Reorganize configuration directory.
[lui-solr.git] / conf / solr / solr-home / lui / conf / schema.xml
diff --git a/conf/solr/solr-home/lui/conf/schema.xml b/conf/solr/solr-home/lui/conf/schema.xml
new file mode 100644 (file)
index 0000000..9253610
--- /dev/null
@@ -0,0 +1,271 @@
+<?xml version="1.0" encoding="UTF-8" ?>
+<!--  
+ This is the Solr schema file for the Index Data Local Unified
+ Index. This file should be named "schema.xml" and should be in the
+ conf directory for the solr core
+ (i.e. /etc/masterkey/lui/solr-config/schema.xml by default).
+ For customization, see https://wiki.apache.org/solr/SchemaXml and the
+ Schema Design section of the Solr Reference Guide.
+-->
+
+<schema name="Local Unified Index" version="1.6">
+
+    <!-- stock Solr field types from the dist schema.xml -->
+    <fieldType name="string" class="solr.StrField" sortMissingLast="true" omitNorms="true"/>
+    <fieldType name="boolean" class="solr.BoolField" sortMissingLast="true" omitNorms="true"/>
+    <fieldtype name="binary" class="solr.BinaryField"/>
+    <fieldType name="int" class="solr.TrieIntField" precisionStep="0" omitNorms="true" positionIncrementGap="0"/>
+    <fieldType name="float" class="solr.TrieFloatField" precisionStep="0" omitNorms="true" positionIncrementGap="0"/>
+    <fieldType name="long" class="solr.TrieLongField" precisionStep="0" omitNorms="true" positionIncrementGap="0"/>
+    <fieldType name="double" class="solr.TrieDoubleField" precisionStep="0" omitNorms="true" positionIncrementGap="0"/>
+    <fieldType name="tint" class="solr.TrieIntField" precisionStep="8" omitNorms="true" positionIncrementGap="0"/>
+    <fieldType name="tfloat" class="solr.TrieFloatField" precisionStep="8" omitNorms="true" positionIncrementGap="0"/>
+    <fieldType name="tlong" class="solr.TrieLongField" precisionStep="8" omitNorms="true" positionIncrementGap="0"/>
+    <fieldType name="tdouble" class="solr.TrieDoubleField" precisionStep="8" omitNorms="true" positionIncrementGap="0"/>
+    <fieldType name="date" class="solr.TrieDateField" omitNorms="true" precisionStep="0" positionIncrementGap="0"/>
+    <fieldType name="tdate" class="solr.TrieDateField" omitNorms="true" precisionStep="6" positionIncrementGap="0"/>
+
+    <fieldType name="facet" class="solr.TextField" sortMissingLast="true" omitNorms="true">
+      <analyzer>
+        <tokenizer class="solr.KeywordTokenizerFactory"/>
+        <filter class="solr.LowerCaseFilterFactory" />
+        <filter class="solr.TrimFilterFactory" />
+      </analyzer>
+    </fieldType>
+
+    <fieldType name="random" class="solr.RandomSortField" indexed="true" />
+
+    <fieldType name="text_ws" class="solr.TextField" positionIncrementGap="100">
+      <analyzer>
+        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
+      </analyzer>
+    </fieldType>
+
+    <fieldType name="text" class="solr.TextField" positionIncrementGap="100">
+      <analyzer type="index">
+        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
+        <filter class="solr.StopFilterFactory"
+                ignoreCase="true"
+                words="stopwords.txt"
+                />
+        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" catenateWords="1" catenateNumbers="1" catenateAll="0" splitOnCaseChange="1"/>
+        <filter class="solr.LowerCaseFilterFactory"/>
+        <filter class="solr.SnowballPorterFilterFactory" language="English" protected="protwords.txt"/>
+      </analyzer>
+      <analyzer type="query">
+        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
+        <filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
+        <filter class="solr.StopFilterFactory"
+                ignoreCase="true"
+                words="stopwords.txt"
+                />
+        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" catenateWords="0" catenateNumbers="0" catenateAll="0" splitOnCaseChange="1"/>
+        <filter class="solr.LowerCaseFilterFactory"/>
+        <filter class="solr.SnowballPorterFilterFactory" language="English" protected="protwords.txt"/>
+      </analyzer>
+    </fieldType>
+
+    <fieldType name="textTight" class="solr.TextField" positionIncrementGap="100" >
+      <analyzer>
+        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
+        <filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="false"/>
+        <filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt"/>
+        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="0" generateNumberParts="0" catenateWords="1" catenateNumbers="1" catenateAll="0"/>
+        <filter class="solr.LowerCaseFilterFactory"/>
+        <filter class="solr.SnowballPorterFilterFactory" language="English" protected="protwords.txt"/>
+        <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>
+      </analyzer>
+    </fieldType>
+
+    <fieldType name="textgen" class="solr.TextField" positionIncrementGap="100">
+      <analyzer type="index">
+        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
+        <filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" />
+        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" catenateWords="1" catenateNumbers="1" catenateAll="0" splitOnCaseChange="0"/>
+        <filter class="solr.LowerCaseFilterFactory"/>
+      </analyzer>
+      <analyzer type="query">
+        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
+        <filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
+        <filter class="solr.StopFilterFactory"
+                ignoreCase="true"
+                words="stopwords.txt"
+                />
+        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" catenateWords="0" catenateNumbers="0" catenateAll="0" splitOnCaseChange="0"/>
+        <filter class="solr.LowerCaseFilterFactory"/>
+      </analyzer>
+    </fieldType>
+
+    <fieldType name="text_rev" class="solr.TextField" positionIncrementGap="100">
+      <analyzer type="index">
+        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
+        <filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" />
+        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" catenateWords="1" catenateNumbers="1" catenateAll="0" splitOnCaseChange="0"/>
+        <filter class="solr.LowerCaseFilterFactory"/>
+        <filter class="solr.ReversedWildcardFilterFactory" withOriginal="true"
+           maxPosAsterisk="3" maxPosQuestion="2" maxFractionAsterisk="0.33"/>
+      </analyzer>
+      <analyzer type="query">
+        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
+        <filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
+        <filter class="solr.StopFilterFactory"
+                ignoreCase="true"
+                words="stopwords.txt"
+                />
+        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" catenateWords="0" catenateNumbers="0" catenateAll="0" splitOnCaseChange="0"/>
+        <filter class="solr.LowerCaseFilterFactory"/>
+      </analyzer>
+    </fieldType>
+
+    <fieldType name="alphaOnlySort" class="solr.TextField" sortMissingLast="true" omitNorms="true">
+      <analyzer>
+        <tokenizer class="solr.KeywordTokenizerFactory"/>
+        <filter class="solr.LowerCaseFilterFactory" />
+        <filter class="solr.TrimFilterFactory" />
+        <filter class="solr.PatternReplaceFilterFactory"
+                pattern="([^a-z])" replacement="" replace="all"
+        />
+      </analyzer>
+    </fieldType>
+    
+    <fieldtype name="phonetic" stored="false" indexed="true" class="solr.TextField" >
+      <analyzer>
+        <tokenizer class="solr.StandardTokenizerFactory"/>
+        <filter class="solr.DoubleMetaphoneFilterFactory" inject="false"/>
+      </analyzer>
+    </fieldtype>
+
+    <fieldtype name="payloads" stored="false" indexed="true" class="solr.TextField" >
+      <analyzer>
+        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
+        <filter class="solr.DelimitedPayloadTokenFilterFactory" encoder="float"/>
+      </analyzer>
+    </fieldtype>
+
+    <fieldType name="lowercase" class="solr.TextField" positionIncrementGap="100">
+      <analyzer>
+        <tokenizer class="solr.KeywordTokenizerFactory"/>
+        <filter class="solr.LowerCaseFilterFactory" />
+      </analyzer>
+    </fieldType>
+
+    <fieldtype name="ignored" stored="false" indexed="false" multiValued="true" class="solr.StrField" /> 
+
+   <!-- defined fields in the lui-solr schema -->
+
+   <field name="id" type="string" indexed="true" stored="true" required="true" /> 
+   <field name="_version_" type="long" indexed="true" stored="true"/>
+
+   <field name="transactionId" type="long" indexed="true" stored="false"/>
+
+   <field name="author"       type="text"   indexed="true" stored="true"  multiValued="true" omitNorms="true"/>
+   <field name="author_exact" type="string" indexed="true" stored="false" multiValued="true" omitNorms="true" docValues="true" />
+   <field name="author-date"  type="text"   indexed="true" stored="true"  multiValued="true" omitNorms="true"/>
+   <field name="author-title" type="text"   indexed="true" stored="true"  multiValued="true" omitNorms="true"/>
+
+   <field name="corporate-date"     type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+   <field name="corporate-location" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+   <field name="corporate-name"     type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+
+   <field name="callnumber" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+   <field name="citation"   type="text" indexed="true" stored="true" multiValued="true"  omitNorms="true"/>
+
+   <field name="date"        type="text"    indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+   <field name="description" type="text_ws" indexed="true" stored="true" multiValued="true"  omitNorms="true"/>
+
+   <field name="edition" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+
+   <field name="electronic-format-instruction" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+   <field name="electronic-format-type"        type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+   <field name="electronic-note"               type="text" indexed="true" stored="true" multiValued="true"  omitNorms="true"/>
+   <field name="electronic-text"               type="text" indexed="true" stored="true" multiValued="true"  omitNorms="true"/>
+   <field name="electronic-url"                type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+
+   <field name="isbn" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+   <field name="issn" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+   <field name="lccn" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+
+   <field name="medium"       type="text"  indexed="true" stored="true"   multiValued="true" omitNorms="true"/>
+   <field name="medium_exact" type="string" indexed="true" stored="false" multiValued="true" omitNorms="true" docValues="true" />
+
+   <field name="meeting-date"     type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+   <field name="meeting-location" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+   <field name="meeting-name"     type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+
+   <field name="series-title" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+
+   <field name="subject"       type="text_ws" indexed="true" stored="true"  multiValued="true" omitNorms="true" />
+   <field name="subject_exact" type="string"  indexed="true" stored="false" multiValued="true" omitNorms="true" docValues="true" />
+   <field name="subject-long"  type="text_ws" indexed="true" stored="true"  multiValued="true" omitNorms="true" />
+
+   <field name="system-control-nr" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
+
+   <field name="tech-rep-nr" type="text" indexed="true" multiValued="true" stored="true"/>
+
+   <field name="title"                  type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-complete"         type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-dates"            type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-medium"           type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-number-section"   type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-remainder"        type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-responsibility"   type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-uniform"          type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-uniform-key"      type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-uniform-media"    type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-uniform-partname" type="text"   indexed="true" multiValued="true" stored="true"/>
+   <field name="title-uniform-parts"    type="text"   indexed="true" multiValued="true" stored="true"/>
+
+   <field name="journal-title"          type="text" indexed="true" multiValued="true"  stored="true" />
+   <field name="journal-title_exact"    type="string" indexed="true" multiValued="true"  stored="false" docValues="true" />
+
+   <field name="physical-accomp"     type="text" indexed="true" stored="true" multiValued="true" />
+   <field name="physical-dimensions" type="text" indexed="true" stored="true" multiValued="true" />
+   <field name="physical-extent"     type="text" indexed="true" stored="true" multiValued="true" />
+   <field name="physical-format"     type="text" indexed="true" stored="true" multiValued="true" />
+   <field name="physical-specified"  type="text" indexed="true" stored="true" multiValued="true" />
+   <field name="physical-unitsize"   type="text" indexed="true" stored="true" multiValued="true" />
+   <field name="physical-unittype"   type="text" indexed="true" stored="true" multiValued="true" />
+
+   <field name="publication-date"  type="text" indexed="true" stored="true" multiValued="true" />
+   <field name="publication-name"  type="text" indexed="true" stored="true" multiValued="true" />
+   <field name="publication-place" type="text" indexed="true" stored="true" multiValued="true" />
+
+   <field name="harvest-timestamp" type="date"   indexed="true" stored="true"/>
+   <field name="harvest-date"      type="string" indexed="true" stored="true"/>
+
+   <!-- default search field, created by the clone-fields update
+        processor chain in solrconfig.xml -->
+   <field name="text" type="text" indexed="true" stored="false" multiValued="true"/>
+
+   <!-- Dynamic field definitions -->
+   <dynamicField name="*_i"  type="int"    indexed="true"  stored="true"/>
+   <dynamicField name="*_s"  type="string"  indexed="true"  stored="true"/>
+   <dynamicField name="*_l"  type="long"   indexed="true"  stored="true"/>
+   <dynamicField name="*_t"  type="text"    indexed="true"  stored="true"/>
+   <dynamicField name="*_b"  type="boolean" indexed="true"  stored="true"/>
+   <dynamicField name="*_f"  type="float"  indexed="true"  stored="true"/>
+   <dynamicField name="*_d"  type="double" indexed="true"  stored="true"/>
+   <dynamicField name="*_dt" type="date"    indexed="true"  stored="true"/>
+   <dynamicField name="*_ti" type="tint"    indexed="true"  stored="true"/>
+   <dynamicField name="*_tl" type="tlong"   indexed="true"  stored="true"/>
+   <dynamicField name="*_tf" type="tfloat"  indexed="true"  stored="true"/>
+   <dynamicField name="*_td" type="tdouble" indexed="true"  stored="true"/>
+   <dynamicField name="*_tdt" type="tdate"  indexed="true"  stored="true"/>
+   <dynamicField name="ignored_*" type="ignored" multiValued="true"/>
+   <dynamicField name="attr_*" type="textgen" indexed="true" stored="true" multiValued="true"/>
+   <dynamicField name="random_*" type="random" />
+   <dynamicField name="*_binary" type="binary" indexed="false" stored="true" multiValued="true"/>
+   <dynamicField name="*_path" type="text" indexed="true" stored="true" multiValued="true"/>
+
+   <!-- catchall dynamic field -->
+   <dynamicField name="*" type="text" multiValued="true"/>
+
+ <uniqueKey>id</uniqueKey>
+
+ <!-- copyField commands -->
+ <copyField source="author"  dest="author_exact"/>
+ <copyField source="subject" dest="subject_exact"/>
+ <copyField source="medium"  dest="medium_exact"/>
+ <copyField source="journal-title"   dest="journal-title_exact"/>
+
+</schema>