close #434: Cannot update metadata: Namespace for prefix 'rdf' has no…

…t been declared
torakiki · Jun 28, 2024 · d1ab11c · d1ab11c
1 parent b3c3339
commit d1ab11c
Show file tree

Hide file tree

Showing 4 changed files with 47 additions and 7 deletions.
diff --git a/sejda-sambox/src/main/java/org/sejda/impl/sambox/SetMetadataTask.java b/sejda-sambox/src/main/java/org/sejda/impl/sambox/SetMetadataTask.java
@@ -141,9 +141,10 @@ public void execute(SetMetadataParameters parameters) throws TaskException {
                         try {
                             updateXmpMetadata(catalog, actualMeta);
                         } catch (RuntimeException ex) {
-                            if(exceptionStackContains("Namespace for prefix 'xmp' has not been declared", ex)) {
+                            if(exceptionStackContains("Namespace for prefix 'xmp' has not been declared", ex) ||
+                                    exceptionStackContains("Namespace for prefix 'rdf' has not been declared", ex)) {
                                 // try to fix the metadata and retry
-                                fixMissingXmpNamespace(catalog);
+                                fixMissingXmpOrRdfNamespace(catalog);
                                 updateXmpMetadata(catalog, actualMeta);
                             } else {
                                 throw ex;
@@ -229,18 +230,27 @@ private void deleteAttr(String path, String attrName, Document document) throws
         }
     }
 
-    private void fixMissingXmpNamespace(PDDocumentCatalog catalog) throws IOException {
+    private void fixMissingXmpOrRdfNamespace(PDDocumentCatalog catalog) throws IOException {
         try(InputStream is = catalog.getMetadata().createInputStream()) {
             String metadataAsString = new String(is.readAllBytes(), StandardCharsets.UTF_8);
+            String metadataAsStringOriginal = metadataAsString;
+
             if (!metadataAsString.contains("xmlns:xmp=\"http://ns.adobe.com/xap/1.0/\"")) {
                 LOG.warn("Metadata seems to be missing xmlns:xmp namespace definition, adding it");
                 metadataAsString = metadataAsString.replaceAll("<rdf:Description", "<rdf:Description xmlns:xmp=\"http://ns.adobe.com/xap/1.0/\"");
-
+            }
+
+            if (!metadataAsString.contains("xmlns:rdf=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#\"")) {
+                LOG.warn("Metadata seems to be missing xmlns:rdf namespace definition, adding it");
+                metadataAsString = metadataAsString.replaceAll("<rdf:RDF", "<rdf:RDF xmlns:rdf=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#\">");
+            }
+
+            if(!metadataAsStringOriginal.equals(metadataAsString)){
                 catalog.setMetadata(new PDMetadata(new ByteArrayInputStream(metadataAsString.getBytes(StandardCharsets.UTF_8))));
-            } 
+            }
         }
     }
-    
+
     private void updateXmpMetadata(PDDocumentCatalog catalog, PDDocumentInformation metadata) {
         try {
             DocumentBuilderFactory f = DocumentBuilderFactory.newInstance();

diff --git a/sejda-sambox/src/test/java/org/sejda/impl/sambox/SetMetadataSamboxTaskTest.java b/sejda-sambox/src/test/java/org/sejda/impl/sambox/SetMetadataSamboxTaskTest.java
@@ -265,7 +265,7 @@ public void invalidXmlChars_updatedXmpMetadata() throws IOException {
 
     @Test
     public void xmpMetadataMissingNamespaceAndWithAttributesInsteadofNodes() throws IOException {
-        setUpParams(customInput("pdf/xmp_metadata_missing_namespace.pdf"));
+        setUpParams(customInput("pdf/xmp_metadata_missing_xmp_namespace.pdf"));
         parameters.put(PdfMetadataFields.CREATOR, "test_creator");
         parameters.put("Producer", "test_producer");
 
@@ -306,6 +306,36 @@ public void xmpMetadataMissingNamespaceAndWithAttributesInsteadofNodes() throws
         });
     }
 
+    @Test
+    public void xmpMetadataMissingNamespace_rdf() throws IOException {
+        setUpParams(customInput("pdf/xmp_metadata_missing_rdf_namespace.pdf"));
+        String random = Long.toString(System.currentTimeMillis());
+        String producer = "test_producer_" + random;
+        String creator = "test_creator_" + random;
+
+        parameters.put(PdfMetadataFields.CREATOR, creator);
+        parameters.put("Producer", producer);
+
+        parameters.setExistingOutputPolicy(ExistingOutputPolicy.OVERWRITE);
+        testContext.directoryOutputTo(parameters);
+
+        execute(parameters);
+
+        testContext.assertTaskCompleted();
+        testContext.forEachPdfOutput(document -> {
+            try {
+                DocumentBuilderFactory f = DocumentBuilderFactory.newInstance();
+                DocumentBuilder b = f.newDocumentBuilder();
+                Document xmlDoc = b.parse(document.getDocumentCatalog().getMetadata().createInputStream());
+
+                assertEquals(producer, getNodeValue(xmlDoc, "//*[name()='pdf:Producer']"));
+
+            } catch (Exception e) {
+                throw new RuntimeException(e);
+            }
+        });
+    }
+
     @Test
     public void updatedXmpMetadataRemoveFields() throws IOException {
         setUpParams(stronglyEncryptedInput());

diff --git a/sejda-tests/src/main/resources/pdf/xmp_metadata_missing_rdf_namespace.pdf b/sejda-tests/src/main/resources/pdf/xmp_metadata_missing_rdf_namespace.pdf
diff --git a/...es/pdf/xmp_metadata_missing_namespace.pdf → ...df/xmp_metadata_missing_xmp_namespace.pdf b/...es/pdf/xmp_metadata_missing_namespace.pdf → ...df/xmp_metadata_missing_xmp_namespace.pdf