Marmelator/layout-parser
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 289 Wiki Activity

RED-8800: revert cropbox adjustment

Browse Source
This commit is contained in:
Kilian Schuettler 2024-07-15 16:04:15 +02:00
parent 785ff0dc4b
commit da8854ba37
1 changed files with 19 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/parsing/LegacyPDFStreamEngine.java
View File

@ -82,6 +82,7 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
private int pageRotation; private int pageRotation;
private PDRectangle pageSize; private PDRectangle pageSize;
private Matrix translateMatrix;
private final GlyphList glyphList; private final GlyphList glyphList;
private final Map<COSDictionary, Float> fontHeightMap = new WeakHashMap<COSDictionary, Float>(); private final Map<COSDictionary, Float> fontHeightMap = new WeakHashMap<COSDictionary, Float>();
@ -133,6 +134,12 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
this.pageRotation = page.getRotation(); this.pageRotation = page.getRotation();
this.pageSize = page.getCropBox(); this.pageSize = page.getCropBox();
if (pageSize.getLowerLeftX() == 0 && pageSize.getLowerLeftY() == 0) {
translateMatrix = null;
} else {
// translation matrix for cropbox
translateMatrix = Matrix.getTranslateInstance(-pageSize.getLowerLeftX(), -pageSize.getLowerLeftY());
}
super.processPage(page); super.processPage(page);
} }
@ -257,13 +264,22 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
return; return;
} }
} }
// adjust for cropbox if needed
Matrix translatedTextRenderingMatrix;
if (translateMatrix == null) {
translatedTextRenderingMatrix = textRenderingMatrix;
} else {
translatedTextRenderingMatrix = Matrix.concatenate(translateMatrix, textRenderingMatrix);
nextX -= pageSize.getLowerLeftX();
nextY -= pageSize.getLowerLeftY();
}
// This is a hack for unicode letter with 2 chars e.g. RA see unicodeProblem.pdf // This is a hack for unicode letter with 2 chars e.g. RA see unicodeProblem.pdf
if (unicodeMapping.length() == 2) { if (unicodeMapping.length() == 2) {
processTextPosition(new TextPosition(pageRotation, processTextPosition(new TextPosition(pageRotation,
pageSize.getWidth(), pageSize.getWidth(),
pageSize.getHeight(), pageSize.getHeight(),
textRenderingMatrix, translatedTextRenderingMatrix,
nextX, nextX,
nextY, nextY,
Math.abs(dyDisplay), Math.abs(dyDisplay),
@ -277,7 +293,7 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
processTextPosition(new TextPosition(pageRotation, processTextPosition(new TextPosition(pageRotation,
pageSize.getWidth(), pageSize.getWidth(),
pageSize.getHeight(), pageSize.getHeight(),
textRenderingMatrix, translatedTextRenderingMatrix,
nextX, nextX,
nextY, nextY,
Math.abs(dyDisplay), Math.abs(dyDisplay),
@ -293,7 +309,7 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
processTextPosition(new TextPosition(pageRotation, processTextPosition(new TextPosition(pageRotation,
pageSize.getWidth(), pageSize.getWidth(),
pageSize.getHeight(), pageSize.getHeight(),
textRenderingMatrix, translatedTextRenderingMatrix,
nextX, nextX,
nextY, nextY,
Math.abs(dyDisplay), Math.abs(dyDisplay),