Merge pull request #349 from sul-dlss/sync-to-sw-prod

Sync gryphon-search relevancy schema changes to -prod.
sul-dlss · Sep 19, 2024 · ad83a3d · ad83a3d
2 parents 4934a84 + fe6eeca
commit ad83a3d
Showing 1 changed file with 24 additions and 0 deletions.
diff --git a/searchworks-prod-20230907/schema.xml b/searchworks-prod-20230907/schema.xml
@@ -44,9 +44,11 @@
 
     <!-- Title Search Fields -->
     <field name="title_245a_exact_search" type="text_anchored" indexed="true" stored="false" />
+    <field name="title_245a_ws_search" type="text_ws" indexed="true" stored="false" />
     <field name="title_245a_search" type="text" indexed="true" stored="true" />
     <field name="vern_title_245a_search" type="textNoStem" indexed="true" stored="true" />
     <field name="title_245a_unstem_search" type="textNoStem" indexed="true" stored="false" />
+    <field name="title_245_ws_search" type="text_ws" indexed="true" stored="false" />
     <field name="title_245_search" type="text" indexed="true" stored="true" />
     <field name="vern_title_245_search" type="textNoStem" indexed="true" stored="true" />
     <field name="title_245_unstem_search" type="textNoStem" indexed="true" stored="false" />
@@ -89,6 +91,7 @@
     <field name="author_title_search" type="text" indexed="true" stored="true" multiValued="true" />
     <field name="author_title_unstem_search" type="textNoStem" indexed="true" stored="false" multiValued="true" />
     <field name="best_author_title_search" type="text" indexed="true" stored="true" multiValued="false" />
+    <field name="best_author_title_unstem_search" type="textNoStem" indexed="true" stored="false" multiValued="false" />
 
     <!-- Author Search Fields -->
     <field name="author_1xx_search" type="text" indexed="true" stored="true" />
@@ -300,14 +303,17 @@
   <copyField source="barcode_search" dest="item_barcodes" />
   <!-- unstemmed and anchored search fields: title -->
   <copyField source="title_245a_search" dest="title_245a_exact_search" />
+  <copyField source="title_245a_search" dest="title_245a_ws_search" />
   <copyField source="title_245a_search" dest="title_245a_unstem_search" />
   <copyField source="title_245_search" dest="title_245_unstem_search" />
+  <copyField source="title_245_search" dest="title_245_ws_search" />
   <copyField source="title_uniform_search" dest="title_uniform_unstem_search" />
   <copyField source="title_variant_search" dest="title_variant_unstem_search" />
   <copyField source="title_related_search" dest="title_related_unstem_search" />
   <copyField source="author_title_search" dest="author_title_unstem_search" />
   <copyField source="author_title_245ac_search" dest="author_title_245ac_unstem_search" />
   <copyField source="author_title_1xx_search" dest="author_title_1xx_unstem_search" />
+  <copyField source="best_author_title_search" dest="best_author_title_unstem_search" />
   <!-- unstemmed search fields: author -->
   <copyField source="author_1xx_search" dest="author_1xx_unstem_search" />
   <copyField source="author_7xx_search" dest="author_7xx_unstem_search" />
@@ -495,6 +501,24 @@
       </analyzer>
     </fieldtype>
 
+    <!-- Analyzed Text, exact string matching -->
+    <fieldtype name="text_ws" class="solr.TextField" positionIncrementGap="10000" autoGeneratePhraseQueries="true">
+      <analyzer type="index">
+        <charFilter class="solr.PatternReplaceCharFilterFactory" pattern="\s+(\p{Punct}+)\s+" replacement=" " />
+        <charFilter class="solr.PatternReplaceCharFilterFactory" pattern="\s+([&quot;\[{(])" replacement=" " />
+        <charFilter class="solr.PatternReplaceCharFilterFactory" pattern="([;:.,&quot;\]})]+)\s+" replacement=" " />
+        <tokenizer class="solr.WhitespaceTokenizerFactory" />
+        <filter class="solr.ICUNormalizer2FilterFactory" />
+      </analyzer>
+      <analyzer type="query">
+        <charFilter class="solr.PatternReplaceCharFilterFactory" pattern="\s+(\p{Punct}+)\s+" replacement=" " />
+        <charFilter class="solr.PatternReplaceCharFilterFactory" pattern="\s+(&quot;)" replacement=" " />
+        <charFilter class="solr.PatternReplaceCharFilterFactory" pattern="([;:.,&quot;]+}+)\s+" replacement=" " />
+        <tokenizer class="solr.WhitespaceTokenizerFactory" />
+        <filter class="solr.ICUNormalizer2FilterFactory" />
+      </analyzer>
+    </fieldtype>
+
     <!-- Left and Right Anchored Analyzed Text, no Stemming -->
     <fieldtype name="text_anchored" class="solr.TextField" positionIncrementGap="10000" autoGeneratePhraseQueries="true">
       <analyzer type="index">