Merged HEAD-BUG-FIX (5.1/Cloud) to HEAD (5.1/Cloud)

104496: Merged 5.0.N (5.0.2) to HEAD-BUG-FIX (5.1/Cloud)
      104336: Merged NESS/5.0.N-2015_03_23 (5.0.2) to 5.0.N (5.0.2)
         103763: MNT-13920 - rewrite the image dimension properties if there is any exif dimensions information available
         104332: MNT-13920 - code changes based on review, improved javadoc and slight modifications on the extract size method


git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@104607 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261

source/java/org/alfresco/repo/content/metadata/TikaAutoMetadataExtracter.java

@@ -18,11 +18,16 @@
  */
 package org.alfresco.repo.content.metadata;
 
+import java.io.Serializable;
 import java.util.ArrayList;
+import java.util.Map;
 
+import org.alfresco.repo.content.MimetypeMap;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.tika.config.TikaConfig;
+import org.apache.tika.metadata.Metadata;
+import org.apache.tika.metadata.TIFF;
 import org.apache.tika.mime.MediaType;
 import org.apache.tika.parser.AutoDetectParser;
 import org.apache.tika.parser.Parser;
@@ -53,6 +58,10 @@ public class TikaAutoMetadataExtracter extends TikaPoweredMetadataExtracter
     protected static Log logger = LogFactory.getLog(TikaAutoMetadataExtracter.class);
     private static AutoDetectParser parser;
     private static TikaConfig config;
+    private static String EXIF_IMAGE_HEIGHT_TAG = "Exif Image Height";
+    private static String EXIF_IMAGE_WIDTH_TAG = "Exif Image Width";
+    private static String JPEG_IMAGE_HEIGHT_TAG = "Image Height";
+    private static String JPEG_IMAGE_WIDTH_TAG = "Image Width";
 
     public static ArrayList<String> SUPPORTED_MIMETYPES;
     private static ArrayList<String> buildMimeTypes(TikaConfig tikaConfig)
@@ -90,4 +99,33 @@ public class TikaAutoMetadataExtracter extends TikaPoweredMetadataExtracter
     {
        return parser;
     }
+    
+    /**
+     * Because some editors use JPEG_IMAGE_HEIGHT_TAG when
+     * saving JPEG images , a more reliable source for
+     * image size are the values provided by Tika
+     * and not the exif/tiff metadata read from the file
+     * This will override the tiff:Image size 
+     * which gets embedded into the alfresco node properties
+     * for jpeg files that contain such exif information
+     */
+    @Override
+    protected Map<String, Serializable> extractSpecific(Metadata metadata,
+         Map<String, Serializable> properties, Map<String,String> headers) 
+    {
+        
+        if(MimetypeMap.MIMETYPE_IMAGE_JPEG.equals(metadata.get(Metadata.CONTENT_TYPE)))                
+        {
+            //check if the image has exif information
+            if(metadata.get(EXIF_IMAGE_WIDTH_TAG) != null && metadata.get(EXIF_IMAGE_HEIGHT_TAG) != null )
+            {    
+                //replace the exif size properties that will be embedded in the node with
+                //the guessed dimensions from Tika
+                putRawValue(TIFF.IMAGE_LENGTH.getName(), extractSize(metadata.get(JPEG_IMAGE_HEIGHT_TAG)), properties);
+                putRawValue(TIFF.IMAGE_WIDTH.getName(), extractSize(metadata.get(JPEG_IMAGE_WIDTH_TAG)), properties);
+            }
+        }
+        return properties;
+    }
+    
 }

source/java/org/alfresco/repo/content/metadata/TikaPoweredMetadataExtracter.java

@@ -530,6 +530,30 @@ public abstract class TikaPoweredMetadataExtracter
         }
     }
     
+    /**
+     * Exif metadata for size also returns the string "pixels"
+     * after the number value , this function will 
+     * stop at the first non digit character found in the text
+     * @param sizeText string text
+     * @return the size value
+     */
+    protected String extractSize(String sizeText)
+    {
+        StringBuilder sizeValue = new StringBuilder();
+        for(char c : sizeText.toCharArray())
+        {
+            if(Character.isDigit(c))
+            {
+                sizeValue.append(c);
+            }
+            else
+            {
+                break;
+            }
+        }
+        return sizeValue.toString();
+    }
+    
     /**
      * This content handler will capture entries from within
      *  the header of the Tika content XHTML, but ignore the

source/test-java/org/alfresco/repo/content/metadata/TikaAutoMetadataExtracterTest.java

@@ -250,6 +250,20 @@ public void testImageVideo() throws Throwable {
       assertEquals("92", p.get("height"));
       assertEquals("8 8 8", p.get("Data BitsPerSample"));
       
+      // Image with wrong tiff:Width property. see MNT-13920 
+      p = openAndCheck("SizeSample.jpg", "image/jpeg");
+      // Check raw EXIF properties
+      assertEquals("1535 pixels", p.get("Image Width"));
+      assertEquals("367 pixels", p.get("Image Height"));
+      
+      // Map and check
+      Map<QName, Serializable> propsJPG = new HashMap<QName, Serializable>();
+      ContentReader readerJPG = new FileContentReader(open("SizeSample.jpg"));
+      readerJPG.setMimetype("image/jpeg");
+      extracter.extract(readerJPG, propsJPG);
+      assertEquals(1535, propsJPG.get(QName.createQName(NamespaceService.EXIF_MODEL_1_0_URI, "pixelXDimension")));
+      assertEquals(367, propsJPG.get(QName.createQName(NamespaceService.EXIF_MODEL_1_0_URI, "pixelYDimension")));
+      
       
       // Geo tagged image
       p = openAndCheck("GEO.jpg", "image/jpeg");