michele.artini
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

updated baseline workflow

This commit is contained in:
Sandro La Bruzzo 2021-07-08 11:11:41 +02:00
parent 0799ac9fb6
commit 8a034e46e1
3 changed files with 16 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/sx/graph/ebi/SparkCreateBaselineDataFrame.scala
View File

@ -45,6 +45,11 @@ object SparkCreateBaselineDataFrame {
parser.parseArgument(args)
val isLookupUrl: String = parser.get("isLookupUrl")
log.info("isLookupUrl: {}", isLookupUrl)
val workingPath = parser.get("workingPath")
log.info("workingPath: {}", workingPath)
val targetPath = parser.get("targetPath")
log.info("targetPath: {}", targetPath)
val isLookupService = ISLookupClientFactory.getLookUpService(isLookupUrl)
val vocabularies = VocabularyGroup.loadVocsFromIS(isLookupService)
@ -59,7 +64,7 @@ object SparkCreateBaselineDataFrame {
val sc = spark.sparkContext
val workingPath = parser.get("workingPath")
implicit val PMEncoder: Encoder[PMArticle] = Encoders.kryo(classOf[PMArticle])
implicit val PMJEncoder: Encoder[PMJournal] = Encoders.kryo(classOf[PMJournal])
@ -81,6 +86,8 @@ object SparkCreateBaselineDataFrame {
exported_dataset
.map(a => PubMedToOaf.convert(a, vocabularies)).as[Result]
.filter(p => p!= null)
.write.mode(SaveMode.Overwrite).save(s"$workingPath/oaf/baseline_oaf")
.write.mode(SaveMode.Overwrite).save(targetPath)
//s"$workingPath/oaf/baseline_oaf"
}
}

3
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/sx/graph/ebi/baseline_to_oaf_params.json
View File

@ -1,5 +1,6 @@
[
{"paramName":"mt", "paramLongName":"master", "paramDescription": "should be local or yarn", "paramRequired": true},
{"paramName":"i", "paramLongName":"isLookupUrl","paramDescription": "isLookupUrl", "paramRequired": true},
{"paramName":"w", "paramLongName":"workingPath","paramDescription": "the path of the sequencial file to read", "paramRequired": true}
{"paramName":"w", "paramLongName":"workingPath","paramDescription": "the path of the sequencial file to read", "paramRequired": true},
{"paramName":"t", "paramLongName":"targetPath","paramDescription": "the oaf path ", "paramRequired": true}
]

5
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/sx/graph/pubmed/oozie_app/workflow.xml
View File

@ -4,6 +4,10 @@
<name>baselineWorkingPath</name>
<description>the Baseline Working Path</description>
</property>
<property>
<name>targetPath</name>
<description>the Target Path</description>
</property>
<property>
<name>isLookupUrl</name>
<description>The IS lookUp service endopoint</description>
@ -34,6 +38,7 @@
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts>
<arg>--workingPath</arg><arg>${baselineWorkingPath}</arg>
<arg>--targetPath</arg><arg>${targetPath}</arg>
<arg>--master</arg><arg>yarn</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
</spark>