Reviewving schema

romanchyla · Nov 25, 2014 · 25f308f · 25f308f
1 parent dc32c6a
commit 25f308f
Show file tree

Hide file tree

Showing 6 changed files with 72 additions and 160 deletions.
diff --git a/.classpath b/.classpath
@@ -1,7 +1,7 @@
 <?xml version="1.0" encoding="UTF-8"?>
 <classpath>
 	<classpathentry kind="lib" path="build/solrjars-extracted/lucene-test-framework-4.8-SNAPSHOT.jar" sourcepath="build/solr-download/apache-solr-4.0.0-SVN/lucene/test-framework/src/java"/>
-	<classpathentry kind="lib" path="build/solrjars-extracted/solr-test-framework-4.8-SNAPSHOT.jar" sourcepath="build/solr-download/apache-solr-4.0.0-SVN/lucene/core/src"/>
+	<classpathentry kind="lib" path="build/solrjars-extracted/solr-test-framework-4.8-SNAPSHOT.jar" sourcepath="/solr-next"/>
 	<classpathentry kind="src" path="contrib/antlrqueryparser/src/java"/>
 	<classpathentry kind="src" path="contrib/adsabs/src/jython"/>
 	<classpathentry kind="src" path="contrib/antlrqueryparser/src/test"/>

diff --git a/.project b/.project
@@ -3,6 +3,7 @@
 	<name>montysolr-next</name>
 	<comment></comment>
 	<projects>
+		<project>solr-next</project>
 	</projects>
 	<buildSpec>
 		<buildCommand>

diff --git a/build.xml b/build.xml
@@ -130,12 +130,14 @@
 
 
 
-	<target name="test" depends="write-properties">
-		   <antcall target="common.test" inheritall="true" inheritrefs="true"/>
-	</target>
+	<target name="test" depends="write-properties" description="Run tests for core and contrib">
+    <antcall target="common.test" inheritall="true" inheritrefs="true"/>
+    <antcall target="test-contrib" inheritall="true" inheritrefs="true"/>
+  </target>
 
 
-	<target name="test-all" depends="test,test-contrib" description="Run tests for core and contrib"/>
+  <!-- obsolete: redirecting to test -->
+  <target name="test-all" depends="test" />
 
 
 	<!-- writes into a simple format various config variables, they can be used by unittests/java -->

diff --git a/contrib/adsabs/src/test/org/adsabs/TestAdsAllFields.java b/contrib/adsabs/src/test/org/adsabs/TestAdsAllFields.java
@@ -195,7 +195,7 @@ public void test() throws Exception {
 				", \"alternate_title\": \"This is of the alternate\"" +
 				", \"abstract\": \"all no-sky survey q'i quotient\"" +
 
-				", \"keyword\": [\"Classical statistical mechanics\"]" +
+				", \"keyword\": [\"Classical statistical mechanics\", \"foo bar\"]" +
 				", \"keyword_norm\": [\"angular momentum\", \"89.20.Hh\"]" +
 				", \"keyword_schema\": [\"ADS\", \"PACS Codes\"]" +
 				", \"keyword_facet\": [\"angular momentum kw\"]" +
@@ -504,6 +504,10 @@ public void test() throws Exception {
 				"//*[@numFound='1']",
 				"//doc/int[@name='recid'][.='100']"
 		);
+		assertQ(req("q", "keyword:\"foo bar\""), 
+				"//*[@numFound='1']",
+				"//doc/int[@name='recid'][.='100']"
+		);
 		assertQ(req("q", "keyword:\"Classical Statistical Mechanics\""), // should be case-insensitive 
 				"//*[@numFound='1']",
 				"//doc/int[@name='recid'][.='100']"
@@ -520,7 +524,7 @@ public void test() throws Exception {
 		);
 
 		assertQ(req("q", "keyword_norm:classical"), "//*[@numFound='0']"); // should not contain keywords
-		assertQ(req("q", "keyword:89.20.Hh"), "//*[@numFound='1']"); // should contain keywords_norm
+		assertQ(req("q", "keyword:89.20.Hh"), "//*[@numFound='0']"); // should not contain keywords_norm
 
 
 		/*

diff --git a/contrib/examples/adsabs/solr/collection1/conf/schema.xml b/contrib/examples/adsabs/solr/collection1/conf/schema.xml
@@ -451,10 +451,12 @@
 
 				<!-- remove stop words - first the case sensitively -->
 				<filter class="solr.StopFilterFactory" ignoreCase="false"
-					words="text.kill_sens" enablePositionIncrements="false" />
+					words="text.kill_sens" enablePositionIncrements="false" 
+					luceneMatchVersion="4.3"/>
 				<!-- remove stop words - then case insensitively -->
 				<filter class="solr.StopFilterFactory" ignoreCase="true"
-					words="text.kill" enablePositionIncrements="false" />
+					words="text.kill" enablePositionIncrements="false" 
+					luceneMatchVersion="4.3"/>
 
 
 
@@ -537,10 +539,12 @@
 
 				<!-- remove stop words - first the case sensitively -->
 				<filter class="solr.StopFilterFactory" ignoreCase="false"
-					words="text.kill_sens" enablePositionIncrements="false" />
+					words="text.kill_sens" enablePositionIncrements="false" 
+					luceneMatchVersion="4.3"/>
 				<!-- remove stop words - then case insensitively -->
 				<filter class="solr.StopFilterFactory" ignoreCase="true"
-					words="text.kill" enablePositionIncrements="false" />
+					words="text.kill" enablePositionIncrements="false" 
+					luceneMatchVersion="4.3"/>
 
 
 
@@ -609,10 +613,12 @@
 
         <!-- remove stop words - first the case sensitively -->
         <filter class="solr.StopFilterFactory" ignoreCase="false"
-          words="text.kill_sens" enablePositionIncrements="false" />
+          words="text.kill_sens" enablePositionIncrements="false" 
+          luceneMatchVersion="4.3"/>
         <!-- remove stop words - then case insensitively -->
         <filter class="solr.StopFilterFactory" ignoreCase="true"
-          words="text.kill" enablePositionIncrements="false" />
+          words="text.kill" enablePositionIncrements="false" 
+          luceneMatchVersion="4.3"/>
 
 
 
@@ -679,6 +685,8 @@
 				<filter class="solr.PatternReplaceFilterFactory" pattern="(doi:|arxiv:)"
 					replacement="" replace="all" />
 				<filter class="solr.TrimFilterFactory" />
+				<filter
+          class="org.apache.lucene.analysis.miscellaneous.RemoveDuplicatesTokenFilterFactory" />
 			</analyzer>
 			<analyzer type="query">
 				<tokenizer class="solr.KeywordTokenizerFactory" />
@@ -773,6 +781,8 @@
 				<filter class="solr.PatternReplaceFilterFactory" pattern="(\-|_| )"
           replacement="" replace="all" />
 				<filter class="solr.TrimFilterFactory" />
+				<filter
+          class="org.apache.lucene.analysis.miscellaneous.RemoveDuplicatesTokenFilterFactory" />
 			</analyzer>
 			<analyzer type="query">
 				<tokenizer class="solr.KeywordTokenizerFactory" />
@@ -1050,6 +1060,8 @@
 
 		<field name="page" type="first_string" indexed="true" stored="true"
 			multiValued="true" omitNorms="true" omitTermFreqAndPositions="true"/>
+	  <field name="eid" type="first_string" indexed="true" stored="true"
+      multiValued="true" omitNorms="true" omitTermFreqAndPositions="true"/>
 
 		<field name="year" type="first_string" indexed="true" stored="true" 
 			multiValued="false" omitNorms="true" omitTermFreqAndPositions="true"/>
@@ -1079,9 +1091,6 @@
 		<field name="pubdate" type="date_string" indexed="true" stored="true"
 			omitNorms="true" omitTermFreqAndPositions="true"/>
 
-		<field name="pubdate_sort" type="int" indexed="true" stored="${storeAll:false}"
-			omitNorms="true" omitTermFreqAndPositions="true"/>
-
 
 		<field name="title" type="ads_text" indexed="true" stored="true" 
       multiValued="true" termVectors="true" termOffsets="true" 
@@ -1460,33 +1469,7 @@
 	<solrQueryParser defaultOperator="AND" />
 
 	<copyField source="id" dest="recid" />
-
-  <copyField source="arxiv_class" dest="keyword" />
-	<copyField source="keyword_norm" dest="keyword" />
 
-	<copyField source="bibgroup" dest="bibgroup_facet" />
-	<copyField source="data" dest="data_facet" />
-	<copyField source="vizier" dest="vizier_facet" />
-
-	<copyField source="a_100" dest="author" />
-	<copyField source="a_700" dest="author" />
-
-	<copyField source="a_100_norm" dest="author_norm" />
-	<copyField source="a_700_norm" dest="author_norm" />
-
-	<copyField source="a_100_norm" dest="author_facet" />
-	<copyField source="a_700_norm" dest="author_facet" />
-
-	<copyField source="a_100" dest="first_author" />
-	<copyField source="a_100_norm" dest="first_author_norm" />
-
-  <copyField source="tmp_aff_100" dest="aff" />
-  <copyField source="tmp_aff_700" dest="aff" />
-
-  <copyField source="tmp_email_100" dest="email" />
-  <copyField source="tmp_email_700" dest="email" />
-
-
   <copyField source="bibcode" dest="identifier" />
   <copyField source="alternate_bibcode" dest="identifier" />
   <copyField source="doi" dest="identifier" />