From Bug 38289: Yegor's support for pictures embeded in PPT files

git-svn-id: https://svn.apache.org/repos/asf/jakarta/poi/trunk@369540 13f79535-47bb-0310-9956-ffa450edef68

diff --git a/src/scratchpad/src/org/apache/poi/hslf/HSLFSlideShow.java b/src/scratchpad/src/org/apache/poi/hslf/HSLFSlideShow.java
index 79e28bf5a8072960e94a7261811f8547d609110e..f12896a140557174e750caf608bc9e3a23cc2bfa 100644 (file)
--- a/src/scratchpad/src/org/apache/poi/hslf/HSLFSlideShow.java
+++ b/src/scratchpad/src/org/apache/poi/hslf/HSLFSlideShow.java
@@ -36,6 +36,7 @@ import org.apache.poi.hpsf.DocumentSummaryInformation;
 import org.apache.poi.util.LittleEndian;
 
 import org.apache.poi.hslf.record.*;
+import org.apache.poi.hslf.usermodel.Picture;
 
 /**
  * This class contains the main functionality for the Powerpoint file 
@@ -337,4 +338,34 @@ public class HSLFSlideShow
   * Fetch the Current User Atom of the document
   */
  public CurrentUserAtom getCurrentUserAtom() { return currentUser; }
+
+       /**
+        *  Read pictures contained in this presentation
+        *
+        *  @return array with the read pictures ot <code>null</code> if the
+        *  presentation doesn't contain pictures.
+        */
+       public Picture[] getPictures() throws IOException {
+               byte[] pictstream;
+
+               try {
+                       DocumentEntry entry = (DocumentEntry)filesystem.getRoot().getEntry("Pictures");
+                       pictstream = new byte[entry.getSize()];
+                       DocumentInputStream is = filesystem.createDocumentInputStream("Pictures");
+                       is.read(pictstream);
+               } catch (FileNotFoundException e){
+                       //silently catch exceptions if the presentation doesn't contain pictures
+                       return null;
+               }
+
+               ArrayList p = new ArrayList();
+               int pos = 0; 
+               while (pos < pictstream.length) {
+                       Picture pict = new Picture(pictstream, pos);
+                       p.add(pict);
+                       pos += Picture.HEADER_SIZE + pict.getSize();
+               }
+
+               return (Picture[])p.toArray(new Picture[p.size()]);
+       }
 }

diff --git a/src/scratchpad/src/org/apache/poi/hslf/usermodel/Picture.java b/src/scratchpad/src/org/apache/poi/hslf/usermodel/Picture.java
new file mode 100644 (file)
index 0000000..64caff7
--- /dev/null
+++ b/src/scratchpad/src/org/apache/poi/hslf/usermodel/Picture.java
@@ -0,0 +1,140 @@
+/* ====================================================================\r
+   Copyright 2002-2004   Apache Software Foundation\r
+\r
+   Licensed under the Apache License, Version 2.0 (the "License");\r
+   you may not use this file except in compliance with the License.\r
+   You may obtain a copy of the License at\r
+\r
+       http://www.apache.org/licenses/LICENSE-2.0\r
+\r
+   Unless required by applicable law or agreed to in writing, software\r
+   distributed under the License is distributed on an "AS IS" BASIS,\r
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\r
+   See the License for the specific language governing permissions and\r
+   limitations under the License.\r
+==================================================================== */\r
+package org.apache.poi.hslf.usermodel;\r
+\r
+import org.apache.poi.util.LittleEndian;\r
+\r
+/**\r
+ * Represents a picture in a PowerPoint document.\r
+ * <p>\r
+ * The information about an image in PowerPoint document is stored in \r
+ * two places:\r
+ *  <li> EscherBSE container in the Document keeps information about image \r
+ *    type, image index to refer by slides etc.\r
+ *  <li> "Pictures" OLE stream holds the actual data of the image.\r
+ * </p>\r
+ * <p>\r
+ *  Data in the "Pictures" OLE stream is organized as follows:<br>\r
+ *  For each image there is an entry: 25 byte header + image data.\r
+ *  Image data is the exact content of the JPEG file, i.e. PowerPoint\r
+ *  puts the whole jpeg file there without any modifications.<br>\r
+ *   Header format:\r
+ *    <li> 2 byte: image type. For JPEGs it is 0x46A0, for PNG it is 0x6E00.\r
+ *    <li> 2 byte: unknown.\r
+ *    <li> 4 byte : image size + 17. Looks like shift from the end of \r
+ *          header but why to add it to the image  size?\r
+ *    <li> next 16 bytes. Unique identifier of this image which is used by \r
+ *          EscherBSE record.\r
+ *  </p>\r
+ *\r
+ * @author Yegor Kozlov\r
+ */\r
+public class Picture {\r
+\r
+       /**\r
+       *  Windows Metafile\r
+       */\r
+       public static final int WMF = 0x2160;\r
+\r
+       /**\r
+       * Macintosh PICT\r
+       */\r
+       public static final int PICT = 0x5420;\r
+\r
+       /**\r
+       *  JPEG\r
+       */\r
+       public static final int JPEG = 0x46A0;\r
+\r
+       /**\r
+       *  PNG\r
+       */\r
+       public static final int PNG = 0x6E00;\r
+\r
+       /**\r
+       * Windows DIB (BMP)\r
+       */\r
+       public static final int DIB = 0x7A80;\r
+\r
+       /**\r
+       * The size of the header\r
+       */\r
+       public static final int HEADER_SIZE = 25;\r
+\r
+       /**\r
+       * Binary data of the picture\r
+       */\r
+       protected byte[] pictdata;\r
+\r
+       /**\r
+       * Header which holds information about this picture\r
+       */\r
+       protected byte[] header;\r
+\r
+       /**\r
+       * Read a picture from "Pictures" OLE stream\r
+       *\r
+       * @param pictstream    the bytes to read\r
+       * @param offset        the index of the first byte to read\r
+       */\r
+       public Picture(byte[] pictstream, int offset){\r
+               header = new byte[Picture.HEADER_SIZE];\r
+               System.arraycopy(pictstream, offset, header, 0, header.length);\r
+\r
+               int size = LittleEndian.getInt(header, 4) - 17;\r
+               pictdata = new byte[size];\r
+               System.arraycopy(pictstream, offset + Picture.HEADER_SIZE, pictdata, 0, pictdata.length);\r
+       }\r
+\r
+       /**\r
+       * @return  the binary data of this picture\r
+       */\r
+       public byte[] getData(){\r
+               return pictdata;\r
+       }\r
+\r
+       /**\r
+       * Return image size in bytes\r
+       *\r
+       * @return the size of the picture in bytes\r
+       */\r
+       public int getSize(){\r
+               return pictdata.length;\r
+       }\r
+\r
+       /**\r
+       * Returns the unique identifier (UID) of this picture.\r
+       * The UID is a checksum of the picture data. Its length is 16 bytes\r
+       * and it must be unique across the presentation.\r
+       *\r
+       * @return the unique identifier of this picture\r
+       */\r
+       public byte[] getUID(){\r
+               byte[] uid = new byte[16];\r
+               System.arraycopy(header, 8, uid, 0, uid.length);\r
+               return uid;\r
+       }\r
+\r
+       /**\r
+       * Returns the type of this picture. Must be one of the static constans defined in this class.\r
+       *\r
+       * @return type of this picture.\r
+       */\r
+       public int getType(){\r
+               int type = LittleEndian.getShort(header, 0);\r
+               return type;\r
+       }\r
+}\r

diff --git a/src/scratchpad/src/org/apache/poi/hslf/usermodel/SlideShow.java b/src/scratchpad/src/org/apache/poi/hslf/usermodel/SlideShow.java
index 15b610ce09b97719d83224ad81f19217e9f0b814..23108d4783e1d287a705f42a7da632969dcb8238 100644 (file)
--- a/src/scratchpad/src/org/apache/poi/hslf/usermodel/SlideShow.java
+++ b/src/scratchpad/src/org/apache/poi/hslf/usermodel/SlideShow.java
@@ -391,4 +391,11 @@ public class SlideShow
    * found in the slideshow
    */
   //public MetaSheet[] getMetaSheets() { return _msheets; }
+
+       /**
+        * Returns all the pictures attached to the SlideShow
+        */
+       public Picture[] getPictures() throws IOException {
+               return _hslfSlideShow.getPictures();
+       }
 }