D-Net
/
dnet-hadoop
13
4
Fork 10
Code Issues 14 Pull Requests 8 Projects 2 Releases 1 Wiki Activity

additional XSLT transformation scripts, enhance methods #97

Closed
andreas.czerniak wants to merge 49 commits from hadoop_aggregator into hadoop_aggregator
pull from: hadoop_aggregator
merge into: D-Net:hadoop_aggregator
Conversation 3 Commits 49 Files Changed 16 +11 -13
4 changed files with 11 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files
Showing only changes of commit d91e21edd8 - Show all commits

3
dhp-common/src/main/java/eu/dnetlib/dhp/common/vocabulary/VocabularyGroup.java
View File

@ -1,3 +1,4 @@
package eu.dnetlib.dhp.common.vocabulary;
import java.io.Serializable;
@ -143,7 +144,7 @@ public class VocabularyGroup implements Serializable {
}
public boolean termExists(final String vocId, final String id, final Boolean caseSensitive) {
if(Boolean.TRUE.equals(caseSensitive)) {
if (Boolean.TRUE.equals(caseSensitive)) {
return vocabularyExists(vocId) && vocs.get(vocId).termExists(id);
}
return vocabularyExists(vocId) && vocs.get(vocId.toLowerCase()).termExists(id);

18
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/transformation/xslt/TransformationFunctionProxy.java
View File

@ -5,13 +5,12 @@ import java.util.LinkedList;
import java.util.List;
import java.util.Properties;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.schema.oaf.Qualifier;
import net.sf.saxon.s9api.*;
import scala.Serializable;
//import eu.dnetlib.data.collective.transformation.engine.functions.ProcessingException;
// import eu.dnetlib.data.collective.transformation.engine.functions.ProcessingException;
// import org.apache.kafka.clients.producer.KafkaProducer;
// import org.apache.kafka.clients.producer.Producer;
@ -77,12 +76,12 @@ public class TransformationFunctionProxy implements ExtensionFunction, Serializa
}
/**
* normalize values given as an input value by using a vocabulary
* normalize values given as an input value by using a vocabulary
* @param aInput - the value as a String
* @param aVocabularyName - the name of the vocabulary, which must be known for the vocabulary registry
* @return
*/
public synchronized String convertString(String aInput, String aVocabularyName){
public synchronized String convertString(String aInput, String aVocabularyName) {
List<String> values = new LinkedList<>();
// Producer<String, String> producer = new KafkaProducer<>(props);
@ -105,15 +104,12 @@ public class TransformationFunctionProxy implements ExtensionFunction, Serializa
// producer.send(new ProducerRecord<String, String>("transformation-vocab", "convert failed", aVocabularyName));
throw new IllegalStateException(e);
}
/* catch (KafkaException e) {
// For all other exceptions, just abort the transaction and try again.
producer.abortTransaction();
}
*/
/*
* catch (KafkaException e) { // For all other exceptions, just abort the transaction and try again.
* producer.abortTransaction(); }
*/
// producer.close();
}
}

1
dhp-workflows/dhp-aggregation/src/test/java/eu/dnetlib/dhp/transformation/TransformationJobTest.java
View File

@ -60,6 +60,7 @@ public class TransformationJobTest extends AbstractVocabularyTest {
// We Load the XSLT transformation Rule from the classpath
XSLTTransformationFunction tr = loadTransformationRule("/eu/dnetlib/dhp/transform/zenodo_tr.xslt");
// XSLTTransformationFunction tr = loadTransformationRule("/eu/dnetlib/dhp/transform/scripts/xslt_cleaning_datarepo_datacite.xsl");
MetadataRecord result = tr.call(mr);

2
dhp-workflows/dhp-aggregation/src/test/resources/eu/dnetlib/dhp/transform/scripts/xslt_cleaning_datarepo_datacite.xsl
View File

@ -15,7 +15,7 @@
<xsl:param name="varMESTD" select="'mestd_______::'" />
<xsl:param name="varMZOS" select="'irb_hr______::'" />
<xsl:param name="varNHMRC" select="'nhmrc_______::'" />
<xsl:pasram name="varNIH" select="'nih_________::'" />
<xsl:param name="varNIH" select="'nih_________::'" />
<xsl:param name="varNSF" select="'nsf_________::'" />
<xsl:param name="varNWO" select="'nwo_________::'" />
<xsl:param name="varRCUK" select="'rcuk________::'" />