Extract text and images from any PDF

Santiago,

Yes it is possible and fairly easy to extract page text. Use the GetPageText function and it will return a StringList filled with all of the text found on a given page. Depending upon extract options you can usually figure out the location of the text on the page. The only problem I’ve encountered is that not all PDF are created to same way. I have see PDFs where every single character is a text element. When that happens it can be very difficult to turn those characters back into words.

Image extraction is also simple:

procedure TForm1.btnExtractImagesClick(Sender: TObject);

var   ImageID, i : Integer;

      imageCount : integer;

      SaveName, FirstName : string;

      sl : TStringList;

begin

      qPDF := TiSEDQuickPDF.Create;

      qPDF.UnlockKey(MY_QDPF_KEY);

      if OpenDialog1.Execute then

            begin

            qPDF.LoadFromFile( OpenDialog1.FileName );

            FirstName := copy(OpenDialog1.FileName,1,pos('.pdf',lowercase(OpenDialog1.FileName)));

            imageCount := qPDF.FindImages; //This is Required before ImageID will work

            ShowMessage('ImageCount ' + IntToStr(imageCount) );

            for i := 1 to imageCount do

                  begin

                  ImageID := qPDF.ImageID( i );

                  qPDF.SelectImage( ImageID );

                  case qPDF.ImageType of

                  1 : SaveName := Format('%s%d.jpg',[FirstName,i]);    // The selected image is a JPEG image.

                  2 : SaveName := Format('%s%d.bmp',[FirstName,i]);    // The selected image is a BMP image.

                  3 : SaveName := Format('%s%d.tif',[FirstName,i]);    // The selected image is a TIFF image.

                        end;

                  qPDF.SaveImageToFile( SaveName );

                  end;

            end;

      qPDF.Free;

end;

regards,

A lot of trial and error.

It helps to be a Delphi programmer and to have the source code and, of course, this wonderful forum.

sb

The knowledge base helped me a lot and I was able to test sb's code successfully.

This component rocks!

My last question would be which version do you suggest using? I saw there were some bug fixes in the last beta.

Thanks,
Santiago