update condition for merging images

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/segmentation/PdfSegmentationService.java

@@ -8,6 +8,7 @@ import com.iqser.red.service.redaction.v1.server.classification.service.Classifi
 import com.iqser.red.service.redaction.v1.server.memory.MemoryStats;
 import com.iqser.red.service.redaction.v1.server.parsing.PDFLinesTextStripper;
 import com.iqser.red.service.redaction.v1.server.parsing.model.TextPositionSequence;
+import com.iqser.red.service.redaction.v1.server.redaction.model.PdfImage;
 import com.iqser.red.service.redaction.v1.server.redaction.service.ImageClassificationService;
 import com.iqser.red.service.redaction.v1.server.tableextraction.model.AbstractTextContainer;
 import com.iqser.red.service.redaction.v1.server.tableextraction.model.CleanRulings;
@@ -22,6 +23,9 @@ import org.apache.pdfbox.pdmodel.PDPage;
 import org.apache.pdfbox.pdmodel.common.PDRectangle;
 import org.springframework.stereotype.Service;
 
+import java.awt.Graphics;
+import java.awt.geom.Rectangle2D;
+import java.awt.image.BufferedImage;
 import java.io.File;
 import java.io.FileOutputStream;
 import java.io.IOException;
@@ -92,7 +96,8 @@ public class PdfSegmentationService {
                 page.setRotation(rotation);
                 page.setLandscape(isLandscape || isRotated);
                 page.setPageNumber(pageNumber);
-                page.setImages(stripper.getImages());
+                List<PdfImage> mergedList = processImages(stripper.getImages());
+                page.setImages(mergedList);
 
                 tableExtractionService.extractTables(cleanRulings, page);
                 buildPageStatistics(page);
@@ -105,6 +110,7 @@ public class PdfSegmentationService {
 
                 pages.add(page);
 
+
             }
 
             document.setPages(pages);
@@ -142,6 +148,109 @@ public class PdfSegmentationService {
         return newPDDocument;
     }
 
+    //merge images, if they are separated during pdf import, return new list of Pdfimages
+    private List<PdfImage> processImages(List<PdfImage> imageList){
+        if (imageList.size() >1) {
+            List<PdfImage> mergedList = new ArrayList<>();
+            int countElementsInList = 0;
+            boolean beginImage = true;
+
+            // a List of Boolean, true = candidate for merging, false = no merging
+            ArrayList<Boolean> candidatesList = getCandidatesList(imageList);
+
+            // loop through list, if there are candidates for merging (true), merge images and add it to mergedList
+            for (int i = 0; i < candidatesList.size(); i++) {
+                if (candidatesList.get(i)) {
+                    if (beginImage) {
+                        //begin of image, merge two parts of imageList
+                        PdfImage mergedImage = mergeTwoImages(imageList.get(i), imageList.get(i + 1));
+                        mergedList.add(mergedImage);
+                        countElementsInList++;
+                    } else {
+                        //middle of an image, merge current piece auf mergedList with image of imageList
+                        PdfImage mergedImage = mergeTwoImages(mergedList.get(countElementsInList - 1), imageList.get(i + 1));
+                        mergedList.set(countElementsInList - 1, mergedImage);
+                    }
+                    beginImage = false;
+                } else {
+                    // if the last candidate is false, then both images i and i+1 must be added
+                    if (i == candidatesList.size()-1){
+                        mergedList.add(imageList.get(i));
+                        mergedList.add(imageList.get(i+1));
+                    }else {
+                        //first image is not splitted, add i to resultlist
+                        if (beginImage){
+                            mergedList.add(imageList.get(i));
+                            countElementsInList++;
+                        }else {
+                            // i is the end of an image, add begin of new image
+                            mergedList.add(imageList.get(i + 1));
+                            countElementsInList++;
+                            beginImage = false;
+                        }
+                    }
+                }
+            }return mergedList;
+        }else {
+            return imageList;
+        }
+    }
+
+    private PdfImage mergeTwoImages(PdfImage image1, PdfImage image2){
+        PdfImage newPdfImage = null;
+        // diese Angaben von getPosition scheinen nicht richtig zu sein, damit werden teile des Bildes abgeschnitten
+        double width = image1.getPosition().getWidth();
+        double height1 = image1.getPosition().getHeight();
+        double height2 = image2.getPosition().getHeight();
+        // mit  den Werten, die unter Image gespeichert sind, funktioniert es
+        double img1height = image1.getImage().getHeight();
+        double img1width = image1.getImage().getWidth();
+        double img2height = image2.getImage().getHeight();
+
+        BufferedImage mergedImage = new BufferedImage((int) img1width, (int) (img1height+ img2height), BufferedImage.TYPE_INT_RGB);
+        Graphics g = mergedImage.getGraphics();
+        try {
+            g.drawImage(image1.getImage(),0,0,null);
+            g.drawImage(image2.getImage(),0, (int) (img1height),null);
+
+            // set Image, Position and type for merged Image
+            //set position for merged image with values of image1 and the height of both
+            Rectangle2D pos = new Rectangle2D.Float();
+            pos.setRect(image1.getPosition().getX(), image1.getPosition().getY(), width,height1+height2);
+            newPdfImage = new PdfImage(mergedImage,pos,0);
+            // Graphics need to be disposed
+            newPdfImage.getImage().flush();
+            newPdfImage.getImage().getGraphics().dispose();
+        } catch (Exception e) {
+            e.printStackTrace();
+        }
+        return newPdfImage;
+    }
+
+    //make a list of true and false, if the image is a candidate for merging
+    private ArrayList<Boolean> getCandidatesList(List<PdfImage> imageList) {
+        ArrayList<Boolean> candidatesList = new ArrayList<>();
+        for (int i = 0; i < imageList.size(); i++) {
+            if (i >= 1) {
+                candidatesList.add(isCandidateForMerging(imageList.get(i - 1), imageList.get(i)));
+            }
+        }return candidatesList;
+    }
+
+    // evaluate if two images are candidates for merging, depending on their coordinates, width and height
+    private boolean isCandidateForMerging(PdfImage image1, PdfImage image2){
+        double x1 = image1.getPosition().getX();
+        double y1 = image1.getPosition().getY();
+        double width1 = image1.getPosition().getWidth();
+        double x2 = image2.getPosition().getX();
+        double y2 = image2.getPosition().getY();
+        double width2 = image2.getPosition().getWidth();
+        double height2 = image2.getPosition().getHeight();
+        //if the x-coordinates and widths of images are equal and the height is equal to difference between y-coordinates,
+        // then it is the same picture and has to be merged -> return true
+        return x1 == x2 && width1 == width2 && Math.ceil(height2) == Math.ceil(y1 - y2) && width2 > (height2 / 6);
+    }
+
 
     private void increaseDocumentStatistics(Page page, Document document) {
 

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/segmentation/SectionsBuilderService.java

@@ -121,6 +121,14 @@ public class SectionsBuilderService {
             paragraphMap.computeIfAbsent(1, x -> new TreeSet<>()).add(paragraph);
         }
 
+        // first page is always a paragraph, else we can't process pages 1..N,
+        // where N is the first found page with a paragraph
+        if (paragraphMap.get(1) == null) {
+            Paragraph paragraph = new Paragraph();
+            document.getParagraphs().add(paragraph);
+            paragraphMap.computeIfAbsent(1, x -> new TreeSet<>()).add(paragraph);
+        }
+
         for (Page page : document.getPages()) {
             for (PdfImage image : page.getImages()) {
                 SortedSet<Paragraph> paragraphsOnPage = paragraphMap.get(page.getPageNumber());

redaction-service-v1/redaction-service-server-v1/src/test/java/com/iqser/red/service/redaction/v1/server/RedactionIntegrationTest.java

@@ -484,7 +484,6 @@ public class RedactionIntegrationTest {
 
 
     @Test
-    @Ignore
     public void noExceptionShouldBeThrownForAnyFiles() throws IOException {
 
         long start = System.currentTimeMillis();

redaction-service-v1/redaction-service-server-v1/src/test/java/com/iqser/red/service/redaction/v1/server/segmentation/PdfSegmentationServiceTest.java

@@ -68,6 +68,23 @@ public class PdfSegmentationServiceTest {
 
     }
 
+    @Test
+    public void testMergeImages() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/270Rotated.pdf");
+
+        Document document = pdfSegmentationService.parseDocument(pdfFileResource.getInputStream());
+
+    }
+    @Test
+    public void testMergeImages2() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/new/Single Study - Oral (Gavage) Mouse.pdf");
+
+        Document document = pdfSegmentationService.parseDocument(pdfFileResource.getInputStream());
+
+    }
+
     @Test
     @Ignore
     public void testExtractImages() throws IOException {