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

tests

Browse Source
This commit is contained in:
maverickstuder 2024-02-14 14:54:10 +01:00
parent ec035aca2f
commit 7d6caabbfb
4 changed files with 160 additions and 93 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/PageContents.java
View File

@ -19,4 +19,5 @@ public class PageContents {
Rectangle2D cropBox;
Rectangle2D mediaBox;
List<Ruling> rulings;
List<Rectangle2D> positions;
}

4
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/PageContentExtractor.java
View File

@ -39,6 +39,8 @@ public class PageContentExtractor {
stripper.setEndPage(pageNumber);
stripper.setPdpage(pdPage);
stripper.getText(pdDocument);
var positions = stripper.getExactPositions();
Map<Float, List<TextPositionSequence>> sortedTextPositionSequencesPerDir = stripper.getTextPositionSequences()
.stream()
@ -49,7 +51,7 @@ public class PageContentExtractor {
textPositionSequencesPerPage.add(new PageContents(sortedTextPositionSequences,
RectangleTransformations.toRectangle2D(pdPage.getCropBox()),
RectangleTransformations.toRectangle2D(pdPage.getMediaBox()),
stripper.getRulings()));
stripper.getRulings(), positions));
}
}

188
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/parsing/LegacyPDFStreamEngine.java
View File

@ -16,13 +16,19 @@
*/
package com.knecon.fforesight.service.layoutparser.processor.services.parsing;
import java.awt.geom.AffineTransform;
import java.awt.geom.Area;
import java.awt.geom.Rectangle2D;
import java.io.IOException;
import java.io.InputStream;
import java.util.LinkedList;
import java.util.List;
import java.util.Map;
import java.util.WeakHashMap;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.fontbox.FontBoxFont;
import org.apache.fontbox.ttf.GlyphData;
import org.apache.fontbox.ttf.TrueTypeFont;
import org.apache.fontbox.util.BoundingBox;
@ -53,12 +59,14 @@ import org.apache.pdfbox.cos.COSDictionary;
import org.apache.pdfbox.pdmodel.PDPage;
import org.apache.pdfbox.pdmodel.common.PDRectangle;
import org.apache.pdfbox.pdmodel.font.PDCIDFont;
import org.apache.pdfbox.pdmodel.font.PDCIDFontType0;
import org.apache.pdfbox.pdmodel.font.PDCIDFontType2;
import org.apache.pdfbox.pdmodel.font.PDFont;
import org.apache.pdfbox.pdmodel.font.PDFontDescriptor;
import org.apache.pdfbox.pdmodel.font.PDSimpleFont;
import org.apache.pdfbox.pdmodel.font.PDTrueTypeFont;
import org.apache.pdfbox.pdmodel.font.PDType0Font;
import org.apache.pdfbox.pdmodel.font.PDType1Font;
import org.apache.pdfbox.pdmodel.font.PDType3Font;
import org.apache.pdfbox.pdmodel.font.encoding.GlyphList;
import org.apache.pdfbox.pdmodel.graphics.state.PDGraphicsState;
@ -66,6 +74,8 @@ import org.apache.pdfbox.text.TextPosition;
import org.apache.pdfbox.util.Matrix;
import org.apache.pdfbox.util.Vector;
import lombok.Getter;
/**
* LEGACY text calculations which are known to be incorrect but are depended on by PDFTextStripper.
* <p>
@ -87,6 +97,9 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
private final GlyphList glyphList;
private final Map<COSDictionary, Float> fontHeightMap = new WeakHashMap<COSDictionary, Float>();
@Getter
private List<Rectangle2D> exactPositions = new LinkedList<>();
/**
* Constructor.
@ -164,6 +177,51 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
float horizontalScaling = state.getTextState().getHorizontalScaling() / 100f;
Matrix textMatrix = getTextMatrix();
TrueTypeFont ttf = null;
if (font instanceof PDTrueTypeFont) {
ttf = ((PDTrueTypeFont) font).getTrueTypeFont();
} else if (font instanceof PDType0Font) {
PDCIDFont cidFont = ((PDType0Font) font).getDescendantFont();
if (cidFont instanceof PDCIDFontType2) {
ttf = ((PDCIDFontType2) cidFont).getTrueTypeFont();
}
} else if (font instanceof PDType1Font) {
FontBoxFont fontBoxFont = ((PDType1Font) font).getFontBoxFont();
if (fontBoxFont instanceof TrueTypeFont) {
ttf = (TrueTypeFont) fontBoxFont;
} else {
System.out.println("What do?");
}
}
if (ttf != null) {
Integer glyphId = null;
if (font instanceof PDTrueTypeFont) {
glyphId = ((PDTrueTypeFont) font).codeToGID(code);
} else if (font instanceof PDType0Font) {
glyphId = ((PDType0Font) font).codeToGID(code);
} else if (font instanceof PDType1Font) {
FontBoxFont fontBoxFont = ((PDType1Font) font).getFontBoxFont();
if (fontBoxFont instanceof TrueTypeFont) {
glyphId = ((TrueTypeFont) fontBoxFont).getUnicodeCmapLookup().getGlyphId(code);
}
}
if (glyphId != null) {
GlyphData glyph = ttf.getGlyph().getGlyph(glyphId);
if (glyph != null) {
BoundingBox boundingBox = glyph.getBoundingBox();
Rectangle2D rect = new Rectangle2D.Double(boundingBox.getLowerLeftX(), boundingBox.getLowerLeftY(), boundingBox.getWidth(), boundingBox.getHeight());
Area area = new Area(rect);
AffineTransform affineTransform = textRenderingMatrix.createAffineTransform();
float factor = 1f / ttf.getUnitsPerEm();
affineTransform.scale(factor, factor);
exactPositions.add(area.createTransformedArea(affineTransform).getBounds2D());
}
}
}
float displacementX = displacement.getX();
// the sorting algorithm is based on the width of the character. As the displacement
// for vertical characters doesn't provide any suitable value for it, we have to
@ -171,45 +229,11 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
if (font.isVertical()) {
displacementX = font.getWidth(code) / 1000;
// there may be an additional scaling factor for true type fonts
TrueTypeFont ttf = null;
if (font instanceof PDTrueTypeFont) {
ttf = ((PDTrueTypeFont) font).getTrueTypeFont();
} else if (font instanceof PDType0Font) {
PDCIDFont cidFont = ((PDType0Font) font).getDescendantFont();
if (cidFont instanceof PDCIDFontType2) {
ttf = ((PDCIDFontType2) cidFont).getTrueTypeFont();
}
}
if (ttf != null && ttf.getUnitsPerEm() != 1000) {
displacementX *= 1000f / ttf.getUnitsPerEm();
}
}
GlyphData glyph = null;
TrueTypeFont ttf = null;
Float actualGlyphMinX = null;
Float actualGlyphMaxX = null;
Float actualGlyphWidth = null;
if (font instanceof PDTrueTypeFont) {
ttf = ((PDTrueTypeFont) font).getTrueTypeFont();
} else if (font instanceof PDType0Font) {
PDType0Font type0Font = (PDType0Font) font;
PDCIDFont cidFont = type0Font.getDescendantFont();
if (cidFont instanceof PDCIDFontType2) {
ttf = ((PDCIDFontType2) cidFont).getTrueTypeFont();
int glyphId = type0Font.codeToGID(code);
glyph = ttf.getGlyph().getGlyph(glyphId);
if (glyph != null && glyph.getBoundingBox() != null) {
var lowerX = glyph.getBoundingBox().getLowerLeftX() * (fontSize / ttf.getUnitsPerEm());
var upperX = glyph.getBoundingBox().getUpperRightX() * (fontSize / ttf.getUnitsPerEm());
actualGlyphMinX = Math.min(lowerX, upperX);
actualGlyphMaxX = Math.max(lowerX, upperX);
actualGlyphWidth = actualGlyphMaxX - actualGlyphMinX;
}
}
}
//
// legacy calculations which were previously in PDFStreamEngine
//
@ -227,11 +251,11 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
// (modified) text rendering matrix
Matrix nextTextRenderingMatrix = td.multiply(textMatrix).multiply(ctm); // text space -> device space
float nextX = nextTextRenderingMatrix.getTranslateX();
float nextY = nextTextRenderingMatrix.getTranslateY();
float endX = nextTextRenderingMatrix.getTranslateX();
float endY = nextTextRenderingMatrix.getTranslateY();
// (modified) width and height calculations
float dxDisplay = nextX - textRenderingMatrix.getTranslateX();
float dxDisplay = endX - textRenderingMatrix.getTranslateX();
Float fontHeight = fontHeightMap.get(font.getCOSObject());
if (fontHeight == null) {
fontHeight = computeFontHeight(font);
@ -249,14 +273,6 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
// Text or Disp to represent if the values are in text or disp units (no glyph units are
// saved).
if(actualGlyphMinX != null) {
var oldDxDisplay = dxDisplay;
dxDisplay = actualGlyphWidth;
var diff = Math.abs(oldDxDisplay - dxDisplay);
//textRenderingMatrix.setValue(2,0, textRenderingMatrix.getTranslateX() + diff/2);
nextX -= diff;
}
float glyphSpaceToTextSpaceFactor = 1 / 1000f;
if (font instanceof PDType3Font) {
glyphSpaceToTextSpaceFactor = font.getFontMatrix().getScaleX();
@ -305,56 +321,56 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
translatedTextRenderingMatrix = textRenderingMatrix;
} else {
translatedTextRenderingMatrix = Matrix.concatenate(translateMatrix, textRenderingMatrix);
nextX -= pageSize.getLowerLeftX();
nextY -= pageSize.getLowerLeftY();
endX -= pageSize.getLowerLeftX();
endY -= pageSize.getLowerLeftY();
}
// This is a hack for unicode letter with 2 chars e.g. RA see unicodeProblem.pdf
if (unicodeMapping.length() == 2) {
processTextPosition(new TextPosition(pageRotation,
pageSize.getWidth(),
pageSize.getHeight(),
translatedTextRenderingMatrix,
nextX,
nextY,
Math.abs(dyDisplay),
dxDisplay,
Math.abs(spaceWidthDisplay),
Character.toString(unicodeMapping.charAt(0)),
new int[]{code},
font,
fontSize,
(int) (fontSize * textMatrix.getScalingFactorX())));
pageSize.getWidth(),
pageSize.getHeight(),
translatedTextRenderingMatrix,
endX,
endY,
Math.abs(dyDisplay),
dxDisplay,
Math.abs(spaceWidthDisplay),
Character.toString(unicodeMapping.charAt(0)),
new int[]{code},
font,
fontSize,
(int) (fontSize * textMatrix.getScalingFactorX())));
processTextPosition(new TextPosition(pageRotation,
pageSize.getWidth(),
pageSize.getHeight(),
translatedTextRenderingMatrix,
nextX,
nextY,
Math.abs(dyDisplay),
dxDisplay,
Math.abs(spaceWidthDisplay),
Character.toString(unicodeMapping.charAt(1)),
new int[]{code},
font,
fontSize,
(int) (fontSize * textMatrix.getScalingFactorX())));
pageSize.getWidth(),
pageSize.getHeight(),
translatedTextRenderingMatrix,
endX,
endY,
Math.abs(dyDisplay),
dxDisplay,
Math.abs(spaceWidthDisplay),
Character.toString(unicodeMapping.charAt(1)),
new int[]{code},
font,
fontSize,
(int) (fontSize * textMatrix.getScalingFactorX())));
} else {
processTextPosition(new TextPosition(pageRotation,
pageSize.getWidth(),
pageSize.getHeight(),
translatedTextRenderingMatrix,
nextX,
nextY,
Math.abs(dyDisplay),
dxDisplay,
Math.abs(spaceWidthDisplay),
unicodeMapping,
new int[]{code},
font,
fontSize,
(int) (fontSize * textMatrix.getScalingFactorX())));
pageSize.getWidth(),
pageSize.getHeight(),
translatedTextRenderingMatrix,
endX,
endY,
Math.abs(dyDisplay),
dxDisplay,
Math.abs(spaceWidthDisplay),
unicodeMapping,
new int[]{code},
font,
fontSize,
(int) (fontSize * textMatrix.getScalingFactorX())));
}
}

60
layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/services/PageContentExtractorTest.java
View File

@ -5,6 +5,12 @@ import java.awt.geom.Rectangle2D;
import java.nio.file.Path;
import java.util.List;
import org.apache.pdfbox.pdmodel.PDDocument;
import org.apache.pdfbox.pdmodel.PDPage;
import org.apache.pdfbox.pdmodel.PDPageContentStream;
import org.apache.pdfbox.pdmodel.font.PDFont;
import org.apache.pdfbox.pdmodel.font.PDType1Font;
import org.apache.pdfbox.pdmodel.font.Standard14Fonts;
import org.junit.jupiter.api.Disabled;
import org.junit.jupiter.api.Test;
@ -24,18 +30,22 @@ class PageContentExtractorTest {
@SneakyThrows
public void testTextPositionSequenceExtraction() {
String fileName = "files/CLEAN-II-4.2.2.01_家畜残留分析法.pdf";
//String fileName = "files/CLEAN-II-4.2.2.01_家畜残留分析法.pdf";
//String fileName = "files/BASF/2013-1110704.pdf";
//String fileName = "files/ImportRedactionTestFile_highlighted.pdf";
String fileName = "files/HelloWorldHelvetica.pdf";
var tmpFileName = Path.of("/tmp/").resolve(Path.of(fileName).getFileName() + "_TEXT_POSITION_SEQUENCES.pdf").toString();
List<PageContents> textPositionPerPage = PageContentExtractor.getSortedPageContents(fileName);
PdfDraw.drawRectanglesPerPageNumberedByLine(fileName,
PdfDraw.drawRectanglesPerPage(fileName,
textPositionPerPage.stream()
.map(t -> t.getSortedTextPositionSequences()
.stream().flatMap(sequence -> sequence.getTextPositions().stream().map(textPosition -> mapRedTextPositionToInitialUserSpace(textPosition, sequence)))
//.map(t -> t.getSortedTextPositionSequences()
.map(t -> t.getPositions())
// .stream().flatMap(sequence -> sequence.getTextPositions().stream().map(textPosition -> mapRedTextPositionToInitialUserSpace(textPosition, sequence)))
//.map(textPositionSequence -> (Rectangle2D) new Rectangle2D.Double(textPositionSequence.getMaxXDirAdj(), textPositionSequence.getMaxYDirAdj(), textPositionSequence.getWidth(), textPositionSequence.getHeight()))
.map(List::of)
.toList())
//.map(List::of)
//.toList())
.toList(), tmpFileName);
}
@ -66,4 +76,42 @@ class PageContentExtractorTest {
return transform.createTransformedShape(rectangle2D).getBounds2D();
}
@Test
@SneakyThrows
public void generatePDF() {
// Create a new PDF document
PDDocument document = new PDDocument();
// Create a blank page
PDPage page = new PDPage();
document.addPage(page);
// Load the Helvetica font
PDFont font = new PDType1Font(Standard14Fonts.FontName.HELVETICA);
// Start a content stream to write text
PDPageContentStream contentStream = new PDPageContentStream(document, page);
contentStream.beginText();
// Set font and font size
contentStream.setFont(font, 12);
// Set text position
contentStream.newLineAtOffset(50, 700);
// Write the text
contentStream.showText("Hello World in Helvetica!");
// Finish writing text
contentStream.endText();
contentStream.close();
// Save the PDF
document.save("/tmp/MyPDF.pdf");
document.close();
System.out.println("PDF created successfully!");
}
}