Dataset link simplification - protect code if http status is null (fo…

…r example, a timeout) and check links from metadata url in capabilities start with http
geonetwork · Dec 17, 2024 · 1654361 · 1654361
1 parent de2ab95
commit 1654361
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 14 deletions.
diff --git a/.../java/net/geocat/database/linkchecker/service/CapabilitiesDatasetMetadataLinkService.java b/.../java/net/geocat/database/linkchecker/service/CapabilitiesDatasetMetadataLinkService.java
@@ -42,6 +42,7 @@
 import org.springframework.beans.factory.annotation.Autowired;
 import org.springframework.context.annotation.Scope;
 import org.springframework.stereotype.Component;
+import org.springframework.util.StringUtils;
 
 import java.util.ArrayList;
 import java.util.List;
@@ -57,19 +58,22 @@ public List<CapabilitiesDatasetMetadataLink> createCapabilitiesDatasetMetadataLi
         List<CapabilitiesDatasetMetadataLink> result = new ArrayList<>();
         for (DatasetLink link : doc.getDatasetLinksList()) {
 
-            CapabilitiesDatasetMetadataLink item = new CapabilitiesDatasetMetadataLink();
-            item.setOgcLayerName(link.getOgcLayerName());
-            item.setLinkState(LinkState.Created);
-            item.setRawURL(link.getRawUrl());
-            item.setLinkCheckJobId(cap.getLinkCheckJobId());
-            item.setFixedURL(datasetLinkFixer.fix(link.getRawUrl()));
-            item.setIdentity(link.getIdentifier());
-            item.setAuthority(link.getAuthority());
-            item.setAuthorityName(link.getAuthorityName());
-            item.setCapabilitiesDocument(cap);
+            // Create the link if the URL starts with http
+            if (StringUtils.hasLength(link.getRawUrl()) && link.getRawUrl().toLowerCase().startsWith("http")) {
+                CapabilitiesDatasetMetadataLink item = new CapabilitiesDatasetMetadataLink();
+                item.setOgcLayerName(link.getOgcLayerName());
+                item.setLinkState(LinkState.Created);
+                item.setRawURL(link.getRawUrl());
+                item.setLinkCheckJobId(cap.getLinkCheckJobId());
+                item.setFixedURL(datasetLinkFixer.fix(link.getRawUrl()));
+                item.setIdentity(link.getIdentifier());
+                item.setAuthority(link.getAuthority());
+                item.setAuthorityName(link.getAuthorityName());
+                item.setCapabilitiesDocument(cap);
 
-            //   item.setCapabilitiesDocument(cap);
-            result.add(item);
+                //   item.setCapabilitiesDocument(cap);
+                result.add(item);
+            }
         }
         return result;
     }

diff --git a/...eventprocessor/processors/postprocess/EventProcessor_PostProcessDatasetDocumentEvent.java b/...eventprocessor/processors/postprocess/EventProcessor_PostProcessDatasetDocumentEvent.java
@@ -331,7 +331,7 @@ private void process() {
             // Dataset link simplification
             if (!localDatasetMetadataRecord.getDocumentLinks().isEmpty()) {
                 List<DocumentLink> viewLinksMetadataOnlineResources = localDatasetMetadataRecord.getDocumentLinks().stream()
-                        .filter(x -> (x.getLinkState().equals(LinkState.Complete) && x.getLinkHTTPStatusCode() == 200) && (DocumentLink.validViewProtocols.contains(x.getProtocol().toLowerCase()) || x.getProtocol().toLowerCase().matches(DocumentLink.VALID_PROTOCOLS_VIEW_REGEX)))
+                        .filter(x -> (x.getLinkState().equals(LinkState.Complete) && x.getLinkHTTPStatusCode() != null && x.getLinkHTTPStatusCode() == 200) && (DocumentLink.validViewProtocols.contains(x.getProtocol().toLowerCase()) || x.getProtocol().toLowerCase().matches(DocumentLink.VALID_PROTOCOLS_VIEW_REGEX)))
                         .collect(Collectors.toList());
 
                 if (!viewLinksMetadataOnlineResources.isEmpty()) {
@@ -348,7 +348,7 @@ private void process() {
             // Dataset link simplification
             if (!localDatasetMetadataRecord.getDocumentLinks().isEmpty()) {
                 List<DocumentLink> downloadLinksMetadataOnlineResources = localDatasetMetadataRecord.getDocumentLinks().stream()
-                        .filter(x -> (x.getLinkState().equals(LinkState.Complete) && x.getLinkHTTPStatusCode() == 200) && (DocumentLink.validDownloadProtocols.contains(x.getProtocol().toLowerCase()) || x.getProtocol().toLowerCase().matches(DocumentLink.VALID_PROTOCOLS_DOWNLOAD_REGEX)))
+                        .filter(x -> (x.getLinkState().equals(LinkState.Complete) && x.getLinkHTTPStatusCode() != null && x.getLinkHTTPStatusCode() == 200) && (DocumentLink.validDownloadProtocols.contains(x.getProtocol().toLowerCase()) || x.getProtocol().toLowerCase().matches(DocumentLink.VALID_PROTOCOLS_DOWNLOAD_REGEX)))
                         .collect(Collectors.toList());
 
                 if (!downloadLinksMetadataOnlineResources.isEmpty()) {