Changeset 26214 for gs2-extensions/video-and-audio/trunk/src/opt/Terrier/ImageDocument.java

Timestamp:

2012-09-19T12:06:28+12:00 (12 years ago)

Author:

jmt12

Message:

New hash based generation for associated files directory - so docno is no longer essential

File:

• gs2-extensions/video-and-audio/trunk/src/opt/Terrier/ImageDocument.java (modified) (3 diffs)

gs2-extensions/video-and-audio/trunk/src/opt/Terrier/ImageDocument.java

@@ -20 +20 @@
 import java.io.StringReader;
 import java.io.Reader;
+import java.nio.charset.Charset;
 import java.nio.file.Files;
 import java.nio.file.Path;
 import java.nio.file.Paths;
+import java.security.MessageDigest;
+import java.security.NoSuchAlgorithmException;
 import java.util.Collections;
 import java.util.Arrays;
@@ -79 +82 @@
     this.properties.put("abstract", "This is an image so here is some dummy text to prevent indexer failing.");
     // B. Properties derived from filename
-    String title = this.properties.get("filename").substring(properties.get("filename").lastIndexOf(System.getProperty("file.separator")) + 1);
+    // - A simple title for the document
+    String filepath = this.properties.get("filename");
+    String title = filepath.substring(filepath.lastIndexOf(System.getProperty("file.separator")) + 1);
     this.properties.put("title", title);
-    String ext = this.properties.get("filename").substring(properties.get("filename").lastIndexOf(".") + 1);
+    String ext = filepath.substring(filepath.lastIndexOf(".") + 1);
+    // - The name of the copy of the original document
     String target_filename = "doc." + ext;
     this.properties.put("source","doc." + ext);
-    String assoc_filename = "D" + this.properties.get("docno");
-    if (assoc_filename.equals("Dnull"))
-    {
-      System.err.println("Error! Bogus assoc dir: " + this.properties.get("docno"));
+    // - A unique associated directory. This gets a little tricky as we need
+    //   to create the directory at the same time if an effort to promote
+    //   synchronous behaviour
+    String unique_id = this.generateHash(filepath);
+    //   - we start with the first 4 characters
+    int offset = 0;
+    String assoc_filename = "D" + unique_id.substring(offset, offset + 4);
+    //   - we add ".dir" as a suffix to the directory that actually contains
+    //     files (so the non-suffixed version contains nested directories)
+    Path assoc_path = Paths.get(ApplicationSetup.TERRIER_SHARE, "images", "assoc", assoc_filename + ".dir");
+    //   - then we continue adding blocks of 4 characters until we get a
+    //     directory that doesn't already exist
+    while (assoc_path.toFile().exists() && offset < unique_id.length())
+    {
+      offset += 4;
+      assoc_filename += System.getProperty("file.separator") + unique_id.substring(offset, offset + 4);
+      assoc_path = Paths.get(ApplicationSetup.TERRIER_SHARE, "images", "assoc", assoc_filename + ".dir");
+    }
+    //   - still not unique? but run out of unique_id... time to complain
+    if (assoc_path.toFile().exists())
+    {
+      logger.error("ImageDoument - can't determine unique assocfilepath");
       System.exit(0);
     }
-
+    //   - create the directories quick... hopefully before someone else does
+    assoc_path.toFile().mkdirs(); // bet there is a nice nio way to do this
     this.properties.put("assocfile", assoc_filename);
 
     // Copy (symlink) the file into place in the shared directory
     Path source_path = Paths.get(properties.get("filename"));
-    Path assoc_path = Paths.get(ApplicationSetup.TERRIER_SHARE, "images", "assoc", assoc_filename);
-    assoc_path.toFile().mkdirs(); // bet there is a nice nio way to do this
     Path target_path = assoc_path.resolve(target_filename);
     if (target_path.toFile().exists())
@@ -264 +287 @@
   }
   /** getReader() **/
+
+  /**
+   */
+  private String generateHash(String string)
+  {
+    StringBuffer sb = new StringBuffer();
+    try
+    {
+      final MessageDigest message_digest = MessageDigest.getInstance("MD5");
+      message_digest.reset();
+      message_digest.update(string.getBytes(Charset.forName("UTF8")));
+      final byte[] result_bytes = message_digest.digest();
+      for (int i = 0; i < result_bytes.length; ++i)
+      {
+        sb.append(Integer.toHexString((result_bytes[i] & 0xFF) | 0x100).substring(1,3));
+      }
+    }
+    catch (NoSuchAlgorithmException e)
+    {
+      System.err.println("Exception: " + e);
+      System.exit(0);
+    }
+    return sb.toString();
+  }
+  /** generateHash(String) **/
 }