From fa42026590cad475e786eae85d568954b30768db Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Tue, 27 Apr 2021 10:10:06 +0200
Subject: [PATCH] fixed PersonCleaner extension functions

---
 .../transformation/TransformSparkJobNode.java |  2 +-
 .../transformation/xslt/PersonCleaner.java    | 36 ++++++++-----------
 ...enaire_datacite_ExchangeLandingpagePid.xsl | 12 -------
 3 files changed, 15 insertions(+), 35 deletions(-)

diff --git a/dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/transformation/TransformSparkJobNode.java b/dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/transformation/TransformSparkJobNode.java
index a4ddd367b..5ef26b7d5 100644
--- a/dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/transformation/TransformSparkJobNode.java
+++ b/dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/transformation/TransformSparkJobNode.java
@@ -156,7 +156,7 @@ public class TransformSparkJobNode {
 	 * @return
 	 */
 	private static int getRepartitionNumber(long totalInput, Integer rpt) {
-		return (int) (totalInput / rpt);
+		return Math.max(1, (int) (totalInput / rpt));
 	}
 
 }
diff --git a/dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/transformation/xslt/PersonCleaner.java b/dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/transformation/xslt/PersonCleaner.java
index 069060722..c4cf345f0 100644
--- a/dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/transformation/xslt/PersonCleaner.java
+++ b/dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/transformation/xslt/PersonCleaner.java
@@ -4,7 +4,6 @@ package eu.dnetlib.dhp.transformation.xslt;
 import static eu.dnetlib.dhp.transformation.xslt.XSLTTransformationFunction.QNAME_BASE_URI;
 
 import java.io.Serializable;
-// import java.nio.charset.Charset;
 import java.nio.charset.StandardCharsets;
 import java.text.Normalizer;
 import java.util.List;
@@ -18,22 +17,10 @@ import com.google.common.hash.Hashing;
 
 import eu.dnetlib.dhp.transformation.xslt.utils.Capitalize;
 import eu.dnetlib.dhp.transformation.xslt.utils.DotAbbreviations;
-import net.sf.saxon.s9api.ExtensionFunction;
-import net.sf.saxon.s9api.ItemType;
-import net.sf.saxon.s9api.OccurrenceIndicator;
-import net.sf.saxon.s9api.QName;
-import net.sf.saxon.s9api.SaxonApiException;
-import net.sf.saxon.s9api.SequenceType;
-import net.sf.saxon.s9api.XdmValue;
-
-//import eu.dnetlib.pace.clustering.NGramUtils;
-//import eu.dnetlib.pace.util.Capitalise;
-//import eu.dnetlib.pace.util.DotAbbreviations;
+import net.sf.saxon.s9api.*;
 
 public class PersonCleaner implements ExtensionFunction, Serializable {
-	/**
-	 *
-	 */
+
 	private static final long serialVersionUID = 1L;
 	private List<String> firstname = Lists.newArrayList();
 	private List<String> surname = Lists.newArrayList();
@@ -45,7 +32,7 @@ public class PersonCleaner implements ExtensionFunction, Serializable {
 
 	}
 
-	public String normalize(String s) {
+	private String normalize(String s) {
 		s = Normalizer.normalize(s, Normalizer.Form.NFD); // was NFD
 		s = s.replaceAll("\\(.+\\)", "");
 		s = s.replaceAll("\\[.+\\]", "");
@@ -184,7 +171,7 @@ public class PersonCleaner implements ExtensionFunction, Serializable {
 
 	@Override
 	public QName getName() {
-		return new QName(QNAME_BASE_URI + "/person", "person");
+		return new QName(QNAME_BASE_URI + "/person", "normalize");
 	}
 
 	@Override
@@ -194,13 +181,18 @@ public class PersonCleaner implements ExtensionFunction, Serializable {
 
 	@Override
 	public SequenceType[] getArgumentTypes() {
-		// TODO Auto-generated method stub
-		return null;
+		return new SequenceType[] {
+			SequenceType.makeSequenceType(ItemType.STRING, OccurrenceIndicator.ZERO_OR_ONE)
+		};
 	}
 
 	@Override
-	public XdmValue call(XdmValue[] arguments) throws SaxonApiException {
-		// TODO Auto-generated method stub
-		return null;
+	public XdmValue call(XdmValue[] xdmValues) throws SaxonApiException {
+		XdmValue r = xdmValues[0];
+		if (r.size() == 0) {
+			return new XdmAtomicValue("");
+		}
+		final String currentValue = xdmValues[0].itemAt(0).getStringValue();
+		return new XdmAtomicValue(normalize(currentValue));
 	}
 }
diff --git a/dhp-workflows/dhp-aggregation/src/test/resources/eu/dnetlib/dhp/transform/scripts/xslt_cleaning_oaiOpenaire_datacite_ExchangeLandingpagePid.xsl b/dhp-workflows/dhp-aggregation/src/test/resources/eu/dnetlib/dhp/transform/scripts/xslt_cleaning_oaiOpenaire_datacite_ExchangeLandingpagePid.xsl
index 56451505e..46fcfd7cb 100644
--- a/dhp-workflows/dhp-aggregation/src/test/resources/eu/dnetlib/dhp/transform/scripts/xslt_cleaning_oaiOpenaire_datacite_ExchangeLandingpagePid.xsl
+++ b/dhp-workflows/dhp-aggregation/src/test/resources/eu/dnetlib/dhp/transform/scripts/xslt_cleaning_oaiOpenaire_datacite_ExchangeLandingpagePid.xsl
@@ -68,12 +68,6 @@
         <xsl:call-template name="validRecord" />
     </xsl:template>
 
-    <xsl:template name="terminate">
-        <xsl:message terminate="yes">
-            record is not compliant, transformation is interrupted.
-        </xsl:message>
-    </xsl:template>
-
     <xsl:template name="validRecord">
         <record>
             <xsl:apply-templates select="//*[local-name() = 'header']" />
@@ -282,9 +276,6 @@
                                 <xsl:value-of select="$varEmbargoEndDate"/>
                             </oaf:embargoenddate>
                         </xsl:when>
-                        <xsl:otherwise>
-                            <xsl:call-template name="terminate"/>
-                        </xsl:otherwise>
                     </xsl:choose>
                 </xsl:if>
 
@@ -310,9 +301,6 @@
                           </dr:CobjCategory>
                          -->
                     </xsl:when>
-                    <xsl:otherwise>
-                        <xsl:call-template name="terminate"/>
-                    </xsl:otherwise>
                 </xsl:choose>
 
                 <!-- review status -->