[Archive-access-cvs] SF.net SVN: archive-access:[2687] trunk/archive-access/projects/nutchwax/ archive/src/java/org/archive/nutchwax/OpenSearchServlet.java

[<bi...@us...>]

Revision: 2687
          http://archive-access.svn.sourceforge.net/archive-access/?rev=2687&view=rev
Author:   binzino
Date:     2009-03-03 18:20:14 +0000 (Tue, 03 Mar 2009)

Log Message:
-----------
Fixed handling of start and end of search results so that we detect
"paging off the end" and return an empty result set rather than an
exception.

Modified Paths:
--------------
    trunk/archive-access/projects/nutchwax/archive/src/java/org/archive/nutchwax/OpenSearchServlet.java

Modified: trunk/archive-access/projects/nutchwax/archive/src/java/org/archive/nutchwax/OpenSearchServlet.java
===================================================================
--- trunk/archive-access/projects/nutchwax/archive/src/java/org/archive/nutchwax/OpenSearchServlet.java	2009-02-28 01:26:25 UTC (rev 2686)
+++ trunk/archive-access/projects/nutchwax/archive/src/java/org/archive/nutchwax/OpenSearchServlet.java	2009-03-03 18:20:14 UTC (rev 2687)
@@ -162,18 +162,30 @@
 
     responseTime = System.nanoTime( ) - responseTime;
 
-    // generate xml results
-    int end = (int)Math.min(hits.getLength(), start + hitsPerPage);
-    int length = end-start;
+    // The 'end' is usually just the end of the current page
+    // (start+hitsPerPage); but if we are on the last page
+    // of de-duped results, then the end is hits.getLength().
+    int end = Math.min( hits.getLength( ), start + hitsPerPage );
 
-    Hit[] show = hits.getHits(start, end-start);
-    HitDetails[] details = bean.getDetails(show);
-    Summary[] summaries = bean.getSummary(details, query);
+    // The length is usually just (end-start), unless the start
+    // position is past the end of the results -- which is common when
+    // de-duping.  The user could easily jump past the true end of the
+    // de-dup'd results.  If the start is past the end, we use a
+    // length of '0' to produce an empty results page.
+    int length = Math.max( end-start, 0 );
 
+    // Usually, the total results is the total number of non-de-duped
+    // results.  Howerver, if we are on last page of de-duped results,
+    // then we know our de-dup'd total is hits.getLength().
+    long totalResults = hits.getLength( ) < (start+hitsPerPage) ? hits.getLength( ) : hits.getTotal( );
+
+    Hit[]        show      = hits.getHits(start, length );
+    HitDetails[] details   = bean.getDetails(show);
+    Summary[]    summaries = bean.getSummary(details, query);
+
     String requestUrl = request.getRequestURL().toString();
     String base = requestUrl.substring(0, requestUrl.lastIndexOf('/'));
       
-
     try {
       DocumentBuilderFactory factory = DocumentBuilderFactory.newInstance();
       factory.setNamespaceAware(true);
@@ -197,8 +209,8 @@
               +"&hitsPerDup="+hitsPerDup
               +params);
 
-      addNode(doc, channel, "opensearch", "totalResults", ""+hits.getTotal());
-      addNode(doc, channel, "opensearch", "startIndex", ""+start);
+      addNode(doc, channel, "opensearch", "totalResults", ""+totalResults);
+      addNode(doc, channel, "opensearch", "startIndex",   ""+start);
       addNode(doc, channel, "opensearch", "itemsPerPage", ""+hitsPerPage);
 
       addNode(doc, channel, "nutch", "query", queryString);


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.