WP6
/
patents_search_php
Watch 1
Star 0
Fork 0
Code Issues 0 Pull Requests 0 Releases 0 Wiki Activity
A search interface for data from the Politics of Patents case study (part of Copim WP6): this parses data from the archive of RTF files and provides additional data from the European Patent Office API. https://patents.copim.ac.uk
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
77 Commits
1 Branch
Branch: main
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'main'
${ noResults }
patents_search_php/solr_config/schema.xml

531 lines
30KB
Raw Permalink Blame History 

  1. <?xml version="1.0" encoding="UTF-8"?>

  2. <!-- Solr managed schema - automatically generated - DO NOT EDIT -->

  3. <schema name="example-data-driven-schema" version="1.6">

  4.   <uniqueKey>id</uniqueKey>

  5.   <fieldType name="ancestor_path" class="solr.TextField">

  6.     <analyzer type="index">

  7.       <tokenizer class="solr.KeywordTokenizerFactory"/>

  8.     </analyzer>

  9.     <analyzer type="query">

  10.       <tokenizer class="solr.PathHierarchyTokenizerFactory" delimiter="/"/>

  11.     </analyzer>

  12.   </fieldType>

  13.   <fieldType name="binary" class="solr.BinaryField"/>

  14.   <fieldType name="boolean" class="solr.BoolField" sortMissingLast="true"/>

  15.   <fieldType name="booleans" class="solr.BoolField" sortMissingLast="true" multiValued="true"/>

  16.   <fieldType name="currency" class="solr.CurrencyFieldType" amountLongSuffix="_l_ns" codeStrSuffix="_s_ns" defaultCurrency="USD" currencyConfig="currency.xml" />

  17.   <fieldType name="descendent_path" class="solr.TextField">

  18.     <analyzer type="index">

  19.       <tokenizer class="solr.PathHierarchyTokenizerFactory" delimiter="/"/>

  20.     </analyzer>

  21.     <analyzer type="query">

  22.       <tokenizer class="solr.KeywordTokenizerFactory"/>

  23.     </analyzer>

  24.   </fieldType>

  25.   <fieldType name="ignored" class="solr.StrField" indexed="false" stored="false" multiValued="true"/>

  26.   <fieldType name="location" class="solr.LatLonType" subFieldSuffix="_coordinate"/>

  27.   <fieldType name="location_rpt" class="solr.SpatialRecursivePrefixTreeFieldType" geo="true" maxDistErr="0.001" distErrPct="0.025" distanceUnits="kilometers"/>

  28.   <fieldType name="lowercase" class="solr.TextField" positionIncrementGap="100">

  29.     <analyzer>

  30.       <tokenizer class="solr.KeywordTokenizerFactory"/>

  31.       <filter class="solr.LowerCaseFilterFactory"/>

  32.     </analyzer>

  33.   </fieldType>

  34.   <fieldType name="phonetic_en" class="solr.TextField" indexed="true" stored="false">

  35.     <analyzer>

  36.       <tokenizer class="solr.StandardTokenizerFactory"/>

  37.       <filter class="solr.DoubleMetaphoneFilterFactory" inject="false"/>

  38.     </analyzer>

  39.   </fieldType>

  40.   <fieldType name="pdate" class="solr.DatePointField" docValues="true"/>

  41.   <fieldType name="pdates" class="solr.DatePointField" docValues="true" multiValued="true"/>

  42.   <fieldType name="pdouble" class="solr.DoublePointField" docValues="true"/>

  43.   <fieldType name="pdoubles" class="solr.DoublePointField" docValues="true" multiValued="true"/>

  44.   <fieldType name="pfloat" class="solr.FloatPointField" docValues="true"/>

  45.   <fieldType name="pfloats" class="solr.FloatPointField" docValues="true" multiValued="true"/>

  46.   <fieldType name="pint" class="solr.IntPointField" docValues="true"/>

  47.   <fieldType name="pints" class="solr.IntPointField" docValues="true" multiValued="true"/>

  48.   <fieldType name="plong" class="solr.LongPointField" docValues="true"/>

  49.   <fieldType name="plongs" class="solr.LongPointField" docValues="true" multiValued="true"/>

  50.   <fieldType name="point" class="solr.PointType" subFieldSuffix="_d" dimension="2"/>

  51.   <fieldType name="random" class="solr.RandomSortField" indexed="true"/>

  52.   <fieldType name="string" class="solr.StrField" sortMissingLast="true"/>

  53.   <fieldType name="strings" class="solr.StrField" sortMissingLast="true" multiValued="true"/>

  54.   <fieldType name="text_ar" class="solr.TextField" positionIncrementGap="100">

  55.     <analyzer>

  56.       <tokenizer class="solr.StandardTokenizerFactory"/>

  57.       <filter class="solr.LowerCaseFilterFactory"/>

  58.       <filter class="solr.StopFilterFactory" words="lang/stopwords_ar.txt" ignoreCase="true"/>

  59.       <filter class="solr.ArabicNormalizationFilterFactory"/>

  60.       <filter class="solr.ArabicStemFilterFactory"/>

  61.     </analyzer>

  62.   </fieldType>

  63.   <fieldType name="text_bg" class="solr.TextField" positionIncrementGap="100">

  64.     <analyzer>

  65.       <tokenizer class="solr.StandardTokenizerFactory"/>

  66.       <filter class="solr.LowerCaseFilterFactory"/>

  67.       <filter class="solr.StopFilterFactory" words="lang/stopwords_bg.txt" ignoreCase="true"/>

  68.       <filter class="solr.BulgarianStemFilterFactory"/>

  69.     </analyzer>

  70.   </fieldType>

  71.   <fieldType name="text_ca" class="solr.TextField" positionIncrementGap="100">

  72.     <analyzer>

  73.       <tokenizer class="solr.StandardTokenizerFactory"/>

  74.       <filter class="solr.ElisionFilterFactory" articles="lang/contractions_ca.txt" ignoreCase="true"/>

  75.       <filter class="solr.LowerCaseFilterFactory"/>

  76.       <filter class="solr.StopFilterFactory" words="lang/stopwords_ca.txt" ignoreCase="true"/>

  77.       <filter class="solr.SnowballPorterFilterFactory" language="Catalan"/>

  78.     </analyzer>

  79.   </fieldType>

  80.   <fieldType name="text_cjk" class="solr.TextField" positionIncrementGap="100">

  81.     <analyzer>

  82.       <tokenizer class="solr.StandardTokenizerFactory"/>

  83.       <filter class="solr.CJKWidthFilterFactory"/>

  84.       <filter class="solr.LowerCaseFilterFactory"/>

  85.       <filter class="solr.CJKBigramFilterFactory"/>

  86.     </analyzer>

  87.   </fieldType>

  88.   <fieldType name="text_cz" class="solr.TextField" positionIncrementGap="100">

  89.     <analyzer>

  90.       <tokenizer class="solr.StandardTokenizerFactory"/>

  91.       <filter class="solr.LowerCaseFilterFactory"/>

  92.       <filter class="solr.StopFilterFactory" words="lang/stopwords_cz.txt" ignoreCase="true"/>

  93.       <filter class="solr.CzechStemFilterFactory"/>

  94.     </analyzer>

  95.   </fieldType>

  96.   <fieldType name="text_da" class="solr.TextField" positionIncrementGap="100">

  97.     <analyzer>

  98.       <tokenizer class="solr.StandardTokenizerFactory"/>

  99.       <filter class="solr.LowerCaseFilterFactory"/>

  100.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_da.txt" ignoreCase="true"/>

  101.       <filter class="solr.SnowballPorterFilterFactory" language="Danish"/>

  102.     </analyzer>

  103.   </fieldType>

  104.   <fieldType name="text_de" class="solr.TextField" positionIncrementGap="100">

  105.     <analyzer>

  106.       <tokenizer class="solr.StandardTokenizerFactory"/>

  107.       <filter class="solr.LowerCaseFilterFactory"/>

  108.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_de.txt" ignoreCase="true"/>

  109.       <filter class="solr.GermanNormalizationFilterFactory"/>

  110.       <filter class="solr.GermanLightStemFilterFactory"/>

  111.     </analyzer>

  112.   </fieldType>

  113.   <fieldType name="text_el" class="solr.TextField" positionIncrementGap="100">

  114.     <analyzer>

  115.       <tokenizer class="solr.StandardTokenizerFactory"/>

  116.       <filter class="solr.GreekLowerCaseFilterFactory"/>

  117.       <filter class="solr.StopFilterFactory" words="lang/stopwords_el.txt" ignoreCase="false"/>

  118.       <filter class="solr.GreekStemFilterFactory"/>

  119.     </analyzer>

  120.   </fieldType>

  121.   <fieldType name="text_en" class="solr.TextField" positionIncrementGap="100">

  122.     <analyzer type="index">

  123.       <tokenizer class="solr.StandardTokenizerFactory"/>

  124.       <filter class="solr.StopFilterFactory" words="lang/stopwords_en.txt" ignoreCase="true"/>

  125.       <filter class="solr.LowerCaseFilterFactory"/>

  126.       <filter class="solr.EnglishPossessiveFilterFactory"/>

  127.       <filter class="solr.KeywordMarkerFilterFactory" protected="protwords.txt"/>

  128.       <filter class="solr.PorterStemFilterFactory"/>

  129.     </analyzer>

  130.     <analyzer type="query">

  131.       <tokenizer class="solr.StandardTokenizerFactory"/>

  132.       <filter class="solr.SynonymGraphFilterFactory" expand="true" ignoreCase="true" synonyms="synonyms.txt"/>

  133.       <filter class="solr.StopFilterFactory" words="lang/stopwords_en.txt" ignoreCase="true"/>

  134.       <filter class="solr.LowerCaseFilterFactory"/>

  135.       <filter class="solr.EnglishPossessiveFilterFactory"/>

  136.       <filter class="solr.KeywordMarkerFilterFactory" protected="protwords.txt"/>

  137.       <filter class="solr.PorterStemFilterFactory"/>

  138.     </analyzer>

  139.   </fieldType>

  140.   <fieldType name="text_en_splitting" class="solr.TextField" autoGeneratePhraseQueries="true" positionIncrementGap="100">

  141.     <analyzer type="index">

  142.       <tokenizer class="solr.WhitespaceTokenizerFactory"/>

  143.       <filter class="solr.StopFilterFactory" words="lang/stopwords_en.txt" ignoreCase="true"/>

  144.       <filter class="solr.WordDelimiterGraphFilterFactory" catenateNumbers="1" generateNumberParts="1" splitOnCaseChange="1" generateWordParts="1" catenateAll="0" catenateWords="1"/>

  145.       <filter class="solr.LowerCaseFilterFactory"/>

  146.       <filter class="solr.KeywordMarkerFilterFactory" protected="protwords.txt"/>

  147.       <filter class="solr.PorterStemFilterFactory"/>

  148.       <filter class="solr.FlattenGraphFilterFactory" />

  149.     </analyzer>

  150.     <analyzer type="query">

  151.       <tokenizer class="solr.WhitespaceTokenizerFactory"/>

  152.       <filter class="solr.SynonymGraphFilterFactory" expand="true" ignoreCase="true" synonyms="synonyms.txt"/>

  153.       <filter class="solr.StopFilterFactory" words="lang/stopwords_en.txt" ignoreCase="true"/>

  154.       <filter class="solr.WordDelimiterGraphFilterFactory" catenateNumbers="0" generateNumberParts="1" splitOnCaseChange="1" generateWordParts="1" catenateAll="0" catenateWords="0"/>

  155.       <filter class="solr.LowerCaseFilterFactory"/>

  156.       <filter class="solr.KeywordMarkerFilterFactory" protected="protwords.txt"/>

  157.       <filter class="solr.PorterStemFilterFactory"/>

  158.     </analyzer>

  159.   </fieldType>

  160.   <fieldType name="text_en_splitting_tight" class="solr.TextField" autoGeneratePhraseQueries="true" positionIncrementGap="100">

  161.     <analyzer type="index">

  162.       <tokenizer class="solr.WhitespaceTokenizerFactory"/>

  163.       <filter class="solr.SynonymGraphFilterFactory" expand="false" ignoreCase="true" synonyms="synonyms.txt"/>

  164.       <filter class="solr.StopFilterFactory" words="lang/stopwords_en.txt" ignoreCase="true"/>

  165.       <filter class="solr.WordDelimiterGraphFilterFactory" catenateNumbers="1" generateNumberParts="0" generateWordParts="0" catenateAll="0" catenateWords="1"/>

  166.       <filter class="solr.LowerCaseFilterFactory"/>

  167.       <filter class="solr.KeywordMarkerFilterFactory" protected="protwords.txt"/>

  168.       <filter class="solr.EnglishMinimalStemFilterFactory"/>

  169.       <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>

  170.       <filter class="solr.FlattenGraphFilterFactory" />

  171.     </analyzer>

  172.     <analyzer type="query">

  173.       <tokenizer class="solr.WhitespaceTokenizerFactory"/>

  174.       <filter class="solr.SynonymGraphFilterFactory" expand="false" ignoreCase="true" synonyms="synonyms.txt"/>

  175.       <filter class="solr.StopFilterFactory" words="lang/stopwords_en.txt" ignoreCase="true"/>

  176.       <filter class="solr.WordDelimiterGraphFilterFactory" catenateNumbers="1" generateNumberParts="0" generateWordParts="0" catenateAll="0" catenateWords="1"/>

  177.       <filter class="solr.LowerCaseFilterFactory"/>

  178.       <filter class="solr.KeywordMarkerFilterFactory" protected="protwords.txt"/>

  179.       <filter class="solr.EnglishMinimalStemFilterFactory"/>

  180.       <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>

  181.     </analyzer>

  182.   </fieldType>

  183.   <fieldType name="text_es" class="solr.TextField" positionIncrementGap="100">

  184.     <analyzer>

  185.       <tokenizer class="solr.StandardTokenizerFactory"/>

  186.       <filter class="solr.LowerCaseFilterFactory"/>

  187.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_es.txt" ignoreCase="true"/>

  188.       <filter class="solr.SpanishLightStemFilterFactory"/>

  189.     </analyzer>

  190.   </fieldType>

  191.   <fieldType name="text_eu" class="solr.TextField" positionIncrementGap="100">

  192.     <analyzer>

  193.       <tokenizer class="solr.StandardTokenizerFactory"/>

  194.       <filter class="solr.LowerCaseFilterFactory"/>

  195.       <filter class="solr.StopFilterFactory" words="lang/stopwords_eu.txt" ignoreCase="true"/>

  196.       <filter class="solr.SnowballPorterFilterFactory" language="Basque"/>

  197.     </analyzer>

  198.   </fieldType>

  199.   <fieldType name="text_fa" class="solr.TextField" positionIncrementGap="100">

  200.     <analyzer>

  201.       <charFilter class="solr.PersianCharFilterFactory"/>

  202.       <tokenizer class="solr.StandardTokenizerFactory"/>

  203.       <filter class="solr.LowerCaseFilterFactory"/>

  204.       <filter class="solr.ArabicNormalizationFilterFactory"/>

  205.       <filter class="solr.PersianNormalizationFilterFactory"/>

  206.       <filter class="solr.StopFilterFactory" words="lang/stopwords_fa.txt" ignoreCase="true"/>

  207.     </analyzer>

  208.   </fieldType>

  209.   <fieldType name="text_fi" class="solr.TextField" positionIncrementGap="100">

  210.     <analyzer>

  211.       <tokenizer class="solr.StandardTokenizerFactory"/>

  212.       <filter class="solr.LowerCaseFilterFactory"/>

  213.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_fi.txt" ignoreCase="true"/>

  214.       <filter class="solr.SnowballPorterFilterFactory" language="Finnish"/>

  215.     </analyzer>

  216.   </fieldType>

  217.   <fieldType name="text_fr" class="solr.TextField" positionIncrementGap="100">

  218.     <analyzer>

  219.       <tokenizer class="solr.StandardTokenizerFactory"/>

  220.       <filter class="solr.ElisionFilterFactory" articles="lang/contractions_fr.txt" ignoreCase="true"/>

  221.       <filter class="solr.LowerCaseFilterFactory"/>

  222.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_fr.txt" ignoreCase="true"/>

  223.       <filter class="solr.FrenchLightStemFilterFactory"/>

  224.     </analyzer>

  225.   </fieldType>

  226.   <fieldType name="text_ga" class="solr.TextField" positionIncrementGap="100">

  227.     <analyzer>

  228.       <tokenizer class="solr.StandardTokenizerFactory"/>

  229.       <filter class="solr.ElisionFilterFactory" articles="lang/contractions_ga.txt" ignoreCase="true"/>

  230.       <filter class="solr.StopFilterFactory" words="lang/hyphenations_ga.txt" ignoreCase="true"/>

  231.       <filter class="solr.IrishLowerCaseFilterFactory"/>

  232.       <filter class="solr.StopFilterFactory" words="lang/stopwords_ga.txt" ignoreCase="true"/>

  233.       <filter class="solr.SnowballPorterFilterFactory" language="Irish"/>

  234.     </analyzer>

  235.   </fieldType>

  236.   <fieldType name="text_general" class="solr.TextField" positionIncrementGap="100" multiValued="true">

  237.     <analyzer type="index">

  238.       <tokenizer class="solr.StandardTokenizerFactory"/>

  239.       <filter class="solr.StopFilterFactory" words="stopwords.txt" ignoreCase="true"/>

  240.       <filter class="solr.LowerCaseFilterFactory"/>

  241.     </analyzer>

  242.     <analyzer type="query">

  243.       <tokenizer class="solr.StandardTokenizerFactory"/>

  244.       <filter class="solr.StopFilterFactory" words="stopwords.txt" ignoreCase="true"/>

  245.       <filter class="solr.SynonymGraphFilterFactory" expand="true" ignoreCase="true" synonyms="synonyms.txt"/>

  246.       <filter class="solr.LowerCaseFilterFactory"/>

  247.     </analyzer>

  248.   </fieldType>

  249.   <fieldType name="text_general_rev" class="solr.TextField" positionIncrementGap="100">

  250.     <analyzer type="index">

  251.       <tokenizer class="solr.StandardTokenizerFactory"/>

  252.       <filter class="solr.StopFilterFactory" words="stopwords.txt" ignoreCase="true"/>

  253.       <filter class="solr.LowerCaseFilterFactory"/>

  254.       <filter class="solr.ReversedWildcardFilterFactory" maxPosQuestion="2" maxFractionAsterisk="0.33" maxPosAsterisk="3" withOriginal="true"/>

  255.     </analyzer>

  256.     <analyzer type="query">

  257.       <tokenizer class="solr.StandardTokenizerFactory"/>

  258.       <filter class="solr.SynonymGraphFilterFactory" expand="true" ignoreCase="true" synonyms="synonyms.txt"/>

  259.       <filter class="solr.StopFilterFactory" words="stopwords.txt" ignoreCase="true"/>

  260.       <filter class="solr.LowerCaseFilterFactory"/>

  261.     </analyzer>

  262.   </fieldType>

  263.   <fieldType name="text_gl" class="solr.TextField" positionIncrementGap="100">

  264.     <analyzer>

  265.       <tokenizer class="solr.StandardTokenizerFactory"/>

  266.       <filter class="solr.LowerCaseFilterFactory"/>

  267.       <filter class="solr.StopFilterFactory" words="lang/stopwords_gl.txt" ignoreCase="true"/>

  268.       <filter class="solr.GalicianStemFilterFactory"/>

  269.     </analyzer>

  270.   </fieldType>

  271.   <fieldType name="text_hi" class="solr.TextField" positionIncrementGap="100">

  272.     <analyzer>

  273.       <tokenizer class="solr.StandardTokenizerFactory"/>

  274.       <filter class="solr.LowerCaseFilterFactory"/>

  275.       <filter class="solr.IndicNormalizationFilterFactory"/>

  276.       <filter class="solr.HindiNormalizationFilterFactory"/>

  277.       <filter class="solr.StopFilterFactory" words="lang/stopwords_hi.txt" ignoreCase="true"/>

  278.       <filter class="solr.HindiStemFilterFactory"/>

  279.     </analyzer>

  280.   </fieldType>

  281.   <fieldType name="text_hu" class="solr.TextField" positionIncrementGap="100">

  282.     <analyzer>

  283.       <tokenizer class="solr.StandardTokenizerFactory"/>

  284.       <filter class="solr.LowerCaseFilterFactory"/>

  285.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_hu.txt" ignoreCase="true"/>

  286.       <filter class="solr.SnowballPorterFilterFactory" language="Hungarian"/>

  287.     </analyzer>

  288.   </fieldType>

  289.   <fieldType name="text_hy" class="solr.TextField" positionIncrementGap="100">

  290.     <analyzer>

  291.       <tokenizer class="solr.StandardTokenizerFactory"/>

  292.       <filter class="solr.LowerCaseFilterFactory"/>

  293.       <filter class="solr.StopFilterFactory" words="lang/stopwords_hy.txt" ignoreCase="true"/>

  294.       <filter class="solr.SnowballPorterFilterFactory" language="Armenian"/>

  295.     </analyzer>

  296.   </fieldType>

  297.   <fieldType name="text_id" class="solr.TextField" positionIncrementGap="100">

  298.     <analyzer>

  299.       <tokenizer class="solr.StandardTokenizerFactory"/>

  300.       <filter class="solr.LowerCaseFilterFactory"/>

  301.       <filter class="solr.StopFilterFactory" words="lang/stopwords_id.txt" ignoreCase="true"/>

  302.       <filter class="solr.IndonesianStemFilterFactory" stemDerivational="true"/>

  303.     </analyzer>

  304.   </fieldType>

  305.   <fieldType name="text_it" class="solr.TextField" positionIncrementGap="100">

  306.     <analyzer>

  307.       <tokenizer class="solr.StandardTokenizerFactory"/>

  308.       <filter class="solr.ElisionFilterFactory" articles="lang/contractions_it.txt" ignoreCase="true"/>

  309.       <filter class="solr.LowerCaseFilterFactory"/>

  310.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_it.txt" ignoreCase="true"/>

  311.       <filter class="solr.ItalianLightStemFilterFactory"/>

  312.     </analyzer>

  313.   </fieldType>

  314.   <fieldType name="text_ja" class="solr.TextField" autoGeneratePhraseQueries="false" positionIncrementGap="100">

  315.     <analyzer>

  316.       <tokenizer class="solr.JapaneseTokenizerFactory" mode="search"/>

  317.       <filter class="solr.JapaneseBaseFormFilterFactory"/>

  318.       <filter class="solr.JapanesePartOfSpeechStopFilterFactory" tags="lang/stoptags_ja.txt"/>

  319.       <filter class="solr.CJKWidthFilterFactory"/>

  320.       <filter class="solr.StopFilterFactory" words="lang/stopwords_ja.txt" ignoreCase="true"/>

  321.       <filter class="solr.JapaneseKatakanaStemFilterFactory" minimumLength="4"/>

  322.       <filter class="solr.LowerCaseFilterFactory"/>

  323.     </analyzer>

  324.   </fieldType>

  325.   <fieldType name="text_ko" class="solr.TextField" positionIncrementGap="100">

  326.     <analyzer>

  327.       <tokenizer class="solr.KoreanTokenizerFactory" decompoundMode="discard" outputUnknownUnigrams="false"/>

  328.       <filter class="solr.KoreanPartOfSpeechStopFilterFactory" />

  329.       <filter class="solr.KoreanReadingFormFilterFactory" />

  330.       <filter class="solr.LowerCaseFilterFactory" />

  331.     </analyzer>

  332.   </fieldType>

  333.   <fieldType name="text_lv" class="solr.TextField" positionIncrementGap="100">

  334.     <analyzer>

  335.       <tokenizer class="solr.StandardTokenizerFactory"/>

  336.       <filter class="solr.LowerCaseFilterFactory"/>

  337.       <filter class="solr.StopFilterFactory" words="lang/stopwords_lv.txt" ignoreCase="true"/>

  338.       <filter class="solr.LatvianStemFilterFactory"/>

  339.     </analyzer>

  340.   </fieldType>

  341.   <fieldType name="text_nl" class="solr.TextField" positionIncrementGap="100">

  342.     <analyzer>

  343.       <tokenizer class="solr.StandardTokenizerFactory"/>

  344.       <filter class="solr.LowerCaseFilterFactory"/>

  345.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_nl.txt" ignoreCase="true"/>

  346.       <filter class="solr.StemmerOverrideFilterFactory" dictionary="lang/stemdict_nl.txt" ignoreCase="false"/>

  347.       <filter class="solr.SnowballPorterFilterFactory" language="Dutch"/>

  348.     </analyzer>

  349.   </fieldType>

  350.   <fieldType name="text_no" class="solr.TextField" positionIncrementGap="100">

  351.     <analyzer>

  352.       <tokenizer class="solr.StandardTokenizerFactory"/>

  353.       <filter class="solr.LowerCaseFilterFactory"/>

  354.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_no.txt" ignoreCase="true"/>

  355.       <filter class="solr.SnowballPorterFilterFactory" language="Norwegian"/>

  356.     </analyzer>

  357.   </fieldType>

  358.   <fieldType name="text_pt" class="solr.TextField" positionIncrementGap="100">

  359.     <analyzer>

  360.       <tokenizer class="solr.StandardTokenizerFactory"/>

  361.       <filter class="solr.LowerCaseFilterFactory"/>

  362.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_pt.txt" ignoreCase="true"/>

  363.       <filter class="solr.PortugueseLightStemFilterFactory"/>

  364.     </analyzer>

  365.   </fieldType>

  366.   <fieldType name="text_ro" class="solr.TextField" positionIncrementGap="100">

  367.     <analyzer>

  368.       <tokenizer class="solr.StandardTokenizerFactory"/>

  369.       <filter class="solr.LowerCaseFilterFactory"/>

  370.       <filter class="solr.StopFilterFactory" words="lang/stopwords_ro.txt" ignoreCase="true"/>

  371.       <filter class="solr.SnowballPorterFilterFactory" language="Romanian"/>

  372.     </analyzer>

  373.   </fieldType>

  374.   <fieldType name="text_ru" class="solr.TextField" positionIncrementGap="100">

  375.     <analyzer>

  376.       <tokenizer class="solr.StandardTokenizerFactory"/>

  377.       <filter class="solr.LowerCaseFilterFactory"/>

  378.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_ru.txt" ignoreCase="true"/>

  379.       <filter class="solr.SnowballPorterFilterFactory" language="Russian"/>

  380.     </analyzer>

  381.   </fieldType>

  382.   <fieldType name="text_sv" class="solr.TextField" positionIncrementGap="100">

  383.     <analyzer>

  384.       <tokenizer class="solr.StandardTokenizerFactory"/>

  385.       <filter class="solr.LowerCaseFilterFactory"/>

  386.       <filter class="solr.StopFilterFactory" format="snowball" words="lang/stopwords_sv.txt" ignoreCase="true"/>

  387.       <filter class="solr.SnowballPorterFilterFactory" language="Swedish"/>

  388.     </analyzer>

  389.   </fieldType>

  390.   <fieldType name="text_th" class="solr.TextField" positionIncrementGap="100">

  391.     <analyzer>

  392.       <tokenizer class="solr.ThaiTokenizerFactory"/>

  393.       <filter class="solr.LowerCaseFilterFactory"/>

  394.       <filter class="solr.StopFilterFactory" words="lang/stopwords_th.txt" ignoreCase="true"/>

  395.     </analyzer>

  396.   </fieldType>

  397.   <fieldType name="text_tr" class="solr.TextField" positionIncrementGap="100">

  398.     <analyzer>

  399.       <tokenizer class="solr.StandardTokenizerFactory"/>

  400.       <filter class="solr.TurkishLowerCaseFilterFactory"/>

  401.       <filter class="solr.StopFilterFactory" words="lang/stopwords_tr.txt" ignoreCase="false"/>

  402.       <filter class="solr.SnowballPorterFilterFactory" language="Turkish"/>

  403.     </analyzer>

  404.   </fieldType>

  405.   <fieldType name="text_ws" class="solr.TextField" positionIncrementGap="100">

  406.     <analyzer>

  407.       <tokenizer class="solr.WhitespaceTokenizerFactory"/>

  408.     </analyzer>

  409.   </fieldType>

  410. 

  411.   <fieldType name="text_email_url" class="solr.TextField">

  412.     <analyzer>

  413.       <tokenizer class="solr.UAX29URLEmailTokenizerFactory"/>

  414.       <filter class="solr.TypeTokenFilterFactory" types="email_url_types.txt" useWhitelist="true"/>

  415.     </analyzer>

  416.   </fieldType>

  417. 

  418.   <fieldType name="text_shingles" class="solr.TextField" positionIncrementGap="100" multiValued="true">

  419.     <analyzer type="index">

  420.       <tokenizer class="solr.StandardTokenizerFactory"/>

  421.       <!-- <filter class="solr.StopFilterFactory" words="lang/stopwords_en.txt" ignoreCase="false" /> -->

  422.       <filter class="solr.LengthFilterFactory" min="2" max="18"/>

  423.       <filter class="solr.LowerCaseFilterFactory"/>

  424.       <filter class="solr.PatternReplaceFilterFactory" pattern="(^[^a-z]+$)" replacement="" replace="all"/>

  425.       <filter class="solr.ShingleFilterFactory" minShingleSize="3"  maxShingleSize="3"

  426.              outputUnigrams="false" outputUnigramsIfNoShingles="false" tokenSeparator=" " fillerToken="*"/>

  427.       <filter class="solr.PatternReplaceFilterFactory" pattern="(.*[\*].*)"  replacement=""/>

  428.       <filter class="solr.TrimFilterFactory"/>

  429. 

  430.       <!-- PRFF could have removed everything down to an empty string, remove if so -->

  431.       <filter class="solr.LengthFilterFactory" min="1" max="100"/>

  432.     </analyzer>

  433.     <analyzer type="query">

  434.       <tokenizer class="solr.KeywordTokenizerFactory"/>

  435.       <filter class="solr.LowerCaseFilterFactory"/>

  436.     </analyzer>

  437.   </fieldType>

  438. 

  439.   <field name="id" type="string" multiValued="false" indexed="true" required="true" stored="true"/>

  440.   <field name="_version_" type="plong" indexed="true" stored="true"/>

  441.   <field name="content_type" type="string" indexed="true" stored="true"/>

  442.   <field name="doc_type" type="string" indexed="true" stored="true"/>

  443.   <field name="title" type="string" indexed="true" stored="true"/>

  444.   <field name="language" type="string" indexed="true" stored="true"/>

  445.   <field name="content" type="text_general" multiValued="false" indexed="true" stored="true"/>

  446.   <field name="text_shingles" type="text_shingles" indexed="true" stored="false"/>

  447.   <field name="_text_" type="text_general" multiValued="true" indexed="true" stored="false"/>

  448. 

  449.   <dynamicField name="*_txt_en_split_tight" type="text_en_splitting_tight" indexed="true" stored="true"/>

  450.   <dynamicField name="*_descendent_path" type="descendent_path" indexed="true" stored="true"/>

  451.   <dynamicField name="*_ancestor_path" type="ancestor_path" indexed="true" stored="true"/>

  452.   <dynamicField name="*_txt_en_split" type="text_en_splitting" indexed="true" stored="true"/>

  453.   <dynamicField name="*_coordinate" type="pdouble" indexed="true" stored="false"/>

  454.   <dynamicField name="ignored_*" type="ignored" multiValued="true"/>

  455.   <dynamicField name="*_txt_rev" type="text_general_rev" indexed="true" stored="true"/>

  456.   <dynamicField name="*_phon_en" type="phonetic_en" indexed="true" stored="true"/>

  457.   <dynamicField name="*_s_lower" type="lowercase" indexed="true" stored="true"/>

  458.   <dynamicField name="*_txt_cjk" type="text_cjk" indexed="true" stored="true"/>

  459.   <dynamicField name="random_*" type="random"/>

  460.   <dynamicField name="*_txt_en" type="text_en" indexed="true" stored="true"/>

  461.   <dynamicField name="*_txt_ar" type="text_ar" indexed="true" stored="true"/>

  462.   <dynamicField name="*_txt_bg" type="text_bg" indexed="true" stored="true"/>

  463.   <dynamicField name="*_txt_ca" type="text_ca" indexed="true" stored="true"/>

  464.   <dynamicField name="*_txt_cz" type="text_cz" indexed="true" stored="true"/>

  465.   <dynamicField name="*_txt_da" type="text_da" indexed="true" stored="true"/>

  466.   <dynamicField name="*_txt_de" type="text_de" indexed="true" stored="true"/>

  467.   <dynamicField name="*_txt_el" type="text_el" indexed="true" stored="true"/>

  468.   <dynamicField name="*_txt_es" type="text_es" indexed="true" stored="true"/>

  469.   <dynamicField name="*_txt_eu" type="text_eu" indexed="true" stored="true"/>

  470.   <dynamicField name="*_txt_fa" type="text_fa" indexed="true" stored="true"/>

  471.   <dynamicField name="*_txt_fi" type="text_fi" indexed="true" stored="true"/>

  472.   <dynamicField name="*_txt_fr" type="text_fr" indexed="true" stored="true"/>

  473.   <dynamicField name="*_txt_ga" type="text_ga" indexed="true" stored="true"/>

  474.   <dynamicField name="*_txt_gl" type="text_gl" indexed="true" stored="true"/>

  475.   <dynamicField name="*_txt_hi" type="text_hi" indexed="true" stored="true"/>

  476.   <dynamicField name="*_txt_hu" type="text_hu" indexed="true" stored="true"/>

  477.   <dynamicField name="*_txt_hy" type="text_hy" indexed="true" stored="true"/>

  478.   <dynamicField name="*_txt_id" type="text_id" indexed="true" stored="true"/>

  479.   <dynamicField name="*_txt_it" type="text_it" indexed="true" stored="true"/>

  480.   <dynamicField name="*_txt_ja" type="text_ja" indexed="true" stored="true"/>

  481.   <dynamicField name="*_txt_ko" type="text_ko" indexed="true" stored="true"/>

  482.   <dynamicField name="*_txt_lv" type="text_lv" indexed="true" stored="true"/>

  483.   <dynamicField name="*_txt_nl" type="text_nl" indexed="true" stored="true"/>

  484.   <dynamicField name="*_txt_no" type="text_no" indexed="true" stored="true"/>

  485.   <dynamicField name="*_txt_pt" type="text_pt" indexed="true" stored="true"/>

  486.   <dynamicField name="*_txt_ro" type="text_ro" indexed="true" stored="true"/>

  487.   <dynamicField name="*_txt_ru" type="text_ru" indexed="true" stored="true"/>

  488.   <dynamicField name="*_txt_sv" type="text_sv" indexed="true" stored="true"/>

  489.   <dynamicField name="*_txt_th" type="text_th" indexed="true" stored="true"/>

  490.   <dynamicField name="*_txt_tr" type="text_tr" indexed="true" stored="true"/>

  491.   <dynamicField name="*_point" type="point" indexed="true" stored="true"/>

  492.   <dynamicField name="*_srpt" type="location_rpt" indexed="true" stored="true"/>

  493.   <dynamicField name="attr_*" type="text_general" multiValued="true" indexed="true" stored="true"/>

  494.   <dynamicField name="*_l_ns" type="plong" indexed="true" stored="false"/>

  495.   <dynamicField name="*_s_ns" type="string" indexed="true" stored="false"/>

  496.   <dynamicField name="*_txt" type="text_general" indexed="true" stored="true"/>

  497.   <dynamicField name="*_dts" type="pdate" multiValued="true" indexed="true" stored="true"/>

  498.   <dynamicField name="*_is" type="pints" indexed="true" stored="true"/>

  499.   <dynamicField name="*_ss" type="strings" indexed="true" stored="true"/>

  500.   <dynamicField name="*_ls" type="plongs" indexed="true" stored="true"/>

  501.   <dynamicField name="*_bs" type="booleans" indexed="true" stored="true"/>

  502.   <dynamicField name="*_fs" type="pfloats" indexed="true" stored="true"/>

  503.   <dynamicField name="*_ds" type="pdoubles" indexed="true" stored="true"/>

  504.   <dynamicField name="*_dt" type="pdate" indexed="true" stored="true"/>

  505.   <dynamicField name="*_ws" type="text_ws" indexed="true" stored="true"/>

  506.   <dynamicField name="*_i" type="pint" indexed="true" stored="true"/>

  507.   <dynamicField name="*_s" type="string" indexed="true" stored="true"/>

  508.   <dynamicField name="*_l" type="plong" indexed="true" stored="true"/>

  509.   <dynamicField name="*_t" type="text_general" indexed="true" stored="true"/>

  510.   <dynamicField name="*_b" type="boolean" indexed="true" stored="true"/>

  511.   <dynamicField name="*_f" type="pfloat" indexed="true" stored="true"/>

  512.   <dynamicField name="*_d" type="pdouble" indexed="true" stored="true"/>

  513.   <dynamicField name="*_p" type="location" indexed="true" stored="true"/>

  514.   <dynamicField name="*_c" type="currency" indexed="true" stored="true"/>

  515. 

  516.   <copyField source="content" dest="text_shingles"/>

  517.   <copyField source="*" dest="_text_"/>

  518. 

  519.   <!-- ADDED BY SIMON BOWIE 2022-04-04 -->

  520.   <copyField source="content" dest="year"/>

  521.   <field name="year" type="year" indexed="true" stored="true"/>

  522. 

  523.   <fieldType name="year" class="solr.TextField" positionIncrementGap="100">

  524.     <analyzer>

  525.         <tokenizer class="solr.PatternTokenizerFactory" pattern="=D[^\s]*\s[^\s]*\s[^\s]*\s[^\s]*\s(\d{4})" group="1" />

  526.     </analyzer>

  527.   </fieldType>

  528.   <!-- END -->

  529. 

  530. </schema>