From 7d14c52c122e301219eac523028846c0046c6b5e Mon Sep 17 00:00:00 2001
From: anupsawant <anupsawant@440791e1-c2bf-4c1e-ad7c-caf3f8b0ebe8>
Date: Wed, 25 May 2011 21:43:32 +0000
Subject: [PATCH]

---
 solr/exampleSolr/conf/schema.xml              |  9 ++++---
 solr/exampleSolr/conf/solrconfig.xml          |  3 ++-
 .../vitro/webapp/search/VitroTermNames.java   |  4 +++
 .../search/beans/SearchQueryHandler.java      |  1 +
 .../controller/SolrPagedSearchController.java |  1 +
 .../search/solr/IndividualToSolrDocument.java | 26 +++++++++++--------
 .../vitro/webapp/search/solr/SolrIndexer.java |  2 +-
 7 files changed, 30 insertions(+), 16 deletions(-)

diff --git a/solr/exampleSolr/conf/schema.xml b/solr/exampleSolr/conf/schema.xml
index 7cd9c70c8..89dd80135 100644
--- a/solr/exampleSolr/conf/schema.xml
+++ b/solr/exampleSolr/conf/schema.xml
@@ -381,7 +381,7 @@
     <fieldtype name="phonetic" stored="false" indexed="true" class="solr.TextField" >
       <analyzer>
         <tokenizer class="solr.StandardTokenizerFactory"/>
-        <filter class="solr.DoubleMetaphoneFilterFactory" inject="false"/>
+        <filter class="solr.PhoneticFilterFactory" encoder="Metaphone" inject="false"/>
       </analyzer>
     </fieldtype>
 
@@ -493,9 +493,11 @@
 <field name="acNameUnstemmed" type="textUnstemmed" indexed="true" stored="false" multiValued="true"/>
 <field name="acNameStemmed" type="text" indexed="true" stored="false" multiValued="true"/>
 <field name="indexedTime" type="string" indexed="true" stored="true" multiValued="true"/>
+<field name="NAME_PHONETIC" type ="phonetic" indexed="true" stored="false" multiValued="true"/>
 
 <field name="ALLTEXT" type="text" indexed="true" stored="false" multiValued="true"/>
 <field name="ALLTEXTUNSTEMMED" type="textgen" indexed="true" stored="false" multiValued="true"/>
+<field name="ALLTEXT_PHONETIC" type="phonetic" indexed="true" stored="false" multiValued="true"/>
 
 <field name="THUMBNAIL" type="string" indexed="true" stored="true"/>
 
@@ -577,7 +579,7 @@
  <uniqueKey>DocId</uniqueKey>
 
  <!-- field for the QueryParser to use when an explicit fieldname is absent -->
- <!-- <defaultSearchField>ALLTEXT</defaultSearchField> -->
+ <defaultSearchField>ALLTEXT</defaultSearchField>
 
  <!-- SolrQueryParser configuration: defaultOperator="AND|OR" -->
  <solrQueryParser defaultOperator="AND"/>
@@ -600,7 +602,8 @@
  <!-- Similarity is the scoring routine for each document vs. a query.
       A custom similarity may be specified here, but the default is fine
       for most applications.  -->
- <!-- <similarity class="org.apache.lucene.search.DefaultSimilarity"/> -->
+<similarity class="org.apache.lucene.search.DefaultSimilarity"/>
+<!-- <similarity class="edu.cornell.mannlib.vitro.webapp.search.lucene.CustomSimilarity"/> -->
  <!-- ... OR ...
       Specify a SimilarityFactory class name implementation
       allowing parameters to be used.
diff --git a/solr/exampleSolr/conf/solrconfig.xml b/solr/exampleSolr/conf/solrconfig.xml
index 30aff6ca6..d341bc202 100644
--- a/solr/exampleSolr/conf/solrconfig.xml
+++ b/solr/exampleSolr/conf/solrconfig.xml
@@ -708,8 +708,9 @@
       -->
      <lst name="defaults">
        <str name="defType">dismax</str>
-       <str name="qf">nameRaw nameLowercase nameUnstemmed nameStemmed ALLTEXT ALLTEXTUNSTEMMED</str>
+       <str name="qf">nameRaw nameLowercase nameUnstemmed nameStemmed ALLTEXT ALLTEXTUNSTEMMED ALLTEXT_PHONETIC NAME_PHONETIC</str>
        <str name="echoParams">explicit</str>
+       <str name="bf">BETA</str>
        <int name="rows">10</int>
      </lst>
     <!-- In addition to defaults, "appends" params can be specified
diff --git a/webapp/src/edu/cornell/mannlib/vitro/webapp/search/VitroTermNames.java b/webapp/src/edu/cornell/mannlib/vitro/webapp/search/VitroTermNames.java
index 35613c02d..2bf92a150 100644
--- a/webapp/src/edu/cornell/mannlib/vitro/webapp/search/VitroTermNames.java
+++ b/webapp/src/edu/cornell/mannlib/vitro/webapp/search/VitroTermNames.java
@@ -57,4 +57,8 @@ public class VitroTermNames {
     
     /** field for beta values of all documents **/
     public static final String BETA = "BETA";
+    
+    /** adding phonetic field **/
+    public static final String ALLTEXT_PHONETIC = "ALLTEXT_PHONETIC";
+    public static final String NAME_PHONETIC = "NAME_PHONETIC";
 }
diff --git a/webapp/src/edu/cornell/mannlib/vitro/webapp/search/beans/SearchQueryHandler.java b/webapp/src/edu/cornell/mannlib/vitro/webapp/search/beans/SearchQueryHandler.java
index ded894f8f..1bf79ed59 100644
--- a/webapp/src/edu/cornell/mannlib/vitro/webapp/search/beans/SearchQueryHandler.java
+++ b/webapp/src/edu/cornell/mannlib/vitro/webapp/search/beans/SearchQueryHandler.java
@@ -658,6 +658,7 @@ public class SearchQueryHandler {
 	        }
 	        
 	        beta = (float)Conn/totalInd;
+	        beta *= 100;
 	        beta += 1;
 		return beta; 
 	}
diff --git a/webapp/src/edu/cornell/mannlib/vitro/webapp/search/controller/SolrPagedSearchController.java b/webapp/src/edu/cornell/mannlib/vitro/webapp/search/controller/SolrPagedSearchController.java
index 4b7e4e44c..fd972d834 100644
--- a/webapp/src/edu/cornell/mannlib/vitro/webapp/search/controller/SolrPagedSearchController.java
+++ b/webapp/src/edu/cornell/mannlib/vitro/webapp/search/controller/SolrPagedSearchController.java
@@ -177,6 +177,7 @@ public class SolrPagedSearchController extends FreemarkerHttpServlet {
             SolrQuery query = getQuery(qtxt, maxHitCount, vreq);            
             SolrServer solr = SolrSetup.getSolrServer(getServletContext());
             QueryResponse response = null;
+           
             
             try {
                 response = solr.query(query);
diff --git a/webapp/src/edu/cornell/mannlib/vitro/webapp/search/solr/IndividualToSolrDocument.java b/webapp/src/edu/cornell/mannlib/vitro/webapp/search/solr/IndividualToSolrDocument.java
index d748c077a..031db994e 100644
--- a/webapp/src/edu/cornell/mannlib/vitro/webapp/search/solr/IndividualToSolrDocument.java
+++ b/webapp/src/edu/cornell/mannlib/vitro/webapp/search/solr/IndividualToSolrDocument.java
@@ -71,9 +71,9 @@ public class IndividualToSolrDocument implements Obj2DocIface {
     	String classPublicNames = "";
     	SolrInputDocument doc = new SolrInputDocument();
     	
-    	//float beta = searchQueryHandler.calculateBeta(ent.getURI());
-    	//doc.addField(term.BETA,beta);
-    	float beta =1;
+    	float beta = searchQueryHandler.calculateBeta(ent.getURI());
+    	doc.addField(term.BETA,beta);
+    	//float beta =1;
     	
     	//DocId
     	String id = ent.getURI();
@@ -156,10 +156,11 @@ public class IndividualToSolrDocument implements Obj2DocIface {
     		value = ent.getLocalName();
     	}
     	
-    	doc.addField(term.NAME_RAW, value, (NAME_BOOST*beta));
-    	doc.addField(term.NAME_LOWERCASE, value.toLowerCase(),(NAME_BOOST*beta));
-    	doc.addField(term.NAME_UNSTEMMED, value,(NAME_BOOST*beta));
-    	doc.addField(term.NAME_STEMMED, value, (NAME_BOOST*beta));
+    	doc.addField(term.NAME_RAW, value, NAME_BOOST+beta);
+    	doc.addField(term.NAME_LOWERCASE, value.toLowerCase(),NAME_BOOST+beta);
+    	doc.addField(term.NAME_UNSTEMMED, value,NAME_BOOST+beta);
+    	doc.addField(term.NAME_STEMMED, value, NAME_BOOST+beta);
+    	doc.addField(term.NAME_PHONETIC, value, PHONETIC_BOOST);
     	
     	long tContextNodes = System.currentTimeMillis();
     	
@@ -245,8 +246,10 @@ public class IndividualToSolrDocument implements Obj2DocIface {
             
         	log.debug("time to include data property statements, object property statements in the index: " + Long.toString(System.currentTimeMillis() - tPropertyStatements));
             
-            doc.addField(term.ALLTEXT, value,(ALL_TEXT_BOOST*beta));
-            doc.addField(term.ALLTEXTUNSTEMMED, value,(ALL_TEXT_BOOST*beta));
+            doc.addField(term.ALLTEXT, value, 4*beta);
+            doc.addField(term.ALLTEXTUNSTEMMED, value, 4*beta);
+            doc.addField(term.ALLTEXT_PHONETIC, value, PHONETIC_BOOST);
+            doc.setDocumentBoost(2*beta);
         }
         
         return doc;
@@ -289,7 +292,8 @@ public class IndividualToSolrDocument implements Obj2DocIface {
         return ent;
     }
 
-    public static float NAME_BOOST = 3.0F;
-    public static float ALL_TEXT_BOOST = 2.0F;
+    public static float NAME_BOOST = 2.0F;
+    public static float PHONETIC_BOOST = 0.1F;
+    
     
 }
diff --git a/webapp/src/edu/cornell/mannlib/vitro/webapp/search/solr/SolrIndexer.java b/webapp/src/edu/cornell/mannlib/vitro/webapp/search/solr/SolrIndexer.java
index d7159e8b1..f720a5652 100644
--- a/webapp/src/edu/cornell/mannlib/vitro/webapp/search/solr/SolrIndexer.java
+++ b/webapp/src/edu/cornell/mannlib/vitro/webapp/search/solr/SolrIndexer.java
@@ -36,7 +36,6 @@ public class SolrIndexer implements IndexerIface {
     
     @Override
     public synchronized void index(Individual ind, boolean newDoc) throws IndexingException {
-
         if( ! indexing )
             throw new IndexingException("SolrIndexer: must call " +
                     "startIndexing() before index().");
@@ -61,6 +60,7 @@ public class SolrIndexer implements IndexerIface {
                             Collection<SolrInputDocument> docs = new ArrayList<SolrInputDocument>();
                             docs.add( solrDoc );
                             server.add( docs );
+                   
 //                            if( !newDoc ){  
 //                                server.add( docs );
 //                                log.debug("updated " + ind.getName() + " " + ind.getURI());