Add cm:geographic Aspect, which has cm:latitude and cm:longitude, and update the Tika auto parser to map to this (plus tests)

git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@20925 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261

config/alfresco/model/contentModel.xml

@@ -1175,6 +1175,20 @@
       <!--  DEPRECATED (end of) -->
       <!--                      -->
 
+      <aspect name="cm:geographic">
+         <title>Geographic</title>
+         <properties>
+            <property name="cm:latitude">
+               <title>Latitude</title>
+               <type>d:double</type>
+            </property>
+            <property name="cm:longitude">
+               <title>Longitude</title>
+               <type>d:double</type>
+            </property>
+         </properties>
+      </aspect>
+      
    </aspects>
 
 </model>

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

@@ -33,12 +33,11 @@ import org.apache.tika.parser.pdf.PDFParser;
  *   <b>title:</b>                  --      cm:title
  *   <b>subject:</b>                --      cm:description
  *   <b>created:</b>                --      cm:created
+ *   <b>(custom metadata):</b>      --
  * </pre>
  * 
  * Uses Apache Tika
  * 
- * TODO - Update Tika to handle custom metadata
- * 
  * @author Jesper Steen Møller
  * @author Derek Hulley
  */

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

@@ -40,6 +40,8 @@ import org.apache.tika.parser.Parser;
  *   <b>subject:</b>                --      cm:description
  *   <b>created:</b>                --      cm:created
  *   <b>comments:</b>
+ *   <p>geo:lat:</b>                --      cm:latitude
+ *   <p>geo:long:</b>               --      cm:longitude
  * </pre>
  * 
  * @author Nick Burch

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

@@ -16,3 +16,8 @@ author=cm:author
 title=cm:title
 description=cm:description
 created=cm:created
+
+geo\:lat=cm:latitude
+geo\:long=cm:longitude
+#tiff\:ImageWidth=cm:imageWidth
+#tiff\:ImageLength=cm:imageHeight

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

@@ -210,16 +210,38 @@ public class TikaAutoMetadataExtracterTest extends AbstractMetadataExtracterTest
       assertEquals("409", p.get("width"));
       assertEquals("92", p.get("height"));
       assertEquals("8 8 8", p.get("Data BitsPerSample"));
+      
+      
+      // Geo tagged image
+      p = openAndCheck("GEO.jpg", "image/jpeg");
+      assertEquals("100 pixels", p.get("Image Width"));
+      assertEquals("68 pixels", p.get("Image Height"));
+      assertEquals("8 bits", p.get("Data Precision"));
+      assertEquals(QUICK_TITLE, p.get("Comments"));
+      assertEquals("12.54321", p.get("geo:lat"));
+      assertEquals("-54.1234", p.get("geo:long"));
+      
+      // Map and check
+      Map<QName, Serializable> properties = new HashMap<QName, Serializable>();
+      ContentReader reader = new FileContentReader(open("GEO.jpg"));
+      reader.setMimetype("image/jpeg");
+      extracter.extract(reader, properties);
+      assertEquals(12.54321, properties.get(QName.createQName("http://www.alfresco.org/model/content/1.0","latitude")));
+      assertEquals(-54.1234, properties.get(QName.createQName("http://www.alfresco.org/model/content/1.0","longitude")));
    }
-   private Map<String, Serializable> openAndCheck(String fileBase, String expMimeType) throws Throwable {
+   private File open(String fileBase) throws Throwable {
       String filename = "quick" + fileBase;
       URL url = AbstractContentTransformerTest.class.getClassLoader().getResource("quick/" + filename);
       File file = new File(url.getFile());
-      
+      assertTrue(file.exists());
+      return file;
+   }
+   private Map<String, Serializable> openAndCheck(String fileBase, String expMimeType) throws Throwable {
       // Cheat and ask Tika for the mime type!
+      File file = open(fileBase);
       AutoDetectParser ap = new AutoDetectParser();
       Metadata metadata = new Metadata();
-      metadata.set(Metadata.RESOURCE_NAME_KEY, filename);
+      metadata.set(Metadata.RESOURCE_NAME_KEY, "quick"+fileBase);
       MediaType mt = ap.getDetector().detect(
             new BufferedInputStream(new FileInputStream(file)), metadata);
       String mimetype = mt.toString();

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

@@ -18,6 +18,7 @@
  */
 package org.alfresco.repo.content.metadata;
 
+import java.io.File;
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.Serializable;
@@ -31,9 +32,11 @@ import java.util.HashSet;
 import java.util.Locale;
 import java.util.Map;
 
+import org.alfresco.repo.content.filestore.FileContentReader;
 import org.alfresco.service.cmr.repository.ContentReader;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
+import org.apache.tika.io.TikaInputStream;
 import org.apache.tika.metadata.Metadata;
 import org.apache.tika.mime.MediaType;
 import org.apache.tika.parser.ParseContext;
@@ -109,6 +112,7 @@ public abstract class TikaPoweredMetadataExtracter extends AbstractMappingMetada
     {
         super(supportedMimeTypes);
         
+        // TODO Once TIKA-451 is fixed this list will get nicer
         this.tikaDateFormats = new DateFormat[] {
               new SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ssZ"),
               new SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ssZ", Locale.US),
@@ -116,6 +120,10 @@ public abstract class TikaPoweredMetadataExtracter extends AbstractMappingMetada
               new SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ss", Locale.US),
               new SimpleDateFormat("yyyy-MM-dd"),
               new SimpleDateFormat("yyyy-MM-dd", Locale.US),
+              new SimpleDateFormat("yyyy/MM/dd HH:mm:ss"),
+              new SimpleDateFormat("yyyy/MM/dd HH:mm:ss", Locale.US),
+              new SimpleDateFormat("yyyy/MM/dd"),
+              new SimpleDateFormat("yyyy/MM/dd", Locale.US),
               new SimpleDateFormat("EEE MMM dd hh:mm:ss zzz yyyy"),
               new SimpleDateFormat("EEE MMM dd hh:mm:ss zzz yyyy", Locale.US)
         };
@@ -169,6 +177,28 @@ public abstract class TikaPoweredMetadataExtracter extends AbstractMappingMetada
        return properties;
     }
     
+    /**
+     * There seems to be some sort of issue with some downstream
+     *  3rd party libraries, and input streams that come from
+     *  a {@link ContentReader}. This happens most often with
+     *  JPEG and Tiff files.
+     * For these cases, buffer out to a local file if not
+     *  already there
+     */
+    private InputStream getInputStream(ContentReader reader) throws IOException {
+       if("image/jpeg".equals(reader.getMimetype()) ||
+             "image/tiff".equals(reader.getMimetype())) {
+          if(reader instanceof FileContentReader) {
+             return TikaInputStream.get( ((FileContentReader)reader).getFile() );
+          } else {
+             File tmpFile = File.createTempFile("tika", "tmp");
+             reader.getContent(tmpFile);
+             return TikaInputStream.get(tmpFile);
+          }
+       }
+       return reader.getContentInputStream(); 
+    }
+    
     @Override
     protected Map<String, Serializable> extractRaw(ContentReader reader) throws Throwable
     {
@@ -177,7 +207,7 @@ public abstract class TikaPoweredMetadataExtracter extends AbstractMappingMetada
         InputStream is = null;
         try
         {
-            is = reader.getContentInputStream();
+            is = getInputStream(reader); 
             Parser parser = getParser();
             Metadata metadata = new Metadata();
             ParseContext context = new ParseContext();