ContentScanningSample.cs :  » PDF » PDF-Clown » it » stefanochizzolini » clown » samples » C# / CSharp Open Source

using it.stefanochizzolini.clown.documents;
using it.stefanochizzolini.clown.documents.contents;
using it.stefanochizzolini.clown.documents.contents.composition;
using it.stefanochizzolini.clown.documents.contents.objects;
using xObjectsit.stefanochizzolini.clown.documents.contents.xObjects;
using it.stefanochizzolini.clown.files;
using it.stefanochizzolini.clown.objects;

using System;
using System.Collections.Generic;
using System.Drawing;

namespace it.stefanochizzolini.clown.samples{
  /**
    <summary>This sample demonstrates how to retrieve the precise position (page and coordinates)
    of each image within a PDF document, using the page content scanning functionality.</summary>
    <remarks>This sample leverages the ContentScanner class, a powerful device for accessing
    each single content object within a page.</remarks>
  */
  public class ContentScanningSample
    : ISample
  {
    #region dynamic
    #region interface
    #region public
    #region ISample
    public void Run(
      SampleLoader loader
      )
    {
      // (boilerplate user choice -- ignore it)
      string filePath = loader.GetPdfFileChoice("Please select a PDF file");

      // 1. Open the PDF file!
      File file = new File(filePath);

      // 2. Parsing the document...
      // Get the PDF document!
      Document document = file.Document;
      Console.WriteLine("\nLooking for images...");
      // Parsing each page of the document...
      foreach(Page page in document.Pages)
      {Scan(page);}
      Console.WriteLine("\nImage search complete.");
    }
    #endregion
    #endregion

    #region private
    /**
      <summary>Scans a page's contents looking for images.</summary>
    */
    private void Scan(
      Page page
      )
    {
      /*
        NOTE: Page contents are represented by a sequence of content objects,
        possibly nested into multiple levels.
      */
      Scan(
        new ContentScanner( // Wraps the page contents into the scanner.
          page.Contents // Gets the page contents.
          ),
        page
        );
    }

    /**
      <summary>Scans a content level looking for images.</summary>
    */
    private void Scan(
      ContentScanner level,
      Page page
      )
    {
      if(level == null)
        return;

      ContentObject obj;
      while((obj = level.Current) != null)
      {
        /*
          NOTE: Images can be represented on a page either as inline objects
          or as external objects (XObject).
        */
        // Is it an operation that shows an external object?
        /*
          NOTE: External objects are represented through PaintXObject operations.
        */
        if(obj is PaintXObject)
        {
          // Get the reference key of the shown external object!
          PdfName xObjectKey = ((PaintXObject)obj).Name;
          // Get the external object associated to the reference key!
          xObjects::XObject xObject = page.Resources.XObjects[xObjectKey];
          // Is the external object an image?
          if(xObject is xObjects::ImageXObject)
          {
            Console.WriteLine(
              "Image '" + xObjectKey + "' (" + xObject.BaseObject + ") " // Image key and indirect reference.
                + "on page " + (page.Index + 1) + " (" + page.BaseObject + ")" // Page index and indirect reference.
                );

            // Get the coordinates of the image!
            double[] ctm = level.State.CTM; // Current transformation matrix.
            SizeF imageSize = xObject.Size; // Image native size.
            Console.WriteLine("  Coordinates:");
            Console.WriteLine("     x: " + Math.Round(ctm[4]));
            Console.WriteLine("     y: " + Math.Round(page.Size.Value.Height - ctm[5]));
            Console.WriteLine("     width: " + Math.Round(ctm[0]) + " (native: " + Math.Round(imageSize.Width) + ")");
            Console.WriteLine("     height: " + Math.Round(Math.Abs(ctm[3])) + " (native: " + Math.Round(imageSize.Height) + ")");
          }
        }
        else if(obj is InlineImage)
        {
          Console.WriteLine(
            "Inline Image (TODO: info extraction)"
            );
        }
        else if(!(obj is Operation)) // Composite object.
        {
          // Scan the inner level!
          Scan(
            level.ChildLevel,
            page
            );
        }

        level.MoveNext();
      }
    }
    #endregion
    #endregion
    #endregion
  }
}