Marmelator/redaction-service
1
0
Fork 0
Code Issues Pull Requests 14 Actions Packages Projects Releases 622 Wiki Activity

RED-3067 Bugfix to show redactions on rotated pages correctly

Browse Source
This commit is contained in:
Philipp Schramm 2021-12-16 12:43:17 +01:00
parent e8dce509f3
commit f7fbe8f0f6
2 changed files with 50 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

82
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/parsing/model/TextPositionSequence.java
View File

@ -1,22 +1,24 @@
package com.iqser.red.service.redaction.v1.server.parsing.model; package com.iqser.red.service.redaction.v1.server.parsing.model;
import static java.util.stream.Collectors.toSet;
import com.fasterxml.jackson.annotation.JsonIgnore;
import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
import com.iqser.red.service.redaction.v1.model.Point;
import com.iqser.red.service.redaction.v1.model.Rectangle;
import lombok.Data;
import lombok.NoArgsConstructor;
import org.apache.pdfbox.text.TextPosition;
import java.util.ArrayList; import java.util.ArrayList;
import java.util.List; import java.util.List;
import java.util.stream.Collectors; import java.util.stream.Collectors;
import org.apache.pdfbox.text.TextPosition;
import com.fasterxml.jackson.annotation.JsonIgnore;
import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
import com.iqser.red.service.redaction.v1.model.Point;
import com.iqser.red.service.redaction.v1.model.Rectangle;
import lombok.Data;
import lombok.NoArgsConstructor;
import lombok.extern.slf4j.Slf4j;
@Slf4j
@Data @Data
@NoArgsConstructor @NoArgsConstructor
@JsonIgnoreProperties({ "empty" }) @JsonIgnoreProperties({"empty"})
public class TextPositionSequence implements CharSequence { public class TextPositionSequence implements CharSequence {
private int page; private int page;
@ -25,12 +27,15 @@ public class TextPositionSequence implements CharSequence {
private float x1; private float x1;
private float x2; private float x2;
public TextPositionSequence(int page) { public TextPositionSequence(int page) {
this.page = page; this.page = page;
} }
public static TextPositionSequence fromData(List<RedTextPosition> textPositions, int page) { public static TextPositionSequence fromData(List<RedTextPosition> textPositions, int page) {
var textPositionSequence = new TextPositionSequence(); var textPositionSequence = new TextPositionSequence();
textPositionSequence.textPositions = textPositions; textPositionSequence.textPositions = textPositions;
textPositionSequence.page = page; textPositionSequence.page = page;
@ -46,9 +51,6 @@ public class TextPositionSequence implements CharSequence {
} }
@Override @Override
public int length() { public int length() {
@ -131,6 +133,7 @@ public class TextPositionSequence implements CharSequence {
} }
} }
@JsonIgnore @JsonIgnore
public float getRotationAdjustedY() { public float getRotationAdjustedY() {
@ -190,10 +193,8 @@ public class TextPositionSequence implements CharSequence {
@JsonIgnore @JsonIgnore
public String getFont() { public String getFont() {
return textPositions.get(0).getFontName()
.toLowerCase() return textPositions.get(0).getFontName().toLowerCase().replaceAll(",bold", "").replaceAll(",italic", "");
.replaceAll(",bold", "")
.replaceAll(",italic", "");
} }
@ -214,27 +215,33 @@ public class TextPositionSequence implements CharSequence {
} }
@JsonIgnore @JsonIgnore
public float getFontSize() { public float getFontSize() {
return textPositions.get(0).getFontSizeInPt(); return textPositions.get(0).getFontSizeInPt();
} }
@JsonIgnore @JsonIgnore
public float getSpaceWidth() { public float getSpaceWidth() {
return textPositions.get(0).getWidthOfSpace(); return textPositions.get(0).getWidthOfSpace();
} }
@JsonIgnore @JsonIgnore
public int getRotation() { public int getRotation() {
return textPositions.get(0).getRotation(); return textPositions.get(0).getRotation();
} }
@JsonIgnore @JsonIgnore
public Rectangle getRectangle() { public Rectangle getRectangle() {
log.debug("Page: '{}', Word: '{}', Rotation: '{}'", page, toString(), textPositions.get(0).getRotation());
float height = getTextHeight(); float height = getTextHeight();
float posXInit = getX1(); float posXInit = getX1();
@ -246,36 +253,45 @@ public class TextPositionSequence implements CharSequence {
posXEnd = textPositions.get(0).getYDirAdj() + 2; posXEnd = textPositions.get(0).getYDirAdj() + 2;
posYInit = getY1(); posYInit = getY1();
posYEnd = textPositions.get(textPositions.size() - 1).getXDirAdj() - height + 4; posYEnd = textPositions.get(textPositions.size() - 1).getXDirAdj() - height + 4;
} else if (textPositions.get(0).getRotation() == 270) {
} else if (textPositions.get(0).getRotation() == 270 && textPositions.size() > 1) {
posYInit = textPositions.get(0).getPageHeight() - getX1(); posYInit = textPositions.get(0).getPageHeight() - getX1();
posYEnd = textPositions.get(0).getPageHeight() - getX2() - textPositions.get(0) posYEnd = textPositions.get(0).getPageHeight() - getX2() - textPositions.get(0)
.getWidth() - textPositions.get(textPositions.size() - 1).getWidth() - 1; .getWidth() - textPositions.get(textPositions.size() - 1).getWidth() - 1;
posXInit = textPositions.get(0).getPageWidth() - textPositions.get(0).getYDirAdj() - 2; posXInit = textPositions.get(0).getPageWidth() - textPositions.get(0).getYDirAdj() - 2;
posXEnd = textPositions.get(0).getPageWidth() - textPositions.get(textPositions.size() - 1) posXEnd = textPositions.get(0).getPageWidth() - textPositions.get(textPositions.size() - 1)
.getYDirAdj() + height; .getYDirAdj() + height;
} else if(textPositions.get(0).getRotation() == 0 && textPositions.get(0).getDir() == 270f) {
} else if (textPositions.get(0).getRotation() == 270 && textPositions.size() == 1) {
posYInit = textPositions.get(0).getPageHeight() - textPositions.get(0).getYDirAdj() - 2;
posYEnd = posYInit + 1;
posXInit = textPositions.get(0).getXDirAdj();
posXEnd = posXInit + textPositions.get(0).getWidthDirAdj() + 0.1f;
} else if (textPositions.get(0).getRotation() == 0 && textPositions.get(0).getDir() == 270f) {
posYInit = textPositions.get(0).getPageHeight() - getX1(); posYInit = textPositions.get(0).getPageHeight() - getX1();
posYEnd = textPositions.get(0).getPageHeight() - getX2() - textPositions.get(0) posYEnd = textPositions.get(0).getPageHeight() - getX2() - textPositions.get(0)
.getWidthDirAdj() - textPositions.get(textPositions.size() - 1).getWidthDirAdj() - 3; .getWidthDirAdj() - textPositions.get(textPositions.size() - 1).getWidthDirAdj() - 3;
posXInit = textPositions.get(0).getPageWidth() - textPositions.get(0).getYDirAdj() - 2; posXInit = textPositions.get(0).getPageWidth() - textPositions.get(0).getYDirAdj() - 2;
posXEnd = textPositions.get(0).getPageWidth() - textPositions.get(textPositions.size() - 1) posXEnd = textPositions.get(0).getPageWidth() - textPositions.get(textPositions.size() - 1)
.getYDirAdj() + height; .getYDirAdj() + height;
} else if(textPositions.get(0).getRotation() == 90 && textPositions.get(0).getDir() == 0.0f){
posXInit = textPositions.get(textPositions.size() - 1) } else if (textPositions.get(0).getRotation() == 90 && textPositions.get(0).getDir() == 0.0f) {
posXInit = textPositions.get(textPositions.size() - 1)
.getXDirAdj() + textPositions.get(textPositions.size() - 1).getHeightDir(); .getXDirAdj() + textPositions.get(textPositions.size() - 1).getHeightDir();
posXEnd = textPositions.get(0).getXDirAdj(); posXEnd = textPositions.get(0).getXDirAdj();
posYInit = textPositions.get(0).getPageHeight() - textPositions.get(0).getYDirAdj() - 2; posYInit = textPositions.get(0).getPageHeight() - textPositions.get(0).getYDirAdj() - 2;
posYEnd = textPositions.get(0).getPageHeight() - textPositions.get(textPositions.size() - 1) posYEnd = textPositions.get(0).getPageHeight() - textPositions.get(textPositions.size() - 1)
.getYDirAdj() + 2; .getYDirAdj() + 2;
} else if(textPositions.get(0).getRotation() == 0 && textPositions.get(0).getDir() == 90f){
posYInit = getX1(); } else if (textPositions.get(0).getRotation() == 0 && textPositions.get(0).getDir() == 90f) {
posYEnd = getX2() + textPositions.get(0) posYInit = getX1();
.getWidthDirAdj() - textPositions.get(textPositions.size() - 1).getWidthDirAdj() - 3; posYEnd = getX2() + textPositions.get(0).getWidthDirAdj() - textPositions.get(textPositions.size() - 1)
posXInit = textPositions.get(0).getYDirAdj() + 2; .getWidthDirAdj() - 3;
posXEnd = textPositions.get(textPositions.size() - 1) posXInit = textPositions.get(0).getYDirAdj() + 2;
.getYDirAdj() - height; posXEnd = textPositions.get(textPositions.size() - 1).getYDirAdj() - height;
}
else { } else {
posXEnd = textPositions.get(textPositions.size() - 1) posXEnd = textPositions.get(textPositions.size() - 1)
.getXDirAdj() + textPositions.get(textPositions.size() - 1).getWidthDirAdj() + 1; .getXDirAdj() + textPositions.get(textPositions.size() - 1).getWidthDirAdj() + 1;
posYInit = textPositions.get(0).getPageHeight() - textPositions.get(0).getYDirAdj() - 2; posYInit = textPositions.get(0).getPageHeight() - textPositions.get(0).getYDirAdj() - 2;
@ -283,7 +299,9 @@ public class TextPositionSequence implements CharSequence {
.getYDirAdj() + 2; .getYDirAdj() + 2;
} }
return new Rectangle(new Point(posXInit, posYInit), posXEnd - posXInit, posYEnd - posYInit + height, page); var rectangle = new Rectangle(new Point(posXInit, posYInit), posXEnd - posXInit, posYEnd - posYInit + height, page);
log.debug("Rectangle: {}", rectangle);
return rectangle;
} }
} }

BIN
redaction-service-v1/redaction-service-server-v1/src/test/resources/files/new/VV-511309.pdf Normal file
View File

Binary file not shown.