update to newer schemas for pubmed/medline and wikipedia

lumongo · Jun 29, 2017 · 27d8c14 · 27d8c14
1 parent 99a822e
commit 27d8c14
Show file tree

Hide file tree

Showing 4 changed files with 460 additions and 192 deletions.
diff --git a/lumongo-example/src/main/java/org/lumongo/example/medline/MedlineJAXBReader.java b/lumongo-example/src/main/java/org/lumongo/example/medline/MedlineJAXBReader.java
@@ -7,6 +7,7 @@
 import org.lumongo.xml.StaxJAXBReader;
 
 import javax.xml.bind.JAXBException;
+import java.lang.Object;
 import java.text.DateFormatSymbols;
 import java.util.Arrays;
 import java.util.Date;
@@ -128,11 +129,11 @@ else if (name instanceof CollectiveName) {
 		document.setJournalIso(journal.getISOAbbreviation());
 
 		JournalIssue journalIssue = journal.getJournalIssue();
-		
-		String title = medlineCitation.getArticle().getArticleTitle();
+
+		ArticleTitle title = medlineCitation.getArticle().getArticleTitle();
 
 		if (title != null) {
-			document.setTitle(title);
+			document.setTitle(title.getvalue());
 		}
 
 

diff --git a/...src/main/schemas/mediawiki-export-0.8.xsd → ...rc/main/schemas/mediawiki-export-0.10.xsd b/...src/main/schemas/mediawiki-export-0.8.xsd → ...rc/main/schemas/mediawiki-export-0.10.xsd
@@ -28,16 +28,20 @@
 	Version 0.8 adds support for a <model> and a <format> tag for
 	each revision. See contenthandler.txt.
 
+	Version 0.9 adds the database name to the site information.
+
+	Version 0.10 moved the <model> and <format> tags before the <text> tag.
+
 	The canonical URL to the schema document is:
-	http://www.mediawiki.org/xml/export-0.8.xsd
+	http://www.mediawiki.org/xml/export-0.10.xsd
 
 	Use the namespace:
-	http://www.mediawiki.org/xml/export-0.8/
+	http://www.mediawiki.org/xml/export-0.10/
 -->
 <schema xmlns="http://www.w3.org/2001/XMLSchema"
-		xmlns:mw="http://www.mediawiki.org/xml/export-0.8/"
-		targetNamespace="http://www.mediawiki.org/xml/export-0.8/"
-		elementFormDefault="qualified">
+        xmlns:mw="http://www.mediawiki.org/xml/export-0.10/"
+        targetNamespace="http://www.mediawiki.org/xml/export-0.10/"
+        elementFormDefault="qualified">
 
 	<annotation>
 		<documentation xml:lang="en">
@@ -79,6 +83,7 @@
 	<complexType name="SiteInfoType">
 		<sequence>
 			<element name="sitename" type="string" minOccurs="0" />
+            <element name="dbname" type="string" minOccurs="0"/>
 			<element name="base" type="anyURI" minOccurs="0" />
 			<element name="generator" type="string" minOccurs="0" />
 			<element name="case" type="mw:CaseType" minOccurs="0" />
@@ -180,10 +185,10 @@
 			<element name="contributor" type="mw:ContributorType" />
 			<element name="minor" minOccurs="0" maxOccurs="1" />
 			<element name="comment" type="mw:CommentType" minOccurs="0" maxOccurs="1" />
-			<element name="text" type="mw:TextType" />
-			<element name="sha1" type="string" />
 			<element name="model" type="mw:ContentModelType" />
 			<element name="format" type="mw:ContentFormatType" />
+            <element name="text" type="mw:TextType"/>
+            <element name="sha1" type="string"/>
 		</sequence>
 	</complexType>
 

diff --git a/lumongo-example/src/main/schemas/nlmmedlinecitationset_130101.dtd b/lumongo-example/src/main/schemas/nlmmedlinecitationset_130101.dtd