refactor code for EBI from dhp-graph-mapper into dhp-aggregation

This commit is contained in:
Sandro La Bruzzo 2021-10-14 14:23:05 +02:00
parent dd568ec88b
commit 51a03c0a50
7 changed files with 109 additions and 238 deletions

View File

@ -66,7 +66,7 @@ object SparkDownloadEBILinks {
val log: Logger = LoggerFactory.getLogger(getClass) val log: Logger = LoggerFactory.getLogger(getClass)
val MAX_ITEM_PER_PARTITION = 20000 val MAX_ITEM_PER_PARTITION = 20000
val conf: SparkConf = new SparkConf() val conf: SparkConf = new SparkConf()
val parser = new ArgumentApplicationParser(IOUtils.toString(getClass.getResourceAsStream("/eu/dnetlib/dhp/sx/graph/ebi/ebi_download_update.json"))) val parser = new ArgumentApplicationParser(IOUtils.toString(getClass.getResourceAsStream("/eu/dnetlib/dhp/sx/bio/ebi/ebi_download_update.json")))
parser.parseArgument(args) parser.parseArgument(args)
val spark: SparkSession = val spark: SparkSession =
SparkSession SparkSession
@ -86,7 +86,7 @@ object SparkDownloadEBILinks {
val workingPath = parser.get("workingPath") val workingPath = parser.get("workingPath")
log.info(s"workingPath -> $workingPath") log.info(s"workingPath -> $workingPath")
log.info("Getting max pubmedId where the links have been requested") log.info("Getting max pubmedId where the links have already requested")
val links: Dataset[EBILinkItem] = spark.read.load(s"$sourcePath/ebi_links_dataset").as[EBILinkItem] val links: Dataset[EBILinkItem] = spark.read.load(s"$sourcePath/ebi_links_dataset").as[EBILinkItem]
val lastPMIDRequested = links.map(l => l.id).select(max("value")).first.getLong(0) val lastPMIDRequested = links.map(l => l.id).select(max("value")).first.getLong(0)

View File

@ -14,7 +14,7 @@ object SparkEBILinksToOaf {
def main(args: Array[String]): Unit = { def main(args: Array[String]): Unit = {
val log: Logger = LoggerFactory.getLogger(getClass) val log: Logger = LoggerFactory.getLogger(getClass)
val conf: SparkConf = new SparkConf() val conf: SparkConf = new SparkConf()
val parser = new ArgumentApplicationParser(IOUtils.toString(getClass.getResourceAsStream("/eu/dnetlib/dhp/sx/graph/ebi/ebi_to_df_params.json"))) val parser = new ArgumentApplicationParser(IOUtils.toString(getClass.getResourceAsStream("/eu/dnetlib/dhp/sx/bio/ebi/ebi_to_df_params.json")))
parser.parseArgument(args) parser.parseArgument(args)
val spark: SparkSession = val spark: SparkSession =
SparkSession SparkSession
@ -31,7 +31,7 @@ object SparkEBILinksToOaf {
log.info(s"targetPath -> $targetPath") log.info(s"targetPath -> $targetPath")
implicit val PMEncoder: Encoder[Oaf] = Encoders.kryo(classOf[Oaf]) implicit val PMEncoder: Encoder[Oaf] = Encoders.kryo(classOf[Oaf])
val ebLinks: Dataset[EBILinkItem] = spark.read.load(s"${sourcePath}_dataset").as[EBILinkItem].filter(l => l.links != null) val ebLinks: Dataset[EBILinkItem] = spark.read.load(sourcePath).as[EBILinkItem].filter(l => l.links != null && l.links.startsWith("{"))
ebLinks.flatMap(j => BioDBToOAF.parse_ebi_links(j.links)) ebLinks.flatMap(j => BioDBToOAF.parse_ebi_links(j.links))
.filter(p => BioDBToOAF.EBITargetLinksFilter(p)) .filter(p => BioDBToOAF.EBITargetLinksFilter(p))

View File

@ -0,0 +1,105 @@
<workflow-app name="Create EBI Dataset" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>sourcePath</name>
<description>the Working Path</description>
</property>
<property>
<name>workingPath</name>
<description>the Working Path</description>
</property>
<property>
<name>targetPath</name>
<description>the OAF MDStore 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>resumeFrom</name>
<value>DownloadEBILinks</value>
<description>node to start</description>
</property>
</parameters>
<start to="resume_from"/>
<decision name="resume_from">
<switch>
<case to="DownloadEBILinks">${wf:conf('resumeFrom') eq 'DownloadEBILinks'}</case>
<case to="CreateEBIDataSet">${wf:conf('resumeFrom') eq 'CreateEBIDataSet'}</case>
<default to="DownloadEBILinks"/>
</switch>
</decision>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<action name="DownloadEBILinks">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>Incremental Download EBI Links</name>
<class>eu.dnetllib.dhp.sx.bio.ebi.SparkDownloadEBILinks</class>
<jar>dhp-aggregation-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.shuffle.partitions=2000
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts>
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--master</arg><arg>yarn</arg>
</spark>
<ok to="OverrideFolders"/>
<error to="Kill"/>
</action>
<action name="OverrideFolders">
<fs>
<delete path="${sourcePath}/ebi_links_dataset_old"/>
<move source="${sourcePath}/ebi_links_dataset" target="${sourcePath}/ebi_links_dataset_old"/>
<move source="${workingPath}/links_final" target="${sourcePath}/ebi_links_dataset"/>
</fs>
<ok to="CreateEBIDataSet"/>
<error to="Kill"/>
</action>
<action name="CreateEBIDataSet">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>Create OAF DataSet</name>
<class>eu.dnetllib.dhp.sx.bio.ebi.SparkEBILinksToOaf</class>
<jar>dhp-aggregation-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.sql.shuffle.partitions=2000
${sparkExtraOPT}
</spark-opts>
<arg>--sourcePath</arg><arg>${sourcePath}/ebi_links_dataset</arg>
<arg>--targetPath</arg><arg>${targetPath}</arg>
<arg>--master</arg><arg>yarn</arg>
</spark>
<ok to="End"/>
<error to="Kill"/>
</action>
<end name="End"/>
</workflow-app>

View File

@ -1,99 +0,0 @@
<workflow-app name="Create EBI Dataset" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>workingPath</name>
<description>the Working 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>
</parameters>
<start to="CreateEBIDataSet"/>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<action name="GenerateBaselineDataset">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>Create Baselnie DataSet</name>
<class>eu.dnetlib.dhp.sx.ebi.SparkCreateBaselineDataFrame</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=1
--driver-memory=${sparkDriverMemory}
--executor-cores=${sparkExecutorCores}
${sparkExtraOPT}
</spark-opts>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--master</arg><arg>yarn</arg>
<arg>--hdfsServerUri</arg><arg>${nameNode}</arg>
</spark>
<ok to="End"/>
<error to="Kill"/>
</action>
<action name="CreateEBIDataSet">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>Create EBI DataSet</name>
<class>eu.dnetlib.dhp.sx.ebi.SparkEBILinksToOaf</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.sql.shuffle.partitions=1000
${sparkExtraOPT}
</spark-opts>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--master</arg><arg>yarn</arg>
</spark>
<ok to="GenerateUpdates"/>
<error to="Kill"/>
</action>
<action name="GenerateUpdates">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>Create Baseline DataSet</name>
<class>eu.dnetlib.dhp.sx.ebi.SparkAddLinkUpdates</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=1
--driver-memory=${sparkDriverMemory}
--executor-cores=${sparkExecutorCores}
${sparkExtraOPT}
</spark-opts>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--master</arg><arg>yarn</arg>
</spark>
<ok to="End"/>
<error to="Kill"/>
</action>
<end name="End"/>
</workflow-app>

View File

@ -1,68 +0,0 @@
<configuration>
<!-- OCEAN -->
<property>
<name>jobTracker</name>
<value>yarnRM</value>
</property>
<property>
<name>nameNode</name>
<value>hdfs://nameservice1</value>
</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>
<!-- GARR -->
<!-- <property>-->
<!-- <name>jobTracker</name>-->
<!-- <value>yarn</value>-->
<!-- </property>-->
<!-- <property>-->
<!-- <name>nameNode</name>-->
<!-- <value>hdfs://hadoop-rm1.garr-pa1.d4science.org:8020</value>-->
<!-- </property>-->
<!-- <property>-->
<!-- <name>hive_metastore_uris</name>-->
<!-- <value>thrift://hadoop-edge3.garr-pa1.d4science.org:9083</value>-->
<!-- </property>-->
<!-- <property>-->
<!-- <name>spark2YarnHistoryServerAddress</name>-->
<!-- <value>http://hadoop-rm2.garr-pa1.d4science.org:19888</value>-->
<!-- </property>-->
<property>
<name>oozie.launcher.mapreduce.user.classpath.first</name>
<value>true</value>
</property>
<property>
<name>oozie.use.system.libpath</name>
<value>true</value>
</property>
<property>
<name>oozie.action.sharelib.for.spark</name>
<value>spark2</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>
</configuration>

View File

@ -1,67 +0,0 @@
<workflow-app name="Create EBI Dataset" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>sourcePath</name>
<description>the Working Path</description>
</property>
<property>
<name>workingPath</name>
<description>the Working 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>
</parameters>
<start to="DownloadEBILinks"/>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<action name="DownloadEBILinks">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>Incremental Download EBI Links</name>
<class>eu.dnetlib.dhp.sx.graph.ebi.SparkDownloadEBILinks</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.shuffle.partitions=2000
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts>
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--master</arg><arg>yarn</arg>
</spark>
<ok to="OverrideFolders"/>
<error to="Kill"/>
</action>
<action name="OverrideFolders">
<fs>
<delete path="${sourcePath}/ebi_links_dataset_old"/>
<move source="${sourcePath}/ebi_links_dataset" target="${sourcePath}/ebi_links_dataset_old"/>
<move source="${workingPath}/links_final" target="${sourcePath}/ebi_links_dataset"/>
</fs>
<ok to="End"/>
<error to="Kill"/>
</action>
<end name="End"/>
</workflow-app>