Update pdfbox to 3.0.* (#177)

philterd · Dec 17, 2024 · ab0c806 · ab0c806
1 parent e86c538
commit ab0c806
Show file tree

Hide file tree

Showing 3 changed files with 6 additions and 5 deletions.
diff --git a/...eas-services/phileas-services-pdf/src/main/java/ai/philterd/services/pdf/PdfRedacter.java b/...eas-services/phileas-services-pdf/src/main/java/ai/philterd/services/pdf/PdfRedacter.java
@@ -24,6 +24,7 @@
 import ai.philterd.services.pdf.model.RedactedRectangle;
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
+import org.apache.pdfbox.Loader;
 import org.apache.pdfbox.pdmodel.PDDocument;
 import org.apache.pdfbox.pdmodel.PDPage;
 import org.apache.pdfbox.pdmodel.PDPageContentStream;
@@ -83,8 +84,7 @@ public PdfRedacter(Policy policy,
 
     @Override
     public byte[] process(byte[] document, MimeType outputMimeType) throws IOException {
-
-        final PDDocument pdDocument = PDDocument.load(document);
+        final PDDocument pdDocument = Loader.loadPDF(document);
 
         setSortByPosition(true);
         setStartPage(0);
@@ -209,7 +209,7 @@ protected void endDocument(PDDocument doc) throws IOException {
         for(int pageNumber : rectangles.keySet()) {
 
             final PDPage page = document.getPage(pageNumber);
-            final PDPageContentStream contentStream = new PDPageContentStream(doc, page, true, true, true);
+            final PDPageContentStream contentStream = new PDPageContentStream(doc, page, PDPageContentStream.AppendMode.APPEND, true, true);
 
             for(final RedactedRectangle rectangle : rectangles.get(pageNumber)) {
 

diff --git a/...ervices/phileas-services-pdf/src/main/java/ai/philterd/services/pdf/PdfTextExtractor.java b/...ervices/phileas-services-pdf/src/main/java/ai/philterd/services/pdf/PdfTextExtractor.java
@@ -19,6 +19,7 @@
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
 import org.apache.pdfbox.pdmodel.PDDocument;
+import org.apache.pdfbox.Loader;
 import org.apache.pdfbox.text.PDFTextStripper;
 import org.apache.pdfbox.text.TextPosition;
 
@@ -45,7 +46,7 @@ public PdfTextExtractor() throws IOException {
     @Override
     public List<String> getLines(byte[] document) throws IOException {
 
-        final PDDocument pdDocument = PDDocument.load(document);
+        final PDDocument pdDocument = Loader.loadPDF(document);
 
         this.setSortByPosition(true);
         this.setStartPage(0);

diff --git a/pom.xml b/pom.xml
@@ -108,7 +108,7 @@
         <mockito.version>5.14.2</mockito.version>
         <nv-i18n.version>1.27</nv-i18n.version>
         <opennlp.version>2.5.0</opennlp.version>
-        <pdfbox.version>2.0.23</pdfbox.version>
+        <pdfbox.version>3.0.3</pdfbox.version>
         <redis-mock.version>0.1.6</redis-mock.version>
         <redisson.version>3.39.0</redisson.version>
         <retrofit.version>2.11.0</retrofit.version>