Minor fixes in workflow.xml and job.properties

This commit is contained in:
Serafeim Chatzopoulos 2023-06-23 12:51:50 +03:00
parent a1b9187039
commit 60f25b780d
2 changed files with 49 additions and 48 deletions

View File

@ -102,4 +102,4 @@ actionSetOutputPath=${workingDir}/bip_actionsets/
# The directory to store project impact indicators
projectImpactIndicatorsOutput=${workingDir}/project_indicators
resume=create-openaire-ranking-graph
resume=entry-point-decision

View File

@ -33,7 +33,7 @@
<default to="create-openaire-ranking-graph" />
</switch>
</decision>
<!-- Script here written by Serafeim: maps openaire ids to their synonyms -->
<action name="create-openaire-ranking-graph">
<!-- This is required as a tag for spark jobs, regardless of programming language -->
@ -44,12 +44,12 @@
<prepare>
<delete path="${synonymFolder}"/>
</prepare>
-->
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
-->
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<mode>cluster</mode>
<!-- This is the name of our job -->
<name>Openaire Ranking Graph Creation</name>
<!-- Script name goes here -->
@ -141,8 +141,8 @@
<!-- This is required as a tag for spark jobs, regardless of programming language -->
<spark xmlns="uri:oozie:spark-action:0.2">
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<mode>cluster</mode>
<!-- This is the name of our job -->
@ -244,8 +244,8 @@
<!-- Reference says: The mode element if present indicates the mode of spark, where to run spark driver program. Ex: client,cluster. | In my case I always have a client -->
<!-- <mode>client</mode> -->
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<mode>cluster</mode>
<!-- This is the name of our job -->
@ -289,8 +289,8 @@
<!-- This is required as a tag for spark jobs, regardless of programming language -->
<spark xmlns="uri:oozie:spark-action:0.2">
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<mode>cluster</mode>
<!-- This is the name of our job -->
@ -335,8 +335,8 @@
</action>
<!-- JOIN ITERATIVE METHODS AND THEN END -->
<join name="join-iterative-rankings" to="end" />
<!-- to="get-file-names"/> -->
<join name="join-iterative-rankings" to="get-file-names" />
<!-- to="get-file-names"/> -->
<!-- This will be a shell action that will output key-value pairs for output files -->
@ -349,7 +349,7 @@
<!-- name of script to run -->
<argument>get_ranking_files.sh</argument>
<!-- We only pass the directory where we expect to find the rankings -->
<argument>/${workingDir}</argument>
<argument>${workingDir}</argument>
<!-- the name of the file run -->
<file>${wfAppPath}/get_ranking_files.sh#get_ranking_files.sh</file>
@ -481,8 +481,8 @@
<delete path="${synonymFolder}"/>
</prepare>
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<mode>cluster</mode>
<!-- This is the name of our job -->
@ -503,7 +503,7 @@
</spark-opts>
<!-- Script arguments here -->
<arg>${openaireDataInput}</arg>
<arg>${openaireDataInput}/</arg>
<!-- number of partitions to be used on joins -->
<arg>${synonymFolder}</arg>
<!-- This needs to point to the file on the hdfs i think -->
@ -523,8 +523,8 @@
<!-- This is required as a tag for spark jobs, regardless of programming language -->
<spark xmlns="uri:oozie:spark-action:0.2">
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<!-- using configs from an example on openaire -->
<master>yarn-cluster</master>
<mode>cluster</mode>
<!-- This is the name of our job -->
@ -561,47 +561,48 @@
<!-- Do this after finishing okay -->
<ok to="deleteOutputPathForActionSet" />
<!-- This is the initial code <ok to="deleteOutputPathForActionSet" /> -->
<!-- Go there if we have an error -->
<error to="map-scores-fail" />
</action>
<action name="deleteOutputPathForActionSet">
<fs>
<delete path="${actionSetOutputPath}/results/"/>
<fs>
<delete path="${actionSetOutputPath}/results/"/>
<delete path="${actionSetOutputPath}/projects/"/>
<mkdir path="${actionSetOutputPath}/results/"/>
<mkdir path="${actionSetOutputPath}/projects/"/>
</fs>
<ok to="createActionSetForResults"/>
<error to="actionset-delete-fail"/>
</action>
<ok to="createActionSetForResults"/>
<error to="actionset-delete-fail"/>
</action>
<action name="createActionSetForResults">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Produces the atomic action with the bip finder scores for publications</name>
<class>eu.dnetlib.dhp.actionmanager.bipfinder.SparkAtomicActionScoreJob</class>
<jar>dhp-aggregation-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkNormalExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkNormalDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
</spark-opts>
<arg>--inputPath</arg><arg>${bipScorePath}</arg>
<arg>--outputPath</arg><arg>${actionSetOutputPath}/results/</arg>
<action name="createActionSetForResults">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Produces the atomic action with the bip finder scores for publications</name>
<class>eu.dnetlib.dhp.actionmanager.bipfinder.SparkAtomicActionScoreJob</class>
<jar>dhp-aggregation-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkNormalExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkNormalDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
</spark-opts>
<arg>--inputPath</arg><arg>${bipScorePath}</arg>
<arg>--outputPath</arg><arg>${actionSetOutputPath}/results/</arg>
<arg>--targetEntity</arg><arg>result</arg>
</spark>
<ok to="project-impact-indicators"/>
<error to="actionset-creation-fail"/>
</action>
<ok to="project-impact-indicators"/>
<error to="actionset-creation-fail"/>
</action>
<action name="project-impact-indicators">
<!-- This is required as a tag for spark jobs, regardless of programming language -->