Fix annotating single letter entities

Merge in RED/redaction-service from RED-264 to master

* commit 'e1adcfb02c4c7d46140884b3d38f8daf1e1aae92':
  Added missing unittest
  RED-264: Avoid phantom cells, by merging line till rounded value of biggest char height/width

bamboo-specs/pom.xml

@@ -5,7 +5,7 @@
   <parent>
     <groupId>com.atlassian.bamboo</groupId>
     <artifactId>bamboo-specs-parent</artifactId>
-    <version>7.0.4</version>
+    <version>7.1.2</version>
     <relativePath/>
   </parent>
 
@@ -34,4 +34,4 @@
 
   <!-- run 'mvn test' to perform offline validation of the plan -->
   <!-- run 'mvn -Ppublish-specs' to upload the plan to your Bamboo server -->
-</project>
+</project>

redaction-service-image-v1/src/main/docker/Dockerfile

@@ -1,4 +1,4 @@
-FROM gin5/platform-base:5.2.0
+FROM red/base-image:1.0.0
 
 ARG PLATFORM_JAR
 
@@ -6,4 +6,4 @@ ENV PLATFORM_JAR ${PLATFORM_JAR}
 
 ENV USES_ELASTICSEARCH false
 
-COPY ["${PLATFORM_JAR}", "/"]
+COPY ["${PLATFORM_JAR}", "/"]

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/classification/service/ClassificationService.java

@@ -84,7 +84,7 @@ public class ClassificationService {
         }
         else if (PositionUtils.isWithinBodyTextFrame(bodyTextFrame, textBlock) && textBlock.getMostPopularWordFontSize() == document.getFontSizeCounter().getMostPopular() && textBlock.getMostPopularWordStyle().equals("italic") && !document.getFontStyleCounter().getMostPopular().equals("italic") && PositionUtils.getApproxLineCount(textBlock) < 2.9) {
             textBlock.setClassification("TextBlock Italic");
-        } else if (PositionUtils.isWithinBodyTextFrame(bodyTextFrame, textBlock) && textBlock.getSequences().size() > 3){
+        } else if (PositionUtils.isWithinBodyTextFrame(bodyTextFrame, textBlock)){
             textBlock.setClassification("TextBlock Unknown");
         }
     }

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/parsing/PDFLinesTextStripper.java

@@ -44,10 +44,10 @@ import lombok.extern.slf4j.Slf4j;
 public class PDFLinesTextStripper extends PDFTextStripper {
 
     @Getter
-    private float minCharWidth = Float.MAX_VALUE;
+    private int maxCharWidths;
 
     @Getter
-    private float minCharHeight = Float.MAX_VALUE;
+    private int maxCharHeight;
 
     @Getter
     private final List<TextPositionSequence> textPositionSequences = new ArrayList<>();
@@ -201,8 +201,16 @@ public class PDFLinesTextStripper extends PDFTextStripper {
 
         int startIndex = 0;
         for (int i = 0; i <= textPositions.size() - 1; i++) {
-            minCharWidth = Math.min(minCharWidth, textPositions.get(i).getWidthDirAdj());
-            minCharHeight = Math.min(minCharHeight, textPositions.get(i).getHeightDir());
+
+            int charHeight = (int) textPositions.get(i).getHeightDir();
+            if(charHeight > maxCharHeight){
+                maxCharHeight = charHeight;
+            }
+
+            int charWidth = (int) textPositions.get(i).getWidthDirAdj();
+            if(charWidth > maxCharWidths){
+                maxCharWidths = charWidth;
+            }
 
             if (i == 0 && textPositions.get(i).getUnicode().equals(" ")) {
                 startIndex++;
@@ -241,8 +249,8 @@ public class PDFLinesTextStripper extends PDFTextStripper {
     @Override
     public String getText(PDDocument doc) throws IOException {
 
-        minCharWidth = Float.MAX_VALUE;
-        minCharHeight = Float.MAX_VALUE;
+        maxCharWidths = 0;
+        maxCharWidths = 0;
         textPositionSequences.clear();
         rulings.clear();
         graphicsPath.clear();

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/parsing/model/ParsedElements.java

@@ -17,6 +17,6 @@ public class ParsedElements {
     private boolean landscape;
     private boolean rotated;
 
-    private float minCharWidth;
-    private float minCharHeight;
+    private float maxCharWidth;
+    private float maxCharHeight;
 }

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/parsing/model/TextPositionSequence.java

@@ -9,9 +9,7 @@ import com.iqser.red.service.redaction.v1.model.Point;
 import com.iqser.red.service.redaction.v1.model.Rectangle;
 
 import lombok.Data;
-import lombok.Getter;
 import lombok.RequiredArgsConstructor;
-import lombok.Setter;
 
 @Data
 @RequiredArgsConstructor
@@ -19,10 +17,6 @@ public class TextPositionSequence implements CharSequence {
 
     private List<TextPosition> textPositions = new ArrayList<>();
 
-    @Getter
-    @Setter
-    private float[] annotationColor;
-
     private final int page;
 
 
@@ -194,22 +188,18 @@ public class TextPositionSequence implements CharSequence {
 
     public Rectangle getRectangle() {
 
-        float height = textPositions.get(0).getHeightDir() + 2;
+        float height = getTextHeight();
 
-        float posXInit;
+        float posXInit = getX1();
         float posXEnd;
         float posYInit;
         float posYEnd;
 
         if (textPositions.get(0).getRotation() == 90) {
-
             posXEnd = textPositions.get(0).getYDirAdj() + 2;
-            posXInit = textPositions.get(0).getYDirAdj() - height;
-            posYInit = textPositions.get(0).getXDirAdj();
+            posYInit = getY1();
             posYEnd = textPositions.get(textPositions.size() - 1).getXDirAdj() - height + 4;
         } else {
-
-            posXInit = textPositions.get(0).getXDirAdj();
             posXEnd = textPositions.get(textPositions.size() - 1)
                     .getXDirAdj() + textPositions.get(textPositions.size() - 1).getWidth() + 1;
             posYInit = textPositions.get(0).getPageHeight() - textPositions.get(0).getYDirAdj() - 2;
@@ -220,4 +210,5 @@ public class TextPositionSequence implements CharSequence {
         return new Rectangle(new Point(posXInit, posYInit), posXEnd - posXInit, posYEnd - posYInit + height, page);
     }
 
+
 }

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/redaction/model/CellValue.java

@@ -1,6 +1,8 @@
 package com.iqser.red.service.redaction.v1.server.redaction.model;
 
 import com.iqser.red.service.redaction.v1.server.classification.model.TextBlock;
+import com.iqser.red.service.redaction.v1.server.parsing.model.TextPositionSequence;
+import com.iqser.red.service.redaction.v1.server.redaction.utils.TextNormalizationUtilities;
 
 import lombok.RequiredArgsConstructor;
 import lombok.Value;
@@ -13,4 +15,29 @@ public class CellValue {
 
     int rowSpanStart;
 
+    @Override
+    public String toString() {
+
+        StringBuilder sb = new StringBuilder();
+
+
+            TextPositionSequence previous = null;
+            for (TextPositionSequence word : textBlock.getSequences()) {
+
+                if (previous != null) {
+                    if (Math.abs(previous.getY1() - word.getY1()) > word.getTextHeight()) {
+                        sb.append('\n');
+                    } else {
+                        sb.append(' ');
+                    }
+                }
+                sb.append(word.toString());
+                previous = word;
+            }
+
+        return TextNormalizationUtilities.removeHyphenLineBreaks(sb.toString())
+                .replaceAll("\n", " ")
+                .replaceAll(" {2}", " ");
+    }
+
 }

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/redaction/model/Section.java

@@ -33,13 +33,14 @@ public class Section {
     private Map<String, CellValue> tabularData;
 
 
-    public boolean rowEquals(String headerName, String value){
-        String cleanHeaderName = headerName.replaceAll("\n", "")
-                .replaceAll(" ", "")
-                .replaceAll("-", "");
+    public boolean rowEquals(String headerName, String value) {
 
-        return tabularData != null && tabularData.containsKey(cleanHeaderName)
-                && tabularData.get(cleanHeaderName).getTextBlock().getText().equals(value);
+        String cleanHeaderName = headerName.replaceAll("\n", "").replaceAll(" ", "").replaceAll("-", "");
+
+        return tabularData != null && tabularData.containsKey(cleanHeaderName) && tabularData.get(cleanHeaderName)
+                .getTextBlock()
+                .getText()
+                .equals(value);
     }
 
 
@@ -172,25 +173,46 @@ public class Section {
 
     public void highlightCell(String cellHeader, int ruleNumber, String type) {
 
-        String cleanHeaderName = cellHeader.replaceAll("\n", "")
-                .replaceAll(" ", "")
-                .replaceAll("-", "");
+        annotateCell(cellHeader, ruleNumber, type, false, null);
+    }
+
+
+    public void redactCell(String cellHeader, int ruleNumber, String type, String reason) {
+
+        annotateCell(cellHeader, ruleNumber, type, true, reason);
+    }
+
+
+    public void redactNotCell(String cellHeader, int ruleNumber, String type, String reason) {
+
+        annotateCell(cellHeader, ruleNumber, type, false, reason);
+    }
+
+
+    private void annotateCell(String cellHeader, int ruleNumber, String type, boolean redact, String reason) {
+
+        String cleanHeaderName = cellHeader.replaceAll("\n", "").replaceAll(" ", "").replaceAll("-", "");
 
         CellValue value = tabularData.get(cleanHeaderName);
         if (value == null) {
             log.warn("Could not find any data for {}.", cellHeader);
         } else {
-            Entity entity = new Entity(value.getTextBlock()
-                    .getText(), type, value.getRowSpanStart(), value.getRowSpanStart() + value.getTextBlock()
-                    .getText()
+            Entity entity = new Entity(value.toString(), type, value.getRowSpanStart(), value.getRowSpanStart() + value.toString()
                     .length(), headline, sectionNumber);
-            entity.setRedaction(false);
+            entity.setRedaction(redact);
             entity.setMatchedRule(ruleNumber);
-            entity.setRedactionReason(cellHeader);
-            entity.setTargetSequences(value.getTextBlock().getSequences()); // Make sure no other cells with same content are highlighted
-            entities.add(entity);
-        }
+            entity.setRedactionReason(reason);
+            entity.setTargetSequences(value.getTextBlock()
+                    .getSequences()); // Make sure no other cells with same content are highlighted
 
+            // HashSet keeps the older value, but we want the new only.
+            if(entities.contains(entity)){
+                entities.remove(entity);
+            }
+            entities.add(entity);
+
+            entities = removeEntitiesContainedInLarger(entities);
+        }
     }
 
 }

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

@@ -27,9 +27,7 @@ import com.iqser.red.service.redaction.v1.server.tableextraction.model.Cell;
 import com.iqser.red.service.redaction.v1.server.tableextraction.model.Table;
 
 import lombok.RequiredArgsConstructor;
-import lombok.extern.slf4j.Slf4j;
 
-@Slf4j
 @Service
 @RequiredArgsConstructor
 public class EntityRedactionService {
@@ -72,7 +70,7 @@ public class EntityRedactionService {
                                     .replaceAll("-", "");
                             tabularData.put(headerName, new CellValue(cell.getTextBlocks().get(0), cellStart));
                         });
-                        start = start + cell.getTextBlocks().get(0).toString().length();
+                        start = start + cell.toString().length();
                         for (TextBlock textBlock : cell.getTextBlocks()) {
                             searchableRow.addAll(textBlock.getSequences());
                         }

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

@@ -21,7 +21,6 @@ import com.iqser.red.service.redaction.v1.server.tableextraction.model.AbstractT
 import com.iqser.red.service.redaction.v1.server.tableextraction.model.CleanRulings;
 import com.iqser.red.service.redaction.v1.server.tableextraction.service.RulingCleaningService;
 import com.iqser.red.service.redaction.v1.server.tableextraction.service.TableExtractionService;
-import com.iqser.red.service.redaction.v1.server.tableextraction.utils.Utils;
 
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
@@ -37,6 +36,7 @@ public class PdfSegmentationService {
     private final ClassificationService classificationService;
     private final SectionsBuilderService sectionsBuilderService;
 
+
     public Document parseDocument(PDDocument pdDocument) throws IOException {
 
         Document document = new Document();
@@ -56,19 +56,21 @@ public class PdfSegmentationService {
             int rotation = pdPage.getRotation();
             boolean isRotated = rotation != 0 && rotation != 360;
 
-            ParsedElements parsedElements = ParsedElements
-                    .builder()
+
+            ParsedElements parsedElements = ParsedElements.builder()
                     .rulings(stripper.getRulings())
                     .sequences(stripper.getTextPositionSequences())
-                    .minCharWidth(Utils.round(stripper.getMinCharWidth(), 2))
-                    .minCharHeight(Utils.round(stripper.getMinCharHeight(), 2))
+                    .maxCharWidth(stripper.getMaxCharWidths())
+                    .maxCharHeight(stripper.getMaxCharWidths())
                     .landscape(isLandscape)
                     .rotated(isRotated)
                     .build();
 
-            CleanRulings cleanRulings = rulingCleaningService.getCleanRulings(parsedElements.getRulings(), parsedElements.getMinCharWidth(), parsedElements.getMinCharHeight());
+            CleanRulings cleanRulings = rulingCleaningService.getCleanRulings(parsedElements.getRulings(), parsedElements
+                    .getMaxCharWidth(), parsedElements.getMaxCharHeight());
 
-            Page page = blockificationService.blockify(parsedElements.getSequences(), cleanRulings.getHorizontal(), cleanRulings.getVertical());
+            Page page = blockificationService.blockify(parsedElements.getSequences(), cleanRulings.getHorizontal(), cleanRulings
+                    .getVertical());
             page.setRotation(rotation);
 
             tableExtractionService.extractTables(cleanRulings, page);
@@ -91,7 +93,10 @@ public class PdfSegmentationService {
 
     }
 
+
+
     private void increaseDocumentStatistics(Page page, Document document) {
+
         if (!page.isLandscape()) {
             document.getFontSizeCounter().addAll(page.getFontSizeCounter().getCountPerValue());
         }
@@ -100,6 +105,7 @@ public class PdfSegmentationService {
         document.getFontStyleCounter().addAll(page.getFontStyleCounter().getCountPerValue());
     }
 
+
     private void buildPageStatistics(Page page) {
 
         // Collect all statistics for the page, except from blocks inside tables, as tables will always be added to BodyTextFrame.

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

@@ -42,7 +42,9 @@ public class SectionsBuilderService {
                 if (prev != null && current.getClassification().startsWith("H ") || !document.isHeadlines()) {
                     Paragraph chunkBlock = buildTextBlock(chunkWords, lastHeadline);
                     chunkBlock.setHeadline(lastHeadline);
-                    lastHeadline = current.getText();
+                    if(document.isHeadlines()) {
+                        lastHeadline = current.getText();
+                    }
                     chunkBlockList.add(chunkBlock);
                     chunkWords = new ArrayList<>();
                     if (CollectionUtils.isNotEmpty(chunkBlock.getTables())) {

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/tableextraction/model/Cell.java

@@ -5,6 +5,8 @@ import java.util.ArrayList;
 import java.util.List;
 
 import com.iqser.red.service.redaction.v1.server.classification.model.TextBlock;
+import com.iqser.red.service.redaction.v1.server.parsing.model.TextPositionSequence;
+import com.iqser.red.service.redaction.v1.server.redaction.utils.TextNormalizationUtilities;
 
 import lombok.Data;
 import lombok.EqualsAndHashCode;
@@ -20,10 +22,10 @@ public class Cell extends Rectangle {
 
     private boolean isHeaderCell;
 
+
     public Cell(Point2D topLeft, Point2D bottomRight) {
 
-        super((float) topLeft.getY(), (float) topLeft.getX(), (float) (bottomRight.getX() - topLeft.getX()),
-                (float) (bottomRight
+        super((float) topLeft.getY(), (float) topLeft.getX(), (float) (bottomRight.getX() - topLeft.getX()), (float) (bottomRight
                 .getY() - topLeft.getY()));
     }
 
@@ -33,4 +35,34 @@ public class Cell extends Rectangle {
         textBlocks.add(textBlock);
     }
 
-}
+
+    @Override
+    public String toString() {
+
+        StringBuilder sb = new StringBuilder();
+
+        TextPositionSequence previous = null;
+        for (TextBlock textBlock : textBlocks) {
+
+
+            for (TextPositionSequence word : textBlock.getSequences()) {
+
+                if (previous != null) {
+                    if (Math.abs(previous.getY1() - word.getY1()) > word.getTextHeight()) {
+                        sb.append('\n');
+                    } else {
+                        sb.append(' ');
+                    }
+                }
+                sb.append(word.toString());
+                previous = word;
+            }
+        }
+
+        return TextNormalizationUtilities.removeHyphenLineBreaks(sb.toString())
+                .replaceAll("\n", " ")
+                .replaceAll(" {2}", " ");
+    }
+
+}
+

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/tableextraction/service/RulingCleaningService.java

@@ -18,9 +18,9 @@ import com.iqser.red.service.redaction.v1.server.tableextraction.utils.Utils;
 @Service
 public class RulingCleaningService {
 
-    public CleanRulings getCleanRulings(List<Ruling> rulings, float minCharWidth, float minCharHeight){
+    public CleanRulings getCleanRulings(List<Ruling> rulings, float maxCharWidth, float maxCharHeight){
         if (!rulings.isEmpty()) {
-            snapPoints(rulings, minCharWidth , minCharHeight);
+            snapPoints(rulings, maxCharWidth , maxCharHeight);
         }
 
         List<Ruling> vrs = new ArrayList<>();

redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/visualization/service/AnnotationHighlightService.java

@@ -2,7 +2,11 @@ package com.iqser.red.service.redaction.v1.server.visualization.service;
 
 import java.awt.Color;
 import java.io.IOException;
+import java.util.ArrayList;
+import java.util.HashSet;
 import java.util.List;
+import java.util.Set;
+import java.util.stream.Collectors;
 
 import org.apache.commons.collections4.CollectionUtils;
 import org.apache.pdfbox.pdmodel.PDDocument;
@@ -14,6 +18,7 @@ import org.apache.pdfbox.pdmodel.graphics.color.PDColor;
 import org.apache.pdfbox.pdmodel.graphics.color.PDDeviceRGB;
 import org.apache.pdfbox.pdmodel.interactive.annotation.PDAnnotation;
 import org.apache.pdfbox.pdmodel.interactive.annotation.PDAnnotationTextMarkup;
+import org.apache.pdfbox.text.TextPosition;
 import org.springframework.stereotype.Service;
 
 import com.iqser.red.service.redaction.v1.model.ManualRedactionEntry;
@@ -33,9 +38,7 @@ import com.iqser.red.service.redaction.v1.server.tableextraction.model.Cell;
 import com.iqser.red.service.redaction.v1.server.tableextraction.model.Table;
 
 import lombok.RequiredArgsConstructor;
-import lombok.extern.slf4j.Slf4j;
 
-@Slf4j
 @Service
 @RequiredArgsConstructor
 public class AnnotationHighlightService {
@@ -43,7 +46,10 @@ public class AnnotationHighlightService {
     private final DictionaryService dictionaryService;
 
 
-    public void highlight(PDDocument document, Document classifiedDoc, boolean flatRedaction, ManualRedactions manualRedactions) throws IOException {
+    public void highlight(PDDocument document, Document classifiedDoc, boolean flatRedaction,
+                          ManualRedactions manualRedactions) throws IOException {
+
+        Set<Integer> manualRedactionPages = getManualRedactionPages(manualRedactions);
 
         for (int page = 1; page <= document.getNumberOfPages(); page++) {
 
@@ -51,17 +57,36 @@ public class AnnotationHighlightService {
 
             drawSectionFrames(document, classifiedDoc, flatRedaction, pdPage, page);
 
-            if (classifiedDoc.getEntities().get(page) == null) {
-                continue;
+            if (classifiedDoc.getEntities().get(page) != null) {
+                addAnnotations(pdPage, classifiedDoc, flatRedaction, manualRedactions, page);
             }
 
-            addAnnotations(pdPage, classifiedDoc, flatRedaction, manualRedactions, page);
-            addManualAnnotations(pdPage, classifiedDoc, manualRedactions, page);
+            if (manualRedactionPages.contains(page)) {
+                addManualAnnotations(pdPage, classifiedDoc, manualRedactions, page);
+            }
         }
     }
 
 
-    private void addAnnotations(PDPage pdPage, Document classifiedDoc, boolean flatRedaction, ManualRedactions manualRedactions, int page) throws IOException {
+    private Set<Integer> getManualRedactionPages(ManualRedactions manualRedactions) {
+
+        Set<Integer> manualRedactionPages = new HashSet<>();
+
+        if (manualRedactions == null) {
+            return manualRedactionPages;
+        }
+
+        manualRedactions.getEntriesToAdd().forEach(entry -> {
+            entry.getPositions().forEach(pos -> {
+                manualRedactionPages.add(pos.getPage());
+            });
+        });
+        return manualRedactionPages;
+    }
+
+
+    private void addAnnotations(PDPage pdPage, Document classifiedDoc, boolean flatRedaction,
+                                ManualRedactions manualRedactions, int page) throws IOException {
 
         List<PDAnnotation> annotations = pdPage.getAnnotations();
 
@@ -77,16 +102,23 @@ public class AnnotationHighlightService {
 
                 if (manualRedactions != null && manualRedactions.getIdsToRemove()
                         .contains(entityPositionSequence.getId())) {
+                    String manualOverrideReason = entity.getRedactionReason() + ", removed by manual override";
                     entity.setRedaction(false);
-                    entity.setRedactionReason(entity.getRedactionReason() + ", removed by manual override");
+                    entity.setRedactionReason(manualOverrideReason);
+                    redactionLogEntry.setRedacted(false);
+                    redactionLogEntry.setReason(manualOverrideReason);
                     redactionLogEntry.setManual(true);
                 }
 
-                for (TextPositionSequence textPositions : entityPositionSequence.getSequences()) {
-
-                    Rectangle rectangle = textPositions.getRectangle();
-                    redactionLogEntry.getPositions().add(rectangle);
-                    annotations.add(createAnnotation(rectangle, entityPositionSequence.getId(), createAnnotationContent(entity), getColor(entity), !flatRedaction && !isHint(entity)));
+                if (CollectionUtils.isNotEmpty(entityPositionSequence.getSequences())) {
+                    for (Rectangle rectangle : getRectanglesPerLine(entityPositionSequence.getSequences()
+                            .stream()
+                            .flatMap(seq -> seq.getTextPositions().stream())
+                            .collect(Collectors.toList()), page)) {
+                        redactionLogEntry.getPositions().add(rectangle);
+                        annotations.add(createAnnotation(rectangle, entityPositionSequence.getId(),
+                                createAnnotationContent(entity), getColor(entity), !flatRedaction && !isHint(entity)));
+                    }
                 }
                 redactionLogEntry.setId(entityPositionSequence.getId());
             }
@@ -95,7 +127,33 @@ public class AnnotationHighlightService {
     }
 
 
-    private void addManualAnnotations(PDPage pdPage, Document classifiedDoc, ManualRedactions manualRedactions, int page) throws IOException {
+    private List<Rectangle> getRectanglesPerLine(List<TextPosition> textPositions, int page) {
+
+        List<Rectangle> rectangles = new ArrayList<>();
+        if (textPositions.size() == 1) {
+            rectangles.add(new TextPositionSequence(textPositions, page).getRectangle());
+        } else {
+            float y = textPositions.get(0).getYDirAdj();
+            int startIndex = 0;
+            for (int i = 1; i < textPositions.size(); i++) {
+                float yDirAdj = textPositions.get(i).getYDirAdj();
+                if (yDirAdj != y) {
+                    rectangles.add(new TextPositionSequence(textPositions.subList(startIndex, i), page).getRectangle());
+                    y = yDirAdj;
+                    startIndex = i;
+                }
+            }
+            if (startIndex != textPositions.size() - 1) {
+                rectangles.add(new TextPositionSequence(textPositions.subList(startIndex, textPositions.size()), page).getRectangle());
+            }
+        }
+
+        return rectangles;
+    }
+
+
+    private void addManualAnnotations(PDPage pdPage, Document classifiedDoc, ManualRedactions manualRedactions,
+                                      int page) throws IOException {
 
         if (manualRedactions == null) {
             return;
@@ -107,28 +165,34 @@ public class AnnotationHighlightService {
 
             String id = IdBuilder.buildId(manualRedactionEntry);
 
-            RedactionLogEntry redactionLogEntry = createRedactionLogEntry(manualRedactionEntry);
+            RedactionLogEntry redactionLogEntry = createRedactionLogEntry(manualRedactionEntry, id);
 
+            boolean foundOnPage = false;
             for (Rectangle rectangle : manualRedactionEntry.getPositions()) {
 
                 if (page != rectangle.getPage()) {
                     continue;
                 }
 
-                PDAnnotationTextMarkup highlight = createAnnotation(rectangle, id, createAnnotationContent(manualRedactionEntry), getColor(manualRedactionEntry
-                        .getType()), true);
+                foundOnPage = true;
+
+                PDAnnotationTextMarkup highlight = createAnnotation(rectangle, id,
+                        createAnnotationContent(manualRedactionEntry), getColor(manualRedactionEntry.getType()), true);
                 annotations.add(highlight);
 
                 redactionLogEntry.getPositions().add(rectangle);
             }
-            classifiedDoc.getRedactionLogEntities().add(redactionLogEntry);
+            if (foundOnPage) {
+                classifiedDoc.getRedactionLogEntities().add(redactionLogEntry);
+            }
         }
     }
 
 
-    private RedactionLogEntry createRedactionLogEntry(ManualRedactionEntry manualRedactionEntry) {
+    private RedactionLogEntry createRedactionLogEntry(ManualRedactionEntry manualRedactionEntry, String id) {
 
         return RedactionLogEntry.builder()
+                .id(id)
                 .color(getColor(manualRedactionEntry.getType()))
                 .reason(manualRedactionEntry.getReason())
                 .value(manualRedactionEntry.getValue())
@@ -157,7 +221,8 @@ public class AnnotationHighlightService {
     }
 
 
-    private PDAnnotationTextMarkup createAnnotation(Rectangle rectangle, String id, String content, float[] color, boolean popup) {
+    private PDAnnotationTextMarkup createAnnotation(Rectangle rectangle, String id, String content, float[] color,
+                                                    boolean popup) {
 
         PDAnnotationTextMarkup annotation = new PDAnnotationTextMarkup(PDAnnotationTextMarkup.SUB_TYPE_HIGHLIGHT);
         annotation.constructAppearances();
@@ -175,25 +240,15 @@ public class AnnotationHighlightService {
 
     private String createAnnotationContent(Entity entity) {
 
-        return new StringBuilder().append("\nRule ")
-                .append(entity.getMatchedRule())
-                .append(" matched")
-                .append("\n\n")
-                .append(entity.getRedactionReason())
-                .append("\n\nIn Section : \"")
-                .append(entity.getHeadline())
-                .append("\"")
-                .toString();
+        return "\nRule " + entity.getMatchedRule() + " matched" + "\n\n" + entity.getRedactionReason() + "\n\nIn " +
+                "Section : \"" + entity
+                .getHeadline() + "\"";
     }
 
 
     private String createAnnotationContent(ManualRedactionEntry entry) {
 
-        return new StringBuilder().append("\nManual Redaction")
-                .append("\n\nIn Section : \"")
-                .append(entry.getSection())
-                .append("\"")
-                .toString();
+        return "\nManual Redaction" + "\n\nIn Section : \"" + entry.getSection() + "\"";
     }
 
 
@@ -213,7 +268,8 @@ public class AnnotationHighlightService {
         // quadPoints is array of x,y coordinates in Z-like order (top-left, top-right, bottom-left,bottom-right)
         // of the area to be highlighted
         return new float[]{rectangle.getTopLeft().getX(), rectangle.getTopLeft().getY(), rectangle.getTopLeft()
-                .getX() + rectangle.getWidth(), rectangle.getTopLeft().getY(), rectangle.getTopLeft().getX(), rectangle.getTopLeft()
+                .getX() + rectangle.getWidth(), rectangle.getTopLeft().getY(), rectangle.getTopLeft().getX(),
+                rectangle.getTopLeft()
                 .getY() + rectangle.getHeight(), rectangle.getTopLeft()
                 .getX() + rectangle.getWidth(), rectangle.getTopLeft().getY() + rectangle.getHeight()};
     }
@@ -224,10 +280,7 @@ public class AnnotationHighlightService {
         if (!entity.isRedaction()) {
             return false;
         }
-        if (isHint(entity)) {
-            return false;
-        }
-        return true;
+        return !isHint(entity);
     }
 
 
@@ -255,20 +308,19 @@ public class AnnotationHighlightService {
     private boolean isHint(Entity entity) {
 
         List<String> hintTypes = dictionaryService.getHintTypes();
-        if (CollectionUtils.isNotEmpty(hintTypes) && hintTypes.contains(entity.getType())) {
-            return true;
-        }
-        return false;
+        return CollectionUtils.isNotEmpty(hintTypes) && hintTypes.contains(entity.getType());
     }
 
 
-    private void drawSectionFrames(PDDocument document, Document classifiedDoc, boolean flatRedaction, PDPage pdPage, int page) throws IOException {
+    private void drawSectionFrames(PDDocument document, Document classifiedDoc, boolean flatRedaction, PDPage pdPage,
+                                   int page) throws IOException {
 
         if (flatRedaction) {
             return;
         }
 
-        PDPageContentStream contentStream = new PDPageContentStream(document, pdPage, PDPageContentStream.AppendMode.APPEND, true);
+        PDPageContentStream contentStream = new PDPageContentStream(document, pdPage,
+                PDPageContentStream.AppendMode.APPEND, true);
         for (Paragraph paragraph : classifiedDoc.getParagraphs()) {
 
             for (int i = 0; i <= paragraph.getPageBlocks().size() - 1; i++) {
@@ -317,7 +369,8 @@ public class AnnotationHighlightService {
                 if (cell != null) {
                     contentStream.setLineWidth(0.5f);
                     contentStream.setStrokingColor(Color.CYAN);
-                    contentStream.addRect((float) cell.getX(), (float) cell.getY(), (float) cell.getWidth(), (float) cell
+                    contentStream.addRect((float) cell.getX(), (float) cell.getY(), (float) cell.getWidth(),
+                            (float) cell
                             .getHeight());
                     contentStream.stroke();
                 }

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

@@ -1,5 +1,6 @@
 package com.iqser.red.service.redaction.v1.server;
 
+import static org.assertj.core.api.Assertions.assertThat;
 import static org.mockito.Mockito.when;
 import static org.springframework.boot.test.context.SpringBootTest.WebEnvironment.DEFINED_PORT;
 
@@ -269,7 +270,7 @@ public class RedactionIntegrationTest {
 
         System.out.println("redactionTest");
         long start = System.currentTimeMillis();
-        ClassPathResource pdfFileResource = new ClassPathResource("files/Trinexapac/96 Trinexapac-ethyl_RAR_09_Volume_3CA_B-7_2018-02-23.pdf");
+        ClassPathResource pdfFileResource = new ClassPathResource("files/Fludioxonil/50 Fludioxonil_RAR_01_Volume_1_2018-02-21.pdf");
 
         RedactionRequest request = RedactionRequest.builder()
                 .document(IOUtils.toByteArray(pdfFileResource.getInputStream()))
@@ -388,8 +389,7 @@ public class RedactionIntegrationTest {
     public void htmlTablesTest() throws IOException {
 
         System.out.println("htmlTablesTest");
-        ClassPathResource pdfFileResource = new ClassPathResource("files/Fludioxonil/51 " +
-                "Fludioxonil_RAR_02_Volume_2_2018-02-21.pdf");
+        ClassPathResource pdfFileResource = new ClassPathResource("files/Metolachlor/S-Metolachlor_RAR_02_Volume_2_2018-09-06.pdf");
 
         RedactionRequest request = RedactionRequest.builder()
                 .document(IOUtils.toByteArray(pdfFileResource.getInputStream()))
@@ -422,6 +422,27 @@ public class RedactionIntegrationTest {
     }
 
 
+    @Test
+    public void phantomCellsDocumentTest() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Phantom Cells.pdf");
+
+        RedactionRequest request = RedactionRequest.builder()
+                .document(IOUtils.toByteArray(pdfFileResource.getInputStream()))
+                .build();
+        request.setFlatRedaction(false);
+
+        RedactionResult result = redactionController.redact(request);
+
+        result.getRedactionLog().getRedactionLogEntry().forEach(entry -> {
+            if(!entry.isHint()){
+                assertThat(entry.getReason()).isEqualTo("Not redacted because row is not a vertebrate study");
+            }
+        });
+    }
+
+
+
     private static String loadFromClassPath(String path) {
 
         URL resource = ResourceLoader.class.getClassLoader().getResource(path);

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

@@ -275,6 +275,27 @@ public class EntityRedactionServiceTest {
             assertThat(classifiedDoc.getEntities().get(1).stream().filter(entity -> entity.getMatchedRule() == 9).count()).isEqualTo(8);
             assertThat(classifiedDoc.getEntities().get(2).stream().filter(entity -> entity.getMatchedRule() == 9).count()).isEqualTo(4);
         }
+
+        pdfFileResource = new ClassPathResource("files/Minimal Examples/Header Propagation2.pdf");
+
+        dictionaryResponse = DictionaryResponse.builder()
+                .entries(Arrays.asList("Tribolet, R.", "Muir, G.", "Kühne-Thu, H.", "Close, C."))
+                .build();
+
+        when(dictionaryClient.getVersion()).thenReturn(DICTIONARY_VERSION.incrementAndGet());
+        when(dictionaryClient.getDictionaryForType(NAME_CODE)).thenReturn(dictionaryResponse);
+        addressResponse = DictionaryResponse.builder()
+                .entries(Collections.singletonList("Novartis Crop Protection AG, Basel, Switzerland"))
+                .build();
+        when(dictionaryClient.getDictionaryForType(ADDRESS_CODE)).thenReturn(addressResponse);
+
+        try (PDDocument pdDocument = PDDocument.load(pdfFileResource.getInputStream())) {
+            Document classifiedDoc = pdfSegmentationService.parseDocument(pdDocument);
+            entityRedactionService.processDocument(classifiedDoc, null);
+            assertThat(classifiedDoc.getEntities()).hasSize(1); // one page
+            assertThat(classifiedDoc.getEntities().get(1).stream().filter(entity -> entity.getMatchedRule() == 9).count()).isEqualTo(3);
+            assertThat(classifiedDoc.getEntities().get(1).stream().filter(entity -> entity.getMatchedRule() == 8).count()).isEqualTo(8);
+        }
     }
 
 
@@ -312,7 +333,7 @@ public class EntityRedactionServiceTest {
                 "global Section section\n" +
                 "rule \"8: Not redacted because Vertebrate Study = N\"\n" +
                 "    when\n" +
-                "        Section(rowEquals(\"Vertebrate study Y/N\", \"N\"))\n" +
+                "        Section(rowEquals(\"Vertebrate study Y/N\", \"N\") || rowEquals(\"Vertebrate study Y/N\", \"No\"))\n" +
                 "    then\n" +
                 "        section.redactNot(\"name\", 8, \"Not redacted because row is not a vertebrate study\");\n" +
                 "        section.redactNot(\"address\", 8, \"Not redacted because row is not a vertebrate study\");\n" +
@@ -320,7 +341,8 @@ public class EntityRedactionServiceTest {
                 "    end\n" +
                 "rule \"9: Redact Authors and Addresses in Reference Table, if it is a Vertebrate study\"\n" +
                 "    when\n" +
-                "        Section(rowEquals(\"Vertebrate study Y/N\", \"Y\"))\n" +
+                "        Section(rowEquals(\"Vertebrate study Y/N\", \"Y\") || rowEquals(\"Vertebrate study Y/N\", " +
+                "\"Yes\"))\n" +
                 "    then\n" +
                 "        section.redact(\"name\", 9, \"Redacted because row is a vertebrate study\");\n" +
                 "        section.redact(\"address\", 9, \"Redacted because row is a vertebrate study\");\n" +

redaction-service-v1/redaction-service-server-v1/src/test/resources/dictionaries/addresses.txt

@@ -227,6 +227,7 @@ CIBA-GEIGY Limited, Toxicology Services, Short-term Toxicology, 4332 Stein, Swit
 CIBA-GEIGY Ltd. CH-4002 Basle, Switzerland
 CIBA-GEIGY Ltd., Product Safety, Ecotoxicology, CH-4002 Basel, Switzerland
 CIBA-GEIGY Ltd., Switzerland
+CibaGeigy Ltd., CH-4002, Basel, Switzerland
 CIP Chemisches Institut Pforzheim GmbH, Pforzheim, Germany
 CL PHARMA AG
 CL PHARMA AG, Linz, Austria
@@ -1563,4 +1564,36 @@ Zyma SA
 Zyma SA, Nyon, Switzerland
 “Bayer CropScience AG” Monheim
 “Biologische Bundesanstalt (BBA)”, Berlin-Dahlem
-“W. Neudorff GmbH KG”, An der Mühle 3, D- 31860 Emmertal
+“W. Neudorff GmbH KG”, An der Mühle 3, D- 31860 Emmertal
+Syngenta Ltd Jealott’s Hill Int. Research Centre Bracknell, Berkshire, RG42 6EY, United Kingdom
+MITOX Consultants Science Park 408, 1098XH Amsterdam, NL
+Eurofins Agrosciences Services Eco-Chem GmbH, Eutinger Str. 24, 75233 Niefern-Öschelbronn, Germany
+Fraunhofer Institute for Molecular Biology and Applied Ecology (IME), 57392 Schmallenberg, Germany
+Syngenta Crop Protection AGRCC Cytotest Cell Research GmbH, Germany
+BASF SE, Ludwigshafen, Germany
+AGRCC Ltd., Switzerland
+AGWIL Research Lab. Inc., USA
+Charles River Laboratories, UK
+Ciba-Geigy Corporation Environmental Health Centre, Farmington
+Syngenta Crop Protection, Inc., Greensboro, United States
+Experimental Toxicology and Ecology, BASF SE
+Wildlife International, Ltd. 8598 Commerce Drive Easton, MD 21601 USA
+Küberich, Wiesentheid / Geesdorf, Germany
+CiToxLAB Hungary Ltd. H8200 Veszprém, Szabadságpuszta, Hungary
+Forellenzucht Trostadt GbR, Trostadt, Germany
+Osage Catfisheries Inc., Osage Beach, MO, USA
+Fischzucht Rhönforelle, Gersfeld, Germany
+Aquatic Biosystems, Ft. Collins, Colorado, USA
+West Aquarium GmbH, 37431 Bad Lauterberg, Germany
+Aquatic BioSystems, Fort Collins, Colorado, USA
+“U.S. Environmental Research Laboratory”, Duluth, Minnesota, USA
+Continuous laboratory cultures
+Wildlife International, Ltd. 8598 Commerce Drive Easton, MD 21601 USA
+Osage Catfisheries, Inc., Osage Beach, MO, USA
+Osage catfisheries, Osage Beach (MO), USA
+Ciba-Geigy Corporation, Environmental Health Centre, Farmington
+Novartis Crop Protection Inc., Greensboro, United States
+Syngenta, Jealott’s Hill International Research Station, UK
+Hillandale Farms, Kenansville, Florida, USA
+Jealott’s Hill International Research Station, UK
+L & C Dairy, Avon Park, Florida, UK

redaction-service-v1/redaction-service-server-v1/src/test/resources/dictionaries/names.txt

@@ -5290,6 +5290,7 @@ Naeb O.
 Nagai K.
 Nagashima Yoshikazu
 Nagata H.|Tanaka T.
+Nagra, B
 Nagra B.S
 Nagra B.|Kingdom N.
 Nagy K
@@ -5640,6 +5641,7 @@ Pedersen Carol
 Pedersen Carol|DuCharme Darlene
 Pedersen C|Ducharme D.
 Pederson C.A.
+Pederson
 Peffer
 Peffer R
 Peffer R.
@@ -6249,6 +6251,7 @@ Roth M.
 Roth Markus
 Roth RN|Heilman RD
 Roth, M.
+Roth
 Rotroff
 Rouaud, J. L.
 Roussel C
@@ -6848,7 +6851,7 @@ Sokolowshi Andrea
 Sokolowski A
 Sokolowski A.
 Sokolowski Andrea
-Sokolowski,
+Sokolowski
 Sole
 Sole C
 Sole, C.
@@ -7248,6 +7251,7 @@ Thanei, P.
 Tharp BE
 Thede B.
 Theophilidis G
+Thevenaz
 Thevenaz P
 Thevenaz P.
 Thevenaz Ph.
@@ -8193,4 +8197,113 @@ Zoriki Hosomi R.
 Zoriki Hosomi Rosana
 Zuberer D
 Zubrod J
-Zwicker R.E.
+Zwicker R.E.
+Ott
+Argus et al
+Mihail
+Clapp
+Murfitt R
+Yang
+Swarbrick
+Hertl
+Breitwieser
+Seville, A
+Trumper, C.
+Gasso-Brown, D.
+Váliczkó, É
+Váliczkó, É.
+Durando, J.
+Tarcai, Z.
+Pooles, A
+Roig, J.
+Schäfers, C.
+Nixon, W.
+Zok, S.
+Durjava M.K. et al
+Jatzek J
+Salinas E
+Sousa J. V
+Caferella, M.A.
+Salina, E.
+Walraven J
+Kaspers et al
+Schneider et al
+Lowrie C
+Macdougall J
+Roberts S
+Madrid, S.O.
+Maynard, M.S.
+Ray, W.
+Ray, W.J.
+Váliczkó É
+Tarcai Z
+Orovecz B
+Kaspers
+Tarcai Z.
+Gizler, A
+Whitlow
+Johnson et al
+Bryden et al
+Váliczkó
+Höger
+Dorgerloh
+Schäfers C
+Caferella M.A
+Salinas, E
+Zok S
+Tummon OJ
+Eckstein CL
+Francis PD
+Ott M
+Richeux F
+Rajsekhar PV
+Webbley
+O’Hagan
+MacDonald
+Jewkes
+Harris
+Tomlinson
+Tomlinson et al
+Nagy
+Török-Bathó
+Hargitai
+Shearer
+Robertson
+Strepka
+Hackford
+Blunt
+Sieber
+Elcombe
+Haines
+Kang
+Sayers
+Harris J
+Baxter A
+Macdonald M
+PetusArpasy M
+TorokBatho M.
+van Ravenzwaa y B
+Cords S
+Dymarkows ka K.
+Bahig M
+Pekari K
+Bercz J
+Russell L
+Kitchen K
+Miyagawa M.
+Sasaki Y
+Yin D
+Huff J
+Blackburn K
+Carlson G
+Pereira M
+Stoner G
+Zelenak V
+Valiczko E
+Tomlinso n J
+Trumper C
+Senciuc M
+Vance C.
+Przeorska J.
+Melville S.
+Wicksted G.

redaction-service-v1/redaction-service-server-v1/src/test/resources/drools/rules.drl

@@ -51,7 +51,7 @@ rule "5: Do not redact in guideline sections"
 
 rule "6: Redact contact information if applicant is found"
 	when
-		eval(section.headlineContainsWord("applicant") || section.getText().contains("Applicant"));
+		eval(section.headlineContainsWord("applicant") || section.getText().contains("Applicant") || section.headlineContainsWord("Primary contact") || section.headlineContainsWord("Alternative contact"));
 	then
 		section.redactLineAfter("Name:", "address", 6, "Applicant information was found");
 		section.redactBetween("Address:", "Contact", "address", 6, "Applicant information was found");
@@ -62,14 +62,20 @@ rule "6: Redact contact information if applicant is found"
 		section.redactLineAfter("Tel:", "address", 6, "Applicant information was found");
 		section.redactLineAfter("E-mail:", "address", 6, "Applicant information was found");
 		section.redactLineAfter("Email:", "address", 6, "Applicant information was found");
+		section.redactLineAfter("e-mail:", "address", 6, "Applicant information was found");
+		section.redactLineAfter("E-mail address:", "address", 6, "Applicant information was found");
 		section.redactLineAfter("Contact:", "address", 6, "Applicant information was found");
+		section.redactLineAfter("Alternative contact:", "address", 6, "Applicant information was found");
 		section.redactLineAfter("Telephone number:", "address", 6, "Applicant information was found");
+		section.redactLineAfter("Telephone No:", "address", 6, "Applicant information was found");
 		section.redactLineAfter("Fax number:", "address", 6, "Applicant information was found");
 		section.redactLineAfter("Telephone:", "address", 6, "Applicant information was found");
+		section.redactLineAfter("Company:", "address", 6, "Applicant information was found");
 		section.redactBetween("No:", "Fax", "address", 6, "Applicant information was found");
 		section.redactBetween("Contact:", "Tel.:", "address", 6, "Applicant information was found");
 	end
 
+
 rule "7: Redact contact information if Producer is found"
 	when
 		eval(section.getText().toLowerCase().contains("producer of the plant protection") || section.getText().toLowerCase().contains("producer of the active substance") || section.getText().contains("Manufacturer of the active substance") || section.getText().contains("Manufacturer:") || section.getText().contains("Producer or producers of the active substance"));
@@ -93,9 +99,9 @@ rule "7: Redact contact information if Producer is found"
 
 rule "8: Not redacted because Vertebrate Study = N"
     when
-        Section(rowEquals("Vertebrate study Y/N", "N"))
+        Section(rowEquals("Vertebrate study Y/N", "N") || rowEquals("Vertebrate study Y/N", "No"))
     then
-        section.redactNot("name", 8, "Not redacted because row is not a vertebrate study");
+        section.redactNotCell("Author(s)", 8, "name", "Not redacted because row is not a vertebrate study");
         section.redactNot("address", 8, "Not redacted because row is not a vertebrate study");
         section.highlightCell("Vertebrate study Y/N", 8, "hint_only");
     end
@@ -112,9 +118,9 @@ rule "9: Redact if must redact entry is found"
 
 rule "10: Redact Authors and Addresses in Reference Table if it is a Vertebrate study"
     when
-        Section(rowEquals("Vertebrate study Y/N", "Y"))
+        Section(rowEquals("Vertebrate study Y/N", "Y") || rowEquals("Vertebrate study Y/N", "Yes"))
     then
-        section.redact("name", 10, "Redacted because row is a vertebrate study");
+        section.redactCell("Author(s)", 10, "name", "Redacted because row is a vertebrate study");
         section.redact("address", 10, "Redacted because row is a vertebrate study");
         section.highlightCell("Vertebrate study Y/N", 10, "must_redact");
     end