Crop and convert to JPG embeeded images

src/main/java/w2phtml/base/BinaryGraphicsDocument.java

@@ -25,12 +25,25 @@
 
 package w2phtml.base;
 
+import static w2phtml.office.XMLString.FO_CLIP;
+
+import java.awt.image.BufferedImage;
+import java.awt.image.RasterFormatException;
+import java.io.ByteArrayInputStream;
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
 import java.io.OutputStream;
 
+import javax.imageio.ImageIO;
+
+import org.apache.commons.imaging.ImageInfo;
+import org.apache.commons.imaging.ImageReadException;
+import org.apache.commons.imaging.Imaging;
+
 import w2phtml.api.OutputFile;
-
+import w2phtml.office.MIMETypes;
-import java.io.IOException;
+import w2phtml.office.StyleWithProperties;
-
+import w2phtml.util.DimensionsConverter;
 
 /** This class is used to represent a binary graphics document to be included in the converter result.
  *  I may also represent a linked image, which should <em>not</em> be included (and will produce an empty file
@@ -49,6 +62,9 @@ public class BinaryGraphicsDocument implements OutputFile {
     private byte[] blob = null;
     private int nOff = 0;
     private int nLen = 0;
+		private boolean cropped = false;
+		private float horizontalPPI = 96.0F;
+		private float verticalPPI = 96.0F;
     
     /**Constructs a new graphics document.
      * Until data is added using the <code>read</code> methods, the document is considered a link to
@@ -75,9 +91,27 @@ public class BinaryGraphicsDocument implements OutputFile {
     	this.blob = bgd.getData();
     	this.bAcceptedFormat = bgd.isAcceptedFormat();
     	this.bRecycled = true;
+  	
     }
     
-    /** Is this graphics document recycled?
+    private void extractPPI() {
+    	try {
+				ImageInfo info = Imaging.getImageInfo(blob);
+				int heightDPI = info.getPhysicalHeightDpi();
+				if (heightDPI > 0) {
+					this.verticalPPI = heightDPI;
+				}
+				int widthDPI = info.getPhysicalWidthDpi();
+				if (widthDPI > 0) {
+					this.horizontalPPI = widthDPI;
+				}
+			} catch (ImageReadException | IOException e) {
+				e.printStackTrace();
+			}
+      
+		}
+  
+		/** Is this graphics document recycled?
      * 
      * @return true if this is the case
      */
@@ -170,8 +204,76 @@ public class BinaryGraphicsDocument implements OutputFile {
     public boolean isMasterDocument() {
 		return false;
 	}
-    
+
-    /** Does this document contain formulas?
+	public boolean isCropped() {
+		return cropped;
+	}
+
+	/* Crop image from byte array */
+	public void cropImage(StyleWithProperties style) {
+		if (blob == null || style == null) {
+			return;
+		}
+		if (sMimeType == null || ( 
+				   !sMimeType.equals(MIMETypes.PNG) 
+				&& !sMimeType.equals(MIMETypes.JPEG)
+				&& !sMimeType.equals(MIMETypes.GIF)
+				&& !sMimeType.equals(MIMETypes.TIFF)
+				)){
+			return;
+		}
+		extractPPI();
+		int[] offsets = getOffsets(style);
+		ByteArrayInputStream bis = new ByteArrayInputStream(blob);
+		
+		try {
+			BufferedImage image = ImageIO.read(bis);
+			int height = image.getHeight();
+			int width = image.getWidth();
+			String formatName = MIMETypes.getFormatType(blob);
+			int xTopLeft = offsets[3];
+			int yTopLeft = offsets[0];
+			int xBottomRight = width - offsets[3] - offsets[1];
+			int yBottomRight = height - offsets[0] - offsets[2];
+			BufferedImage croppedImage = image.getSubimage(xTopLeft, yTopLeft, xBottomRight , yBottomRight );
+			ByteArrayOutputStream baos = new ByteArrayOutputStream();
+			ImageIO.write(croppedImage, "JPG", baos);
+			sMimeType = MIMETypes.JPEG;
+			blob = baos.toByteArray();
+			cropped = true;
+		} catch (IOException e) {
+			// TODO Auto-generated catch block
+			e.printStackTrace();
+		} catch (RasterFormatException e ) {
+			e.printStackTrace();
+		}
+
+	}
+	
+	private int[] getOffsets(StyleWithProperties style) {
+		/* top, right, bottom, left */
+		int[] offsets = { 0, 0, 0, 0 };
+		String fo_clip = style.getProperty(FO_CLIP);
+			if (!fo_clip.startsWith("rect(") || !fo_clip.endsWith(")")) {
+				return offsets;
+			}
+			fo_clip = fo_clip.replaceAll("^rect\\(", "");
+			fo_clip = fo_clip.replaceAll("\\)", "");
+			String[] inputOffsets = fo_clip.split("\\s*,\\s*");
+			for (int i = 0; i < inputOffsets.length; i++) {
+				String inputOffset = inputOffsets[i];
+				float ppi;
+				if (i % 2 == 0) {
+					ppi = verticalPPI;
+				} else {
+					ppi = horizontalPPI;
+				}
+				offsets[i] = Math.round(DimensionsConverter.length2px(inputOffset, ppi));
+			}
+			return offsets;
+		}
+
+		/** Does this document contain formulas?
      * 
      *  @return false - a graphics file does not contain formulas
      */

src/main/java/w2phtml/office/MIMETypes.java

@@ -124,6 +124,18 @@ public final class MIMETypes extends w2phtml.api.MIMETypes {
         if (isSVG(blob)) { return SVG; }
         return "";
     }
+    
+    public static final String getFormatType(byte[] blob) {
+      if (isType(blob,PNG_SIG)) { return "PNG"; }
+      if (isType(blob,JPEG_SIG)) { return "JPG"; }
+      if (isType(blob,JPEG_EXIF_SIG)) { return "JPG"; }
+      if (isType(blob,GIF87_SIG)) { return "GIF"; }
+      if (isType(blob,GIF89_SIG)) { return "GIF"; }
+      if (isType(blob,TIFF_SIG)) { return "TIFF"; }
+      if (isType(blob,BMP_SIG)) { return "BMP"; }
+      if (isSVG(blob)) { return "SVG"; }
+      return "";
+  }
 	
     public static final String getFileExtension(String sMIME) {
         if (PNG.equals(sMIME)) { return PNG_EXT; }

src/main/java/w2phtml/office/XMLString.java

@@ -434,6 +434,7 @@ public class XMLString {
 	// fo namespace
 	public static final String FO_LANGUAGE = "fo:language";
 	public static final String FO_COUNTRY = "fo:country";
+	public static final String FO_CLIP = "fo:clip";
 	public static final String FO_TEXT_SHADOW = "fo:text-shadow";
 	public static final String FO_COLOR = "fo:color";
 	public static final String FO_BACKGROUND_COLOR = "fo:background-color";

src/main/java/w2phtml/xhtml/content/DrawParser.java

@@ -539,6 +539,8 @@ public class DrawParser extends Parser {
         String sFileName = null;
         bgd = converter.getImageCv().getImage(onode);
         if (bgd!=null) {
+      		StyleWithProperties style = ofr.getFrameStyle(frame.getAttribute(XMLString.DRAW_STYLE_NAME));
+        	bgd.cropImage(style);
         	if (!bgd.isLinked()) { // embedded image
                 sFileName = bgd.getFileName();
                 // If this is the cover image, add it to the converter result