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

cleanup: mapping contents in the graph already defined in the OAF graph model doesn't require to be aware of the vocabularies

This commit is contained in:
Claudio Atzori 2021-10-20 14:04:45 +02:00
parent c01dd0c925
commit 00b78b9c58
3 changed files with 1 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/CopyHdfsOafApplication.java
View File

@ -67,23 +67,16 @@ public class CopyHdfsOafApplication extends AbstractMigrationApplication {
final String hdfsPath = parser.get("hdfsPath");
log.info("hdfsPath: {}", hdfsPath);
final String isLookupUrl = parser.get("isLookupUrl");
log.info("isLookupUrl: {}", isLookupUrl);
final ISLookUpService isLookupService = ISLookupClientFactory.getLookUpService(isLookupUrl);
final VocabularyGroup vocs = VocabularyGroup.loadVocsFromIS(isLookupService);
final Set<String> paths = mdstorePaths(mdstoreManagerUrl, mdFormat, mdLayout, mdInterpretation);
final SparkConf conf = new SparkConf();
conf.set("spark.serializer", "org.apache.spark.serializer.KryoSerializer");
conf.registerKryoClasses(ModelSupport.getOafModelClasses());
runWithSparkSession(conf, isSparkSessionManaged, spark -> processPaths(spark, vocs, hdfsPath, paths));
runWithSparkSession(conf, isSparkSessionManaged, spark -> processPaths(spark, hdfsPath, paths));
}
public static void processPaths(final SparkSession spark,
final VocabularyGroup vocs,
final String outputPath,
final Set<String> paths) {

6
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/copy_hdfs_oaf_parameters.json
View File

@ -28,11 +28,5 @@
"paramLongName": "mdInterpretation",
"paramDescription": "metadata interpretation",
"paramRequired": true
},
{
"paramName": "isu",
"paramLongName": "isLookupUrl",
"paramDescription": "the url of the ISLookupService",
"paramRequired": true
}
]

1
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/raw_all/oozie_app/workflow.xml
View File

@ -569,7 +569,6 @@
<arg>--mdFormat</arg><arg>OAF</arg>
<arg>--mdLayout</arg><arg>store</arg>
<arg>--mdInterpretation</arg><arg>graph</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
</spark>
<ok to="wait_graphs"/>
<error to="Kill"/>