[RELEASE] iText pdfOCR 4.1.2

Author: iText-CI

pdfocr-api/pom.xml

@@ -5,7 +5,7 @@
   <parent>
     <groupId>com.itextpdf</groupId>
     <artifactId>pdfocr-root</artifactId>
-    <version>4.1.1</version>
+    <version>4.1.2</version>
   </parent>
 
   <properties>

pdfocr-onnxtr/pom.xml

@@ -5,7 +5,7 @@
   <parent>
     <groupId>com.itextpdf</groupId>
     <artifactId>pdfocr-root</artifactId>
-    <version>4.1.1</version>
+    <version>4.1.2</version>
   </parent>
 
   <artifactId>pdfocr-onnxtr</artifactId>

pdfocr-onnxtr/src/main/java/com/itextpdf/pdfocr/onnxtr/actions/data/PdfOcrOnnxTrProductData.java

@@ -30,7 +30,7 @@ This file is part of the iText (R) project.
 public final class PdfOcrOnnxTrProductData {
     private static final String PDF_OCR_ONNXTR_PRODUCT_NAME = "pdfOcr-onnxtr";
     private static final String PDF_OCR_ONNXTR_PUBLIC_PRODUCT_NAME = "pdfOCR-OnnxTR";
-    private static final String PDF_OCR_VERSION = "4.1.1";
+    private static final String PDF_OCR_VERSION = "4.1.2";
     private static final int PDF_OCR_COPYRIGHT_SINCE = 2000;
     private static final int PDF_OCR_COPYRIGHT_TO = 2025;
 

pdfocr-onnxtr/src/test/java/com/itextpdf/pdfocr/onnxtr/OnnxTRCmykIntegrationTest.java

@@ -29,8 +29,10 @@ This file is part of the iText (R) project.
 import com.itextpdf.kernel.pdf.PdfWriter;
 import com.itextpdf.pdfocr.OcrPdfCreator;
 import com.itextpdf.pdfocr.OcrPdfCreatorProperties;
+import com.itextpdf.pdfocr.exceptions.PdfOcrInputException;
 import com.itextpdf.pdfocr.onnxtr.detection.IDetectionPredictor;
 import com.itextpdf.pdfocr.onnxtr.detection.OnnxDetectionPredictor;
+import com.itextpdf.pdfocr.onnxtr.exceptions.PdfOcrOnnxTrExceptionMessageConstant;
 import com.itextpdf.pdfocr.onnxtr.orientation.IOrientationPredictor;
 import com.itextpdf.pdfocr.onnxtr.orientation.OnnxOrientationPredictor;
 import com.itextpdf.pdfocr.onnxtr.recognition.IRecognitionPredictor;
@@ -83,16 +85,19 @@ public void rainbowInvertedCmykTest() throws IOException {
         String dest = TARGET_DIRECTORY + "rainbowInvertedCmykTest.pdf";
         String cmpTxt = TEST_DIRECTORY + "cmp_rainbowInvertedCmykTest.txt";
 
-        if (isFixedInJdk()) {
+        try {
             doOcrAndCreatePdf(src, dest, creatorProperties("Text1", DeviceCmyk.MAGENTA));
             try (PdfDocument pdfDocument = new PdfDocument(new PdfReader(dest))) {
                 ExtractionStrategy extractionStrategy = OnnxTestUtils.extractTextFromLayer(pdfDocument, 1, "Text1");
                 Assertions.assertEquals(DeviceCmyk.MAGENTA, extractionStrategy.getFillColor());
                 Assertions.assertEquals(getCmpText(cmpTxt), extractionStrategy.getResultantText());
             }
-        } else {
-            Exception e = Assertions.assertThrows(Exception.class, () -> doOcrAndCreatePdf(src, dest, null));
-            Assertions.assertEquals("Failed to read image.", e.getMessage());
+        } catch (PdfOcrInputException e) {
+            // CMYK bug https://bugs.openjdk.org/browse/JDK-8274735 in openJDK:
+            // fixed for jdk8 from 351 onwards, for jdk11 from 16 onwards and for jdk17 starting from 4.
+            // Amazon corretto jdk started support CMYK for JPEG from 11 version.
+            // Temurin 8 does not support CMYK for JPEG either.
+            Assertions.assertEquals(PdfOcrOnnxTrExceptionMessageConstant.FAILED_TO_READ_IMAGE, e.getMessage());
         }
     }
 
@@ -102,7 +107,7 @@ public void rainbowAdobeCmykTest() throws IOException {
         String dest = TARGET_DIRECTORY + "rainbowAdobeCmykTest.pdf";
         String cmpTxt = TEST_DIRECTORY + "cmp_rainbowAdobeCmykTest.txt";
 
-        if (isFixedInJdk()) {
+        try {
             doOcrAndCreatePdf(src, dest, creatorProperties("Text1", DeviceCmyk.MAGENTA));
             try (PdfDocument pdfDocument = new PdfDocument(new PdfReader(dest))) {
                 ExtractionStrategy extractionStrategy = OnnxTestUtils.extractTextFromLayer(pdfDocument, 1, "Text1");
@@ -111,9 +116,12 @@ public void rainbowAdobeCmykTest() throws IOException {
                         extractionStrategy.getResultantText()) / getCmpText(cmpTxt).length();
                 Assertions.assertTrue(relativeDistance < 0.05);
             }
-        } else {
-            Exception e = Assertions.assertThrows(Exception.class, () -> doOcrAndCreatePdf(src, dest, null));
-            Assertions.assertEquals("Failed to read image.", e.getMessage());
+        } catch (PdfOcrInputException e) {
+            // CMYK bug https://bugs.openjdk.org/browse/JDK-8274735 in openJDK:
+            // fixed for jdk8 from 351 onwards, for jdk11 from 16 onwards and for jdk17 starting from 4.
+            // Amazon corretto jdk started support CMYK for JPEG from 11 version.
+            // Temurin 8 does not support CMYK for JPEG either.
+            Assertions.assertEquals(PdfOcrOnnxTrExceptionMessageConstant.FAILED_TO_READ_IMAGE, e.getMessage());
         }
     }
 
@@ -123,64 +131,22 @@ public void rainbowCmykNoProfileTest() throws IOException {
         String dest = TARGET_DIRECTORY + "rainbowCmykNoProfileTest.pdf";
         String cmpTxt = TEST_DIRECTORY + "cmp_rainbowCmykNoProfileTest.txt";
 
-        if (isFixedInJdk()) {
+        try {
             doOcrAndCreatePdf(src, dest, creatorProperties("Text1", DeviceCmyk.MAGENTA));
             try (PdfDocument pdfDocument = new PdfDocument(new PdfReader(dest))) {
                 ExtractionStrategy extractionStrategy = OnnxTestUtils.extractTextFromLayer(pdfDocument, 1, "Text1");
                 Assertions.assertEquals(DeviceCmyk.MAGENTA, extractionStrategy.getFillColor());
                 Assertions.assertEquals(getCmpText(cmpTxt), extractionStrategy.getResultantText());
             }
-        } else {
-            Exception e = Assertions.assertThrows(Exception.class, () -> doOcrAndCreatePdf(src, dest, null));
-            Assertions.assertEquals("Failed to read image.", e.getMessage());
+        } catch (PdfOcrInputException e) {
+            // CMYK bug https://bugs.openjdk.org/browse/JDK-8274735 in openJDK:
+            // fixed for jdk8 from 351 onwards, for jdk11 from 16 onwards and for jdk17 starting from 4.
+            // Amazon corretto jdk started support CMYK for JPEG from 11 version.
+            // Temurin 8 does not support CMYK for JPEG either.
+            Assertions.assertEquals(PdfOcrOnnxTrExceptionMessageConstant.FAILED_TO_READ_IMAGE, e.getMessage());
         }
     }
 
-    private static boolean isFixedInJdk() {
-        //Fixed CMYK bug https://bugs.openjdk.org/browse/JDK-8274735 for openJDK:
-        //jdk8 from 351 onwards, for jdk11 from 16 onwards and for jdk17 starting from 4.
-        //Amazon corretto jdk started support CMYK for JPEG from 11 version.
-        //Temurin 8 does not support CMYK for JPEG either.
-        String versionStr = System.getProperty("java.version");
-        String vendorStr = System.getProperty("java.vendor");
-        boolean isFixed = false;
-        int majorVer = getMajorVer(versionStr);
-        String[] split = versionStr.split("[._-]");
-        int minorVer = Integer.parseInt(split[split.length - 1]);
-
-        switch (majorVer) {
-            case 8:
-                if ("Amazon.com Inc.".equals(vendorStr) || "Temurin".equals(vendorStr)) {
-                    return false;
-                }
-
-                isFixed = minorVer >= 351;
-                break;
-            case 11:
-                isFixed = minorVer >= 16;
-                break;
-            case 17:
-                isFixed = minorVer >= 4;
-                break;
-            default:
-                isFixed = true;
-        }
-
-        return isFixed;
-    }
-
-    private static int getMajorVer(String versionStr) {
-        int majorVer = 0;
-        String[] split = versionStr.split("\\.");
-        if (versionStr.startsWith("1.")) {
-            //jdk versions 1 - 8 have 1. as prefix
-            majorVer = Integer.parseInt(split[1]);
-        } else {
-            majorVer = Integer.parseInt(split[0]);
-        }
-        return majorVer;
-    }
-
     private OcrPdfCreatorProperties creatorProperties(String layerName, Color color) {
         OcrPdfCreatorProperties ocrPdfCreatorProperties = new OcrPdfCreatorProperties();
         ocrPdfCreatorProperties.setTextLayerName(layerName);

pdfocr-tesseract4/pom.xml

@@ -5,7 +5,7 @@
   <parent>
     <groupId>com.itextpdf</groupId>
     <artifactId>pdfocr-root</artifactId>
-    <version>4.1.1</version>
+    <version>4.1.2</version>
   </parent>
 
   <artifactId>pdfocr-tesseract4</artifactId>

pdfocr-tesseract4/src/main/java/com/itextpdf/pdfocr/tesseract4/actions/data/PdfOcrTesseract4ProductData.java

@@ -30,7 +30,7 @@ This file is part of the iText (R) project.
 public class PdfOcrTesseract4ProductData {
     private static final String PDF_OCR_TESSERACT4_PRODUCT_NAME = "pdfOcr-tesseract4";
     private static final String PDF_OCR_TESSERACT4_PUBLIC_PRODUCT_NAME = "pdfOCR-Tesseract4";
-    private static final String PDF_OCR_VERSION = "4.1.1";
+    private static final String PDF_OCR_VERSION = "4.1.2";
     private static final int PDF_OCR_COPYRIGHT_SINCE = 2000;
     private static final int PDF_OCR_COPYRIGHT_TO = 2025;
 

pdfocr-tesseract4/src/test/java/com/itextpdf/pdfocr/tesseract4/ImageIntegrationTest.java

@@ -27,6 +27,7 @@ This file is part of the iText (R) project.
 import com.itextpdf.pdfocr.TextInfo;
 import org.junit.jupiter.api.Assertions;
 import org.junit.jupiter.api.BeforeEach;
+import org.junit.jupiter.api.Disabled;
 import org.junit.jupiter.api.Test;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
@@ -88,6 +89,7 @@ public void testHocrRotatedImage() throws IOException {
     }
 
     @Test
+    @Disabled("DEVSIX-9261 Investigate test failures on Windows Server 2025 and Windows 11")
     public void compareRotatedImage() throws InterruptedException, IOException {
         String testName = "compareRotatedImage";
         String filename = "90_degrees_rotated";
@@ -109,7 +111,7 @@ public void compareRotatedImage() throws InterruptedException, IOException {
                 Arrays.<String>asList("eng"), Arrays.<String>asList(NOTO_SANS_FONT_PATH),
                 null, true);
 
-        // Because of difference of tesseract 5 and tesseract 4 there're some differences in text recognition.
+        // Because of difference of tesseract 5 and tesseract 4 there are some differences in text recognition.
         // So the goal of this test is to make text invisible and check if image is rotated.
         // Proper text recognition is compared in testHocrRotatedImage test by checking HOCR file.
         boolean javaTest = new CompareTool().compareVisually(resultPdfPath, expectedPdfPathJava,

pom.xml

@@ -5,12 +5,12 @@
   <parent>
     <groupId>com.itextpdf</groupId>
     <artifactId>root</artifactId>
-    <version>9.4.0</version>
+    <version>9.5.0</version>
     <relativePath />
   </parent>
 
   <artifactId>pdfocr-root</artifactId>
-  <version>4.1.1</version>
+  <version>4.1.2</version>
   <packaging>pom</packaging>
 
   <name>pdfOCR</name>
@@ -23,7 +23,7 @@
   </modules>
 
   <properties>
-    <itext.version>9.4.0</itext.version>
+    <itext.version>9.5.0</itext.version>
     <java.version>1.8</java.version>
     <jdkLevel>${java.version}</jdkLevel>
     <maven.compiler.source>${java.version}</maven.compiler.source>

sharpenConfiguration.xml

@@ -21,10 +21,6 @@
                 <file path="com/itextpdf/pdfocr/onnxtr/OnnxTRPdfAIntegrationTest.java" />
                 <file path="com/itextpdf/pdfocr/onnxtr/OnnxTRRotationIntegrationTest.java" />
             </fileset>
-            <!-- TODO DEVSIX-9305: fileset should be removed and OnnxTRCmykIntegrationTest made autoportable-->
-            <fileset reason="Bad handling of JPEG CMYK in java versions without JDK-8274735 fix">
-                <file path="com/itextpdf/pdfocr/onnxtr/OnnxTRCmykIntegrationTest.java" />
-            </fileset>
             <fileset reason="Difference in Iterators in java and enumerators in .net">
                 <file path="com/itextpdf/pdfocr/onnxtr/util/Batching.java" />
                 <file path="com/itextpdf/pdfocr/onnxtr/util/BatchProcessingGeneratorTest.java" />