UBO-382 Allow to set preconfigured classification values during scopu…

…s import
MyCoRe-Org · Dec 11, 2024 · 0184ad4 · 0184ad4
1 parent 55e9530
commit 0184ad4
Show file tree

Hide file tree

Showing 2 changed files with 33 additions and 1 deletion.
diff --git a/ubo-common/src/main/java/org/mycore/ubo/importer/scopus/ScopusImporter.java b/ubo-common/src/main/java/org/mycore/ubo/importer/scopus/ScopusImporter.java
@@ -8,6 +8,8 @@
 import org.jdom2.Document;
 import org.jdom2.Element;
 import org.jdom2.filter.ElementFilter;
+import org.jdom2.filter.Filters;
+import org.jdom2.xpath.XPathExpression;
 import org.mycore.access.MCRAccessException;
 import org.mycore.common.MCRConstants;
 import org.mycore.common.MCRException;
@@ -16,9 +18,11 @@
 import org.mycore.common.MCRSessionMgr;
 import org.mycore.common.config.MCRConfiguration2;
 import org.mycore.common.xml.MCRURIResolver;
+import org.mycore.common.xml.MCRXMLFunctions;
 import org.mycore.datamodel.metadata.MCRMetadataManager;
 import org.mycore.datamodel.metadata.MCRObject;
 import org.mycore.datamodel.metadata.MCRObjectID;
+import org.mycore.frontend.MCRFrontendUtil;
 import org.mycore.mods.MCRMODSWrapper;
 import org.mycore.solr.MCRSolrClientFactory;
 import org.mycore.solr.MCRSolrUtils;
@@ -31,6 +35,9 @@
 import java.util.List;
 import java.util.Locale;
 
+import static org.mycore.common.MCRConstants.MODS_NAMESPACE;
+import static org.mycore.common.MCRConstants.XPATH_FACTORY;
+
 class ScopusImporter {
 
     private static final Logger LOGGER = LogManager.getLogger();
@@ -49,6 +56,8 @@ class ScopusImporter {
 
     private static String MAIL_XSL;
 
+    private static XPathExpression<Element> MODS_XPATH = XPATH_FACTORY
+        .compile(".//mods:mods", Filters.element(), null, MODS_NAMESPACE);
     private List<MCRObject> importedObjects = new ArrayList<>();
 
     static {
@@ -57,7 +66,6 @@ class ScopusImporter {
         IMPORT_URI = MCRConfiguration2.getString(prefix + "ImportURI").get();
         PROJECT_ID = MCRConfiguration2.getString(prefix + "ProjectID").get();
         STATUS = MCRConfiguration2.getString(prefix + "Status").get();
-
         prefix += "Mail.";
         MAIL_TO = MCRConfiguration2.getString(prefix + "To").get();
         MAIL_PARAM = MCRConfiguration2.getString(prefix + "Param").get();
@@ -110,6 +118,8 @@ private static boolean shouldIgnore(Element publication) {
     private final static SimpleDateFormat ID_BUILDER = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss", Locale.ROOT);
 
     private MCRObject buildMCRObject(Element publicationXML) {
+        setPreconfiguredClasses(publicationXML);
+
         MCRObject obj = new MCRObject(new Document(publicationXML));
         MCRMODSWrapper wrapper = new MCRMODSWrapper(obj);
         wrapper.setServiceFlag("status", STATUS);
@@ -119,6 +129,27 @@ private MCRObject buildMCRObject(Element publicationXML) {
         return obj;
     }
 
+    private void setPreconfiguredClasses(Element publicationXML) {
+        List<String> list = MCRConfiguration2.getString("UBO.Scopus.Importer.Default.Classifications")
+            .stream()
+            .flatMap(MCRConfiguration2::splitValue)
+            .toList();
+
+        list.forEach(classification -> {
+            String[] parts = classification.split(":");
+            if (parts.length == 2 && MCRXMLFunctions.isCategoryID(parts[0], parts[1])) {
+                MODS_XPATH.evaluate(publicationXML).forEach(element -> {
+                    Element c = new Element("classification", MODS_NAMESPACE);
+                    c.setAttribute("valueURI",
+                        MCRFrontendUtil.getBaseURL() + "classifications/" + parts[0] + "#" + parts[1]);
+                    c.setAttribute("authorityURI",
+                        MCRFrontendUtil.getBaseURL() + "classifications/" + parts[0]);
+                    publicationXML.addContent(c);
+                });
+            }
+        });
+    }
+
     private String getImportID() {
         return ID_BUILDER.format(new Date(MCRSessionMgr.getCurrentSession().getLoginTime()));
     }

diff --git a/ubo-common/src/main/resources/config/ubo-common/mycore.properties b/ubo-common/src/main/resources/config/ubo-common/mycore.properties
@@ -686,6 +686,7 @@ UBO.Scopus.API.URL=https://api.elsevier.com/content/
 UBO.Scopus.Importer.Field2QueryID=id_scopus
 UBO.Scopus.Importer.ImportURI.Scopus=%UBO.Scopus.API.URL%abstract/scopus_id/{0}?apikey=%UBO.Scopus.API.Key%
 UBO.Scopus.Importer.ImportURI=xslStyle:mods2mycoreobject:enrich:scopus-import:xslStyle:import/scopus2mods,import/genre2genre,%UBO.SupportedMODSFilterName%:ratelimit:ScopusWeek:ratelimit:ScopusSecond:%UBO.Scopus.Importer.ImportURI.Scopus%
+UBO.Scopus.Importer.Default.Classifications=
 UBO.Scopus.Importer.ProjectID=%UBO.projectid.default%
 UBO.Scopus.Importer.Status=imported
 UBO.Scopus.Importer.Mail.To=%MCR.Mail.Address%