Using PDFbox to determine the coordinates of words in a document

For those who still need assistance, this is what I used in my code and should be useful to start with. It uses PDFBox 2.0.16

public class PDFTextLocator extends PDFTextStripper {
    
    private static String key_string;
    private static float x;
    private static float y;
    
    public PDFTextLocator() throws IOException {
        x = -1;
        y = -1;
    }
    
    /**
     * Takes in a PDF Document, phrase to find, and page to search and returns the x,y in float array
     * @param document
     * @param phrase
     * @param page
     * @return
     * @throws IOException
     */
    public static float[] getCoordiantes(PDDocument document, String phrase, int page) throws IOException {
        key_string = phrase;
        PDFTextStripper stripper = new PDFTextLocator();
        stripper.setSortByPosition(true);
        stripper.setStartPage(page);
        stripper.setEndPage(page);
        stripper.writeText(document, new OutputStreamWriter(new ByteArrayOutputStream()));
        y = document.getPage(page).getMediaBox().getHeight()-y;
        
        return new float[]{x,y};
    }
    
    /**
     * Override the default functionality of PDFTextStripper.writeString()
     */
    @Override
    protected void writeString(String string, List<TextPosition> textPositions) throws IOException {
        if(string.contains(key_string)) {
            TextPosition text = textPositions.get(0);
            if(x == -1) {
                x = text.getXDirAdj();
                y = text.getYDirAdj();
            }        
        }
    }
}

Below is the Maven dependency details,

<dependency>
    <groupId>org.apache.pdfbox</groupId>
    <artifactId>pdfbox</artifactId>
    <version>2.0.16</version>
</dependency>

I got this working using the IKVM conversion PDFBox.NET 1.8.9. in C# and .NET.

I finally figured out the character (glyph) coordinates are private to the .NET assembly, but can be accessed using System.Reflection.

I posted a full example of getting the coordinates of WORDS and drawing them back on images of PDF's using SVG and HTML here: https://github.com/tsamop/PDF_Interpreter

For the example below you need PDFbox.NET: http://www.squarepdf.net/pdfbox-in-net, and include references to it in your project.

It took me quite a while to figure it out, so I really hope it saves someone else time!!

If you just need to know where to look for the characters & coordinates, a very abridged version would be:

    using System;
using System.Reflection;

using org.apache.pdfbox.pdmodel;
using org.apache.pdfbox.util;

  // to test run pdfTest.RunTest(@"C:\temp\test_2.pdf");

class pdfTest
{
    //simple example for getting character (gliph) coordinates out of a pdf doc.
    // a more complete example is here:   https://github.com/tsamop/PDF_Interpreter
    public static void RunTest(string sFilename)
    {
    //probably a better way to get page count, but I cut this out of a bigger project.
    PDDocument oDoc = PDDocument.load(sFilename);
        object[] oPages = oDoc.getDocumentCatalog().getAllPages().toArray();

        int iPageNo = 0; //1's based!!
        foreach (object oPage in oPages)
        {
            iPageNo++;

            //feed the stripper a page.
            PDFTextStripper tStripper = new PDFTextStripper();
            tStripper.setStartPage(iPageNo);
            tStripper.setEndPage(iPageNo);
            tStripper.getText(oDoc);

            //This gets the "charactersByArticle" private object in PDF Box.
            FieldInfo charactersByArticleInfo = typeof(PDFTextStripper).GetField("charactersByArticle", BIndingFlags.NonPublic | BindingFlags.Instance);
            object charactersByArticle = charactersByArticleInfo.GetValue(tStripper);

            object[] aoArticles = (object[])charactersByArticle.GetField("elementData");

            foreach (object oArticle in aoArticles)
            {
                if (oArticle != null)
                {

                    //THE CHARACTERS within the article
                    object[] aoCharacters = (object[])oArticle.GetField("elementData");

                    foreach (object oChar in aoCharacters)
                    {

                        /*properties I caulght using reflection:
                         * endX, endY, font, fontSize, fontSizePt, maxTextHeight, pageHeight, pageWidth, rot, str textPos, unicodCP, widthOfSpace, widths, wordSpacing, x, y
                         * 
                         */
                        if (oChar != null)
                        {
                            //this is a really quick test.  
                            // for a more complete solution that pulls the characters into words and displays the word positions on the page, try this:  https://github.com/tsamop/PDF_Interpreter
                            //the Y's appear to be the bottom of the char?
                            double mfMaxTextHeight = Convert.ToDouble(oChar.GetField("maxTextHeight")); //I think this is the height of the character/word

                            char mcThisChar = oChar.GetField("str").ToString().ToCharArray()[0];
                            double mfX = Convert.ToDouble(oChar.GetField("x"));
                            double mfY = Convert.ToDouble(oChar.GetField("y")) - mfMaxTextHeight;

                            //CALCULATE THE OTHER SIDE OF THE GLIPH
                            double mfWidth0 = ((Single[])oChar.GetField("widths"))[0];
                            double mfXend = mfX + mfWidth0; // Convert.ToDouble(oChar.GetField("endX"));

                            //CALCULATE THE BOTTOM OF THE GLIPH.
                            double mfYend = mfY + mfMaxTextHeight; //  Convert.ToDouble(oChar.GetField("endY"));

                            double mfPageHeight = Convert.ToDouble(oChar.GetField("pageHeight"));
                            double mfPageWidth = Convert.ToDouble(oChar.GetField("pageWidth"));

                            System.Diagnostics.Debug.Print(@"add some stuff to test {0}, {1}, {2}", mcThisChar, mfX, mfY);

                        }
                    }

                }
            }

        }
    }
}



using System.Reflection;

/// <summary>
/// To deal with the Java interface hiding necessary properties! ~mwr
/// </summary>
public static class GetField_Extension
{
    public static object GetField(this object randomPDFboxObject, string sFieldName)
    {
            FieldInfo itemInfo = randomPDFboxObject.GetType().GetField(sFieldName, BindingFlags.NonPublic | BindingFlags.Instance);
            return itemInfo.GetValue(randomPDFboxObject);
    }

}