D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 13 Projects 2 Releases 2 Wiki Activity

master #11

Manually merged
claudio.atzori merged 275 commits from :master into enrichment_wfs 2020-05-11 15:14:56 +02:00
Conversation 1 Commits 275 Files Changed 156 -54
1 changed files with 0 additions and 54 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files
Showing only changes of commit 026f297e49 - Show all commits

54
dhp-workflows/dhp-propagation/src/main/java/eu/dnetlib/dhp/resulttoorganizationfrominstrepo/SparkResultToOrganizationFromIstRepoJob2.java
View File

@ -61,12 +61,6 @@ public class SparkResultToOrganizationFromIstRepoJob2 {
final String resultType = resultClassName.substring(resultClassName.lastIndexOf(".") + 1).toLowerCase();
log.info("resultType: {}", resultType);
final Boolean writeUpdates = Optional
.ofNullable(parser.get("writeUpdate"))
.map(Boolean::valueOf)
.orElse(Boolean.TRUE);
log.info("writeUpdate: {}", writeUpdates);
final Boolean saveGraph = Optional
.ofNullable(parser.get("saveGraph"))
.map(Boolean::valueOf)
@ -93,7 +87,6 @@ public class SparkResultToOrganizationFromIstRepoJob2 {
outputPath,
resultClazz,
resultType,
writeUpdates,
saveGraph);
});
}
@ -106,7 +99,6 @@ public class SparkResultToOrganizationFromIstRepoJob2 {
String outputPath,
Class<? extends Result> resultClazz,
String resultType,
Boolean writeUpdates,
Boolean saveGraph) {
final JavaSparkContext sc = new JavaSparkContext(spark.sparkContext());
@ -124,10 +116,6 @@ public class SparkResultToOrganizationFromIstRepoJob2 {
broadcast_datasourceorganizationassoc)
.as(Encoders.bean(ResultOrganizationSet.class));
if (writeUpdates) {
createUpdateForRelationWrite(potentialUpdates, outputPath + "/" + resultType);
}
if (saveGraph) {
getNewRelations(
spark
@ -229,48 +217,6 @@ public class SparkResultToOrganizationFromIstRepoJob2 {
Encoders.bean(DatasourceOrganization.class));
}
private static void createUpdateForRelationWrite(
Dataset<ResultOrganizationSet> toupdaterelation, String outputPath) {
toupdaterelation
.flatMap(
s -> {
List<Relation> relationList = new ArrayList<>();
List<String> orgs = s.getOrganizationSet();
String resId = s.getResultId();
for (String org : orgs) {
relationList
.add(
getRelation(
org,
resId,
RELATION_ORGANIZATION_RESULT_REL_CLASS,
RELATION_RESULTORGANIZATION_REL_TYPE,
RELATION_RESULTORGANIZATION_SUBREL_TYPE,
PROPAGATION_DATA_INFO_TYPE,
PROPAGATION_RELATION_RESULT_ORGANIZATION_INST_REPO_CLASS_ID,
PROPAGATION_RELATION_RESULT_ORGANIZATION_INST_REPO_CLASS_NAME));
relationList
.add(
getRelation(
resId,
org,
RELATION_RESULT_ORGANIZATION_REL_CLASS,
RELATION_RESULTORGANIZATION_REL_TYPE,
RELATION_RESULTORGANIZATION_SUBREL_TYPE,
PROPAGATION_DATA_INFO_TYPE,
PROPAGATION_RELATION_RESULT_ORGANIZATION_INST_REPO_CLASS_ID,
PROPAGATION_RELATION_RESULT_ORGANIZATION_INST_REPO_CLASS_NAME));
}
return relationList.iterator();
},
Encoders.bean(Relation.class))
.toJSON()
.write()
.mode(SaveMode.Overwrite)
.option("compression", "gzip")
.text(outputPath);
}
private static org.apache.spark.sql.Dataset<ResultOrganizationSet> organizationPropagationAssoc(
SparkSession spark,
Broadcast<org.apache.spark.sql.Dataset<DatasourceOrganization>> broadcast_datasourceorganizationassoc) {