2

Iam using the PDF BOX for getting color information of the text in PDF. I could able to get the output by using the following code. But my doubt is what StrokingColor represents, what Non stroking color represents. Based on this how will i decide which text is having which color. Anyone suggest me? My cuurent output is like this:DeviceRGB DeviceCMYK java.awt.Color[r=63,g=240,b=0] java.awt.Color[r=35,g=31,b=32] 34.934998 31.11 31.875

PDDocument doc = null;
        try {
            doc = PDDocument.load(strFilepath);
            PDFStreamEngine engine = new PDFStreamEngine(ResourceLoader.loadProperties("org/apache/pdfbox/resources/PageDrawer.properties"));
            PDPage page = (PDPage)doc.getDocumentCatalog().getAllPages().get(1);
            engine.processStream(page, page.findResources(), page.getContents().getStream());
            PDGraphicsState graphicState = engine.getGraphicsState();
            System.out.println(graphicState.getStrokingColor().getColorSpace().getName());
            System.out.println(graphicState.getNonStrokingColor().getColorSpace().getName());
            System.out.println(graphicState.getNonStrokingColor().getJavaColor()); 
            System.out.println(graphicState.getStrokingColor().getJavaColor());
            float colorSpaceValues[] = graphicState.getStrokingColor().getColorSpaceValue();
            for (float c : colorSpaceValues) {
                System.out.println(c * 255);
            }
        }
        finally {
            if (doc != null) {
                doc.close();
            }
        }
Lucene1
  • 105
  • 1
  • 12
  • Your code first processes the whole selected page and then looks at the final graphics state. Thus, the colors you get are the colors which would be used for text drawn next (which does not exist because you scanned the whole page content) but not necessarily the colors used for anything actually drawn on the page. You find a sample of a text stripper looking at text color at the end of [this answer](http://stackoverflow.com/questions/20878170/how-to-determine-artificial-bold-style-artificial-italic-style-and-artificial-o/20924898#20924898). – mkl Jan 29 '14 at 12:26
  • I want to determine the color for the particular text area or sentence. Can you help me in doing that? – Lucene1 Jan 29 '14 at 12:36
  • How are you given that area? Usual rectangle coordinates? Or that sentence, how is that sentence to be recognized? – mkl Jan 29 '14 at 14:57
  • iam new to pdfbox. Iam not sure about how to give data. Above is my requirement. Please help me in doing this. – Lucene1 Jan 30 '14 at 05:10
  • But shouldn't your requirement especially clarify inputs and expected outputs of your code? – mkl Jan 30 '14 at 05:42
  • My Requirement is like: I have to compare the font colors of one pdf page to another pdf page. If I can get the font colors of text then it will be easy for me to compare. – Lucene1 Jan 30 '14 at 06:33
  • Just used text colors in general or do the pages in question contain the same text and you are expected to check whether the corresponding letters on those pages habe the same color? – mkl Jan 30 '14 at 06:44
  • I need color of the text..example , if there is a text "Sample" in black color and some other text "sample1" in grey color....i need to know that sample--> black color, sample1-->grey color like this..i want the full text and its color...bit comprehensive – Lucene1 Jan 30 '14 at 06:51

1 Answers1

6

According to the clarifications in comments the OP wants to

compare the font colors of one pdf page to another pdf page [...] if there is a text "Sample" in black color and some other text "sample1" in grey color....i need to know that sample--> black color, sample1-->grey color like this..i want the full text and its color

PDFBox has a text extraction engine, the PDFTextStripper. There are some challenges in using it for the task at hand, though, among them:

  • Originally it is not designed for extracting color information alongside the text; the TextPosition objects it uses don't even have any attribute for color. Thus, we will have to extend it somewhat.

    • We will first register listeners for color operations to keep track of colors at all.

    • We will furthermore store the color information for a TextPosition object in another structure (I would prefer to extend text position accordingly but due to several inaccessible private members that would have meant quite some hassle).

    • This has already been shown in details in this answer; for the backgrounds, look there.

  • PDF allows many ways of drawing text. The letters may be filled with one color and its border may be stroked with another. Their border may even serve as clipping path for following drawing operations. We will only consider filling and stroking colors.

  • Text drawn may later on be covered by other drawings, either completely hiding it or changing its apparent color. We will ignore this for now.

For PDFBox 1.8.x

As indicated, we extend the PDFTextStripper like this:

import java.io.IOException;
import java.util.Arrays;
import java.util.HashMap;
import java.util.List;
import java.util.Map;

import org.apache.pdfbox.util.PDFTextStripper;
import org.apache.pdfbox.util.TextPosition;

public class ColorTextStripper extends PDFTextStripper
{
    public ColorTextStripper() throws IOException
    {
        super();
        setSuppressDuplicateOverlappingText(false);

        registerOperatorProcessor("CS", new org.apache.pdfbox.util.operator.SetStrokingColorSpace());
        registerOperatorProcessor("cs", new org.apache.pdfbox.util.operator.SetNonStrokingColorSpace());
        registerOperatorProcessor("SC", new org.apache.pdfbox.util.operator.SetStrokingColor());
        registerOperatorProcessor("sc", new org.apache.pdfbox.util.operator.SetNonStrokingColor());
        registerOperatorProcessor("SCN", new org.apache.pdfbox.util.operator.SetStrokingColor());
        registerOperatorProcessor("scn", new org.apache.pdfbox.util.operator.SetNonStrokingColor());
        registerOperatorProcessor("G", new org.apache.pdfbox.util.operator.SetStrokingGrayColor());
        registerOperatorProcessor("g", new org.apache.pdfbox.util.operator.SetNonStrokingGrayColor());
        registerOperatorProcessor("RG", new org.apache.pdfbox.util.operator.SetStrokingRGBColor());
        registerOperatorProcessor("rg", new org.apache.pdfbox.util.operator.SetNonStrokingRGBColor());
        registerOperatorProcessor("K", new org.apache.pdfbox.util.operator.SetStrokingCMYKColor());
        registerOperatorProcessor("k", new org.apache.pdfbox.util.operator.SetNonStrokingCMYKColor());
    }

    @Override
    protected void processTextPosition(TextPosition text)
    {
        renderingMode.put(text, getGraphicsState().getTextState().getRenderingMode());
        strokingColor.put(text, getGraphicsState().getStrokingColor().getColorSpaceValue());
        nonStrokingColor.put(text, getGraphicsState().getNonStrokingColor().getColorSpaceValue());

        super.processTextPosition(text);
    }

    Map<TextPosition, Integer> renderingMode = new HashMap<TextPosition, Integer>();
    Map<TextPosition, float[]> strokingColor = new HashMap<TextPosition, float[]>();
    Map<TextPosition, float[]> nonStrokingColor = new HashMap<TextPosition, float[]>();

    final static List<Integer> FILLING_MODES = Arrays.asList(0, 2, 4, 6);
    final static List<Integer> STROKING_MODES = Arrays.asList(1, 2, 5, 6);
    final static List<Integer> CLIPPING_MODES = Arrays.asList(4, 5, 6, 7);

    @Override
    protected void writeString(String text, List<TextPosition> textPositions) throws IOException
    {
        for (TextPosition textPosition: textPositions)
        {
            Integer charRenderingMode = renderingMode.get(textPosition);
            float[] charStrokingColor = strokingColor.get(textPosition);
            float[] charNonStrokingColor = nonStrokingColor.get(textPosition);

            StringBuilder textBuilder = new StringBuilder();
            textBuilder.append(textPosition.getCharacter())
                       .append("{");

            if (FILLING_MODES.contains(charRenderingMode))
            {
                textBuilder.append("FILL:")
                           .append(toString(charNonStrokingColor))
                           .append(';');
            }

            if (STROKING_MODES.contains(charRenderingMode))
            {
                textBuilder.append("STROKE:")
                           .append(toString(charStrokingColor))
                           .append(';');
            }

            if (CLIPPING_MODES.contains(charRenderingMode))
            {
                textBuilder.append("CLIP;");
            }

            textBuilder.append("}");
            writeString(textBuilder.toString());
        }
    }

    String toString(float[] values)
    {
        if (values == null)
            return "null";
        StringBuilder builder = new StringBuilder();
        switch(values.length)
        {
        case 1:
            builder.append("GRAY"); break;
        case 3:
            builder.append("RGB"); break;
        case 4:
            builder.append("CMYK"); break;
        default:
            builder.append("UNKNOWN");
        }
        for (float f: values)
        {
            builder.append(' ')
                   .append(f);
        }

        return builder.toString();
    }
}

You can call it like this:

PDFTextStripper stripper = new ColorTextStripper();

PDDocument document = PDDocument.load(SOURCE_FILE);

String text = stripper.getText(document);

The resulting text contains something like this:

P{FILL:RGB 0.803 0.076 0.086;}e{FILL:RGB 0.803 0.076 0.086;}l{FILL:RGB 0.803 0.076 0.086;}l{FILL:RGB 0.803 0.076 0.086;}e{FILL:RGB 0.803 0.076 0.086;}

and 

G{FILL:RGB 0.102 0.101 0.095;}r{FILL:RGB 0.102 0.101 0.095;}a{FILL:RGB 0.102 0.101 0.095;}z{FILL:RGB 0.102 0.101 0.095;}i{FILL:RGB 0.102 0.101 0.095;}e{FILL:RGB 0.102 0.101 0.095;}

for the Pelle and Grazie from this

Pelle and Grazie

or 

K{FILL:RGB 0.0 0.322 0.573;}E{FILL:RGB 0.0 0.322 0.573;}Y{FILL:RGB 0.0 0.322 0.573;}

and 

C{FILL:GRAY 0.0;}o{FILL:GRAY 0.0;}m{FILL:GRAY 0.0;}b{FILL:GRAY 0.0;}i{FILL:GRAY 0.0;}n{FILL:GRAY 0.0;}e{FILL:GRAY 0.0;}d{FILL:GRAY 0.0;}

for KEY and Combined from this:

KEY and Combined

Instead of serializing all the information into a String result, you can of course also create some class containing both the color and the character information in a structured way. Just like now the String result is created in writeString, you can change this method to add instances of such a class to some list in it.

Requirements

At least PDFBox version 1.8.4 is required to make this work. I tested it using 2.0.0-SNAPSHOT but 1.8.4 should suffice. 1.8.3, on the other hand, has a bug which sometimes forwards the wrong TextPosition objects to writeString, cf. PDFBOX-1804, and earlier versions don't provide a TextPosition collection to writeString at all.

For PDFBox 2.x

There were multiple refactorings and other changes in PDFBox 2.x which also concern the code above.

Ported to PDFBox 2.x it may look like this:

public class ColorTextStripper extends PDFTextStripper {
    public ColorTextStripper() throws IOException {
        super();
        setSuppressDuplicateOverlappingText(false);

        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetStrokingColorSpace());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetNonStrokingColorSpace());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetStrokingColor());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetNonStrokingColor());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetStrokingColorN());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetNonStrokingColorN());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetStrokingDeviceGrayColor());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetNonStrokingDeviceGrayColor());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetStrokingDeviceRGBColor());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetNonStrokingDeviceRGBColor());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetStrokingDeviceCMYKColor());
        addOperator(new org.apache.pdfbox.contentstream.operator.color.SetNonStrokingDeviceCMYKColor());
    }

    @Override
    protected void processTextPosition(TextPosition text) {
        renderingMode.put(text, getGraphicsState().getTextState().getRenderingMode());
        strokingColor.put(text, getGraphicsState().getStrokingColor().getComponents());
        nonStrokingColor.put(text, getGraphicsState().getNonStrokingColor().getComponents());

        super.processTextPosition(text);
    }

    Map<TextPosition, RenderingMode> renderingMode = new HashMap<TextPosition, RenderingMode>();
    Map<TextPosition, float[]> strokingColor = new HashMap<TextPosition, float[]>();
    Map<TextPosition, float[]> nonStrokingColor = new HashMap<TextPosition, float[]>();

    final static List<RenderingMode> FILLING_MODES = Arrays.asList(RenderingMode.FILL, RenderingMode.FILL_STROKE, RenderingMode.FILL_CLIP, RenderingMode.FILL_STROKE_CLIP);
    final static List<RenderingMode> STROKING_MODES = Arrays.asList(RenderingMode.STROKE, RenderingMode.FILL_STROKE, RenderingMode.STROKE_CLIP, RenderingMode.FILL_STROKE_CLIP);
    final static List<RenderingMode> CLIPPING_MODES = Arrays.asList(RenderingMode.FILL_CLIP, RenderingMode.STROKE_CLIP, RenderingMode.FILL_STROKE_CLIP, RenderingMode.NEITHER_CLIP);

    @Override
    protected void writeString(String text, List<TextPosition> textPositions) throws IOException {
        for (TextPosition textPosition: textPositions) {
            RenderingMode charRenderingMode = renderingMode.get(textPosition);
            float[] charStrokingColor = strokingColor.get(textPosition);
            float[] charNonStrokingColor = nonStrokingColor.get(textPosition);

            StringBuilder textBuilder = new StringBuilder();
            textBuilder.append(textPosition.getUnicode()).append("{");

            if (FILLING_MODES.contains(charRenderingMode)) {
                textBuilder.append("FILL:").append(toString(charNonStrokingColor)).append(';');
            }

            if (STROKING_MODES.contains(charRenderingMode)) {
                textBuilder.append("STROKE:").append(toString(charStrokingColor)).append(';');
            }

            if (CLIPPING_MODES.contains(charRenderingMode)) {
                textBuilder.append("CLIP;");
            }

            textBuilder.append("}");
            writeString(textBuilder.toString());
        }
    }

    String toString(float[] values)
    {
        if (values == null)
            return "null";
        StringBuilder builder = new StringBuilder();
        switch(values.length) {
        case 1:
            builder.append("GRAY"); break;
        case 3:
            builder.append("RGB"); break;
        case 4:
            builder.append("CMYK"); break;
        default:
            builder.append("UNKNOWN");
        }
        for (float f: values) {
            builder.append(' ')
                   .append(f);
        }

        return builder.toString();
    }
}

(ColorTextStripper)

mkl
  • 90,588
  • 15
  • 125
  • 265
  • @mkl is there anyway that the numbers can be transposed into the actual colours they represent? E.g. P{FILL:RGB 0.803 0.076 0.086;} - We know from your above answer that is Red or thereabouts. If i didn't have the example, how would i know it was Red? I've tried putting it into a hex to RGB converter but no luck (as i suspect it not Hex). Likewise, {FILL:CMYK 1.0 0.87 0.34 0.3;} - How would i work out what this colour actual is? – Rusty Shackleford Mar 09 '21 at 00:14
  • 1
    Pdf uses floating point ranges from 0 to 1 for the color components. Hex representations use int ranges from 0 to 255 for them. Thus, you have to multiply each value by 255 and then round when you want to determine a hex color representation. – mkl Mar 09 '21 at 07:46
  • 1
    *"E.g. P{FILL:RGB 0.803 0.076 0.086;} - We know from your above answer that is Red or thereabouts. If i didn't have the example, how would i know it was Red?"* - the .8 Red is much larger than the .07 green and the .08 blue. Thus, it's some flavor of red. – mkl Mar 09 '21 at 07:57