Merge pull request #121 from europeana/EA-testingPageXML

Ea testing page xml
europeana · Jun 15, 2023 · 1c51870 · 1c51870
2 parents 4625ade + 291f503
commit 1c51870
Show file tree

Hide file tree

Showing 3 changed files with 30 additions and 18 deletions.
diff --git a/common/src/main/java/eu/europeana/fulltext/pageXML/PageXMLParser.java b/common/src/main/java/eu/europeana/fulltext/pageXML/PageXMLParser.java
@@ -3,33 +3,37 @@
 import eu.europeana.edm.media.MediaReference;
 import eu.europeana.fulltext.alto.model.AltoPage;
 import eu.europeana.fulltext.alto.parser.AltoParser;
+import eu.europeana.fulltext.exception.XmlParsingException;
 import org.xml.sax.InputSource;
 
 import javax.xml.XMLConstants;
 import javax.xml.transform.*;
 import javax.xml.transform.dom.DOMResult;
 import javax.xml.transform.dom.DOMSource;
 import javax.xml.transform.stream.StreamSource;
+import java.io.IOException;
 import java.io.InputStream;
+import java.net.URL;
 
 /**
  * @author Hugo
  * @since 4 Apr 2023
  */
 public class PageXMLParser extends AltoParser {
     private static final String XSLT_PATH = "etc/PageToAlto.xsl";
-    private final Transformer _transformer;
+    private final Transformer transformer;
 
-    public PageXMLParser() throws TransformerConfigurationException {
-        InputStream is = ClassLoader.getSystemClassLoader()
-                .getResourceAsStream(XSLT_PATH);
-        if (is == null) {
-            is = this.getClass().getResourceAsStream(XSLT_PATH);
+    public PageXMLParser() throws TransformerConfigurationException, IOException, XmlParsingException {
+        URL file = PageXMLParser.class.getClassLoader().getResource(XSLT_PATH);
+        if (file == null) {
+            throw new XmlParsingException("Unable to find file " +  XSLT_PATH);
+        }
+        try (InputStream is = file.openStream()) {
+            TransformerFactory tf = TransformerFactory.newInstance();
+            tf.setAttribute(XMLConstants.ACCESS_EXTERNAL_DTD, "");
+            tf.setAttribute(XMLConstants.ACCESS_EXTERNAL_STYLESHEET, "");
+            transformer = tf.newTransformer(new StreamSource(is));
         }
-        TransformerFactory tf = TransformerFactory.newInstance();
-        tf.setAttribute(XMLConstants.ACCESS_EXTERNAL_DTD, "");
-        tf.setAttribute(XMLConstants.ACCESS_EXTERNAL_STYLESHEET, "");
-        _transformer = tf.newTransformer(new StreamSource(is));
     }
 
     public AltoPage processPage(InputSource source, MediaReference ref) {
@@ -39,7 +43,7 @@ public AltoPage processPage(InputSource source, MediaReference ref) {
     public AltoPage processPage(Source source, MediaReference ref) {
         try {
             DOMResult result = new DOMResult();
-            _transformer.transform(source, result);
+            transformer.transform(source, result);
             return super.processPage(new DOMSource(result.getNode()), ref);
         } catch (TransformerException e) {
             throw new RuntimeException(e);

diff --git a/common/src/main/java/eu/europeana/fulltext/service/PageXmlFulltextConverter.java b/common/src/main/java/eu/europeana/fulltext/service/PageXmlFulltextConverter.java
@@ -13,6 +13,7 @@
 import javax.xml.transform.TransformerConfigurationException;
 import javax.xml.transform.stream.StreamSource;
 import java.io.ByteArrayInputStream;
+import java.io.IOException;
 import java.nio.charset.StandardCharsets;
 
 
@@ -31,9 +32,8 @@ public FullTextPackage convert(AnnotationPreview annotationPreview) throws Europ
             // page xml first converts the xslt into alto then we can convert alto to EDM
             return new AltoToFulltextConverter().getAltoToEDM(altoPage, annotationPreview, reference);
 
-        } catch (TransformerConfigurationException e) {
-            e.printStackTrace();
-            throw new XmlParsingException("Error configuring the transformer for type " +annotationPreview.getFulltextType().getMimeType());
+        } catch (TransformerConfigurationException | IOException e) {
+            throw new XmlParsingException("Error configuring the transformer for type " +annotationPreview.getFulltextType());
         }
 
     }

diff --git a/k8s/overlays/cloud/cronjob.yaml.template b/k8s/overlays/cloud/cronjob.yaml.template
@@ -10,6 +10,10 @@ spec:
     spec:
       ttlSecondsAfterFinished: ${CRON_TTL_AFTER_FINISHED}
       template:
+        metadata:
+          annotations:
+            fluentd/include: '${COLLECT_LOGS}'
+            fluentd/multiline: 'true'
         spec:
           containers:
             - name: fulltext-annosync
@@ -56,13 +60,13 @@ spec:
   suspend: ${SUSPEND_INDEXING}
   concurrencyPolicy: Forbid
   jobTemplate:
-    metadata:
-      annotations:
-        fluentd/include: '${COLLECT_LOGS}'
-        fluentd/multiline: 'true'    
     spec:
       ttlSecondsAfterFinished: ${CRON_TTL_AFTER_FINISHED}
       template:
+        metadata:
+          annotations:
+            fluentd/include: '${COLLECT_LOGS}'
+            fluentd/multiline: 'true'
         spec:
           containers:
             - name: fulltext-indexing
@@ -117,6 +121,10 @@ spec:
     spec:
       ttlSecondsAfterFinished: ${CRON_TTL_AFTER_FINISHED}
       template:
+        metadata:
+          annotations:
+            fluentd/include: '${COLLECT_LOGS}'
+            fluentd/multiline: 'true'
         spec:
           containers:
             - name: fulltext-indexing