jsonHarvester: extract uuid from identifier

https://data-atmo-hdf.opendata.arcgis.com/datasets/bac17d7d05a34242a8b22c535ecdb13d will extract bac17d7d05a34242a8b22c535ecdb13d
geonetwork · Sep 9, 2021 · 332ff96 · 332ff96
1 parent c42d4ca
commit 332ff96
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 4 deletions.
diff --git a/harvesters/src/main/java/org/fao/geonet/kernel/harvest/harvester/simpleUrl/Harvester.java b/harvesters/src/main/java/org/fao/geonet/kernel/harvest/harvester/simpleUrl/Harvester.java
@@ -136,8 +136,9 @@ public HarvestResult harvest(Logger log) throws Exception {
                         log.debug("Number of records in response: " + nodes.size());
 
                         nodes.forEach(record -> {
-                            Element xml = convertRecordToXml(record);
-                            uuids.put(record.get(params.recordIdPath).asText(), xml);
+                            String uuid = this.extractUuidFromIdentifier(record.get(params.recordIdPath).asText());
+                            Element xml = convertRecordToXml(record, uuid);
+                            uuids.put(uuid, xml);
                         });
                         aligner.align(uuids, errors);
                         allUuids.putAll(uuids);
@@ -168,6 +169,14 @@ public HarvestResult harvest(Logger log) throws Exception {
         return result;
     }
 
+    private String extractUuidFromIdentifier(final String identifier ) {
+        String uuid = identifier;
+        if (Lib.net.isUrlValid(uuid)) {
+            uuid = uuid.replaceFirst(".*/([^/?]+).*", "$1");
+        }
+        return uuid;
+    }
+
     @VisibleForTesting
     protected List<String> buildListOfUrl(SimpleUrlParams params, int numberOfRecordsToHarvest) {
         List<String> urlList = new ArrayList<String>();
@@ -217,14 +226,15 @@ protected List<String> buildListOfUrl(SimpleUrlParams params, int numberOfRecord
         return urlList;
     }
 
-    private Element convertRecordToXml(JsonNode record) {
+    private Element convertRecordToXml(JsonNode record, String uuid) {
         ObjectMapper objectMapper = new ObjectMapper();
         try {
             String recordAsXml = XML.toString(
                 new JSONObject(
                     objectMapper.writeValueAsString(record)), "record");
             recordAsXml = Xml.stripNonValidXMLCharacters(recordAsXml).replace("<@", "<").replace("</@", "</");
             Element recordAsElement = Xml.loadString(recordAsXml, false);
+            recordAsElement.addContent(new Element("uuid").setText(uuid));
             Path importXsl = context.getAppPath().resolve(Geonet.Path.IMPORT_STYLESHEETS);
             final Path xslPath = importXsl.resolve(params.toISOConversion + ".xsl");
             return Xml.transform(recordAsElement, xslPath);

diff --git a/schemas/iso19115-3.2018/src/main/plugin/iso19115-3.2018/convert/fromJsonLdEsri.xsl b/schemas/iso19115-3.2018/src/main/plugin/iso19115-3.2018/convert/fromJsonLdEsri.xsl
@@ -70,7 +70,7 @@
           <mcc:MD_Identifier>
             <mcc:code>
               <gco:CharacterString>
-                <xsl:value-of select="identifier"/>
+                <xsl:value-of select="uuid"/>
               </gco:CharacterString>
             </mcc:code>
           </mcc:MD_Identifier>