> <description>organization community map</description>
</property> </property>
<property> <property>
<name>sparkDriverMemory</name> <name>writeUpdate</name>
<description>memory for driver process</description> <description>writes the information found for the update. No double check done if the information is already present</description>
</property> </property>
<property> <property>
<name>sparkExecutorMemory</name> <name>saveGraph</name>
<description>memory for individual executor</description> <description>writes new version of the graph after the propagation step</description>
</property>
<property>
<name>sparkExecutorCores</name>
<description>number of cores used by single executor</description>
</property> </property>
</parameters> </parameters>
<start to="CommunityToResultFromOrganizationPropagation"/> <start to="reset-outputpath"/>
<kill name="Kill"> <kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message> <message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill> </kill>
<action name="CommunityToResultFromOrganizationPropagation"> <action name="reset-outputpath">
<spark xmlns="uri:oozie:spark-action:0.2"> <fs>
<job-tracker>${jobTracker}</job-tracker> <delete path='${workingDir}/resulttocommunityorganization_propagation'/>
<name-node>${nameNode}</name-node> </fs>
<master>yarn-cluster</master> <ok to="copy_relation"/>
<mode>cluster</mode>
<name>ProjectToResultPropagation</name>
<class>eu.dnetlib.dhp.projecttoresult.SparkResultToProjectThroughSemRelJob</class>
<jar>dhp-propagation-${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"
</spark-opts>
<arg>-mt</arg> <arg>yarn-cluster</arg>
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--organizationtoresultcommunitymap</arg><arg>${organizationtoresultcommunitymap}</arg>
</spark>
<ok to="End"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<action name="copy_relation">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<arg>${nameNode}/${sourcePath}/relation</arg>
<arg>${nameNode}/${workingDir}/projecttoresult_propagation/relation</arg>
</distcp>
<ok to="prepare_result_communitylist"/>
<error to="Kill"/>
</action>
<action name="prepare_result_communitylist">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Prepare-Community-Result-Organization</name>
<class>eu.dnetlib.dhp.resulttocommunityfromorganization.PrepareResultCommunitySet</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${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}${spark2EventLogDir}
--conf spark.dynamicAllocation.enabled=true
--conf spark.dynamicAllocation.maxExecutors=${spark2MaxExecutors}
</spark-opts>
<arg>--sourcePath</arg><arg>${sourcePath}/relation</arg>
<arg>--hive_metastore_uris</arg><arg>${hive_metastore_uris}</arg>
<arg>--outputPath</arg><arg>${workingDir}/preparedInfo/resultCommunityList</arg>
<arg>--organizationtoresultcommunitymap</arg><arg>${organizationtoresultcommunitymap}</arg>
</spark>
<ok to="fork-join-exec-propagation"/>
<!-- <ok to="End"/>-->
<error to="Kill"/>
</action>
<fork name="fork-join-exec-propagation">
<path start="join_propagate_publication"/>
<path start="join_propagate_dataset"/>
<path start="join_propagate_otherresearchproduct"/>
<path start="join_propagate_software"/>
</fork>
<action name="join_propagate_publication">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>community2resultfromorganization-Publication</name>
<class>eu.dnetlib.dhp.resulttocommunityfromorganization.SparkResultToCommunityFromOrganizationJob2</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${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}${spark2EventLogDir}
--conf spark.dynamicAllocation.enabled=true
--conf spark.dynamicAllocation.maxExecutors=${spark2MaxExecutors}
</spark-opts>
<arg>--preparedInfoPath</arg><arg>${workingDir}/preparedInfo/resultCommunityList</arg>
<arg>--sourcePath</arg><arg>${sourcePath}/publication</arg>
<arg>--hive_metastore_uris</arg><arg>${hive_metastore_uris}</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
<arg>--outputPath</arg><arg>${workingDir}/publication</arg>
</spark>
<ok to="wait2"/>
<error to="Kill"/>
</action>
<action name="join_propagate_dataset">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>community2resultfromorganization-Dataset</name>
<class>eu.dnetlib.dhp.resulttocommunityfromorganization.SparkResultToCommunityFromOrganizationJob2</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${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}${spark2EventLogDir}
--conf spark.dynamicAllocation.enabled=true
--conf spark.dynamicAllocation.maxExecutors=${spark2MaxExecutors}
</spark-opts>
<arg>--preparedInfoPath</arg><arg>${workingDir}/preparedInfo/resultCommunityList</arg>
<arg>--sourcePath</arg><arg>${sourcePath}/dataset</arg>
<arg>--hive_metastore_uris</arg><arg>${hive_metastore_uris}</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
<arg>--outputPath</arg><arg>${workingDir}/dataset</arg>
</spark>
<ok to="wait2"/>
<error to="Kill"/>
</action>
<action name="join_propagate_otherresearchproduct">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>community2resultfromorganization-ORP</name>
<class>eu.dnetlib.dhp.resulttocommunityfromorganization.SparkResultToCommunityFromOrganizationJob2</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${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}${spark2EventLogDir}
--conf spark.dynamicAllocation.enabled=true
--conf spark.dynamicAllocation.maxExecutors=${spark2MaxExecutors}
</spark-opts>
<arg>--preparedInfoPath</arg><arg>${workingDir}/preparedInfo/resultCommunityList</arg>
<arg>--sourcePath</arg><arg>${sourcePath}/otherresearchproduct</arg>
<arg>--hive_metastore_uris</arg><arg>${hive_metastore_uris}</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
<arg>--outputPath</arg><arg>${workingDir}/otherresearchproduct</arg>
</spark>
<ok to="wait2"/>
<error to="Kill"/>
</action>
<action name="join_propagate_software">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>community2resultfromorganization-Software</name>
<class>eu.dnetlib.dhp.resulttocommunityfromorganization.SparkResultToCommunityFromOrganizationJob2</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${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}${spark2EventLogDir}
--conf spark.dynamicAllocation.enabled=true
--conf spark.dynamicAllocation.maxExecutors=${spark2MaxExecutors}
</spark-opts>
<arg>--preparedInfoPath</arg><arg>${workingDir}/preparedInfo/resultCommunityList</arg>
<arg>--sourcePath</arg><arg>${sourcePath}/software</arg>
<arg>--hive_metastore_uris</arg><arg>${hive_metastore_uris}</arg>
<arg>--resultTableName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
<arg>--outputPath</arg><arg>${workingDir}/software</arg>
</spark>
<ok to="wait2"/>
<error to="Kill"/>
</action>
<join name="wait2" to="End"/>
<end name="End"/> <end name="End"/>
</workflow-app> </workflow-app>