exhibits_solr_conf 0.0.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/.gitignore +12 -0
- data/.hound.yml +2 -0
- data/.rspec +1 -0
- data/.rubocop.yml +8 -0
- data/.rubocop_todo.yml +24 -0
- data/.travis.yml +11 -0
- data/Gemfile +4 -0
- data/LICENSE +13 -0
- data/README.md +78 -0
- data/Rakefile +17 -0
- data/bin/console +14 -0
- data/bin/setup +7 -0
- data/exhibits_solr_conf.gemspec +29 -0
- data/lib/exhibits_solr_conf/tasks/configure_solr.rake +11 -0
- data/lib/exhibits_solr_conf/version.rb +3 -0
- data/lib/exhibits_solr_conf.rb +6 -0
- data/solr_conf_4_testing/_rest_managed.json +3 -0
- data/solr_conf_4_testing/schema.xml +529 -0
- data/solr_conf_4_testing/solrconfig.xml +881 -0
- data/solr_conf_4_testing/stopwords_punctuation.txt +22 -0
- data/solr_conf_4_testing/synonyms.txt +73 -0
- data/solr_conf_4_testing/synonyms_both_anchors.txt +47 -0
- data/solr_conf_4_testing/synonyms_left_anchor.txt +45 -0
- data/solr_conf_4_testing/synonyms_right_anchor.txt +47 -0
- metadata +168 -0
|
@@ -0,0 +1,529 @@
|
|
|
1
|
+
<?xml version="1.0" encoding="UTF-8" ?>
|
|
2
|
+
<schema name="Stanford Searchworks" version="1.5">
|
|
3
|
+
<uniqueKey>id</uniqueKey>
|
|
4
|
+
|
|
5
|
+
<fields>
|
|
6
|
+
<!-- needed by some of Solr 4.0 functionality like transaction log or partial documents update -->
|
|
7
|
+
<field name="_version_" type="long" indexed="true" stored="true"/>
|
|
8
|
+
<field name="timestamp" type="date" indexed="true" stored="true" default="NOW" />
|
|
9
|
+
|
|
10
|
+
<field name="id" type="string_punct_stop" indexed="true" stored="true" required="true" />
|
|
11
|
+
<field name="created" type="date" indexed="true" stored="true" default="NOW/SECOND" />
|
|
12
|
+
<field name="last_updated" type="date" indexed="true" stored="true" default="NOW/SECOND" />
|
|
13
|
+
<!-- entire marc bib record -->
|
|
14
|
+
<field name="marcxml" type="string" indexed="false" stored="true" />
|
|
15
|
+
<!-- all_search: catch-all field for searchable text; stored for hit highlighting />-->
|
|
16
|
+
<field name="all_search" type="text" indexed="true" stored="true" termVectors="true" termPositions="true" termOffsets="true" multiValued="true" />
|
|
17
|
+
<field name="all_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
18
|
+
<field name="vern_all_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
19
|
+
|
|
20
|
+
<!-- Format Field: facet and display -->
|
|
21
|
+
<field name="format" type="string" indexed="true" stored="true" multiValued="true" />
|
|
22
|
+
|
|
23
|
+
<!-- Language Field: facet and display -->
|
|
24
|
+
<field name="language" type="string" indexed="true" stored="true" multiValued="true" />
|
|
25
|
+
|
|
26
|
+
<!-- Standard Number Fields -->
|
|
27
|
+
<!-- allow end users to search our index by ISBN and ISSN; more values than display isbn/issn.-->
|
|
28
|
+
<!-- isbn_search is type text to accommodate hyphens that could be entered by end users -->
|
|
29
|
+
<field name="isbn_search" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true" />
|
|
30
|
+
<!-- text for issn to accommodate hyphen present or not -->
|
|
31
|
+
<field name="issn_search" type="text" indexed="true" stored="true" multiValued="true" omitNorms="true"/>
|
|
32
|
+
<!-- display fields to allow external lookups (e.g. Google book search, xISBN, xISSN ...) -->
|
|
33
|
+
<field name="isbn_display" type="string" indexed="false" stored="true" multiValued="true" />
|
|
34
|
+
<field name="issn_display" type="string" indexed="false" stored="true" multiValued="true" />
|
|
35
|
+
<field name="lccn" type="string" indexed="false" stored="true" />
|
|
36
|
+
<field name="oclc" type="string_punct_stop" indexed="true" stored="true" multiValued="true" />
|
|
37
|
+
|
|
38
|
+
<!-- Title Search Fields -->
|
|
39
|
+
<field name="title_245a_exact_search" type="text_anchored" indexed="true" stored="true" />
|
|
40
|
+
<field name="title_245a_search" type="text" indexed="true" stored="true" />
|
|
41
|
+
<field name="vern_title_245a_search" type="textNoStem" indexed="true" stored="true" />
|
|
42
|
+
<field name="title_245a_unstem_search" type="textNoStem" indexed="true" stored="true" />
|
|
43
|
+
<field name="title_245_search" type="text" indexed="true" stored="true" />
|
|
44
|
+
<field name="vern_title_245_search" type="textNoStem" indexed="true" stored="true" />
|
|
45
|
+
<field name="title_245_unstem_search" type="textNoStem" indexed="true" stored="true" />
|
|
46
|
+
<field name="title_uniform_search" type="text" indexed="true" stored="true" />
|
|
47
|
+
<field name="vern_title_uniform_search" type="textNoStem" indexed="true" stored="true" />
|
|
48
|
+
<field name="title_uniform_unstem_search" type="textNoStem" indexed="true" stored="true" />
|
|
49
|
+
<field name="title_variant_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
50
|
+
<field name="vern_title_variant_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
51
|
+
<field name="title_variant_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
52
|
+
<field name="title_related_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
53
|
+
<field name="vern_title_related_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
54
|
+
<field name="title_related_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
55
|
+
|
|
56
|
+
<!-- Title Display fields -->
|
|
57
|
+
<field name="title_display" type="string" indexed="false" stored="true" />
|
|
58
|
+
<field name="vern_title_display" type="string" indexed="false" stored="true" />
|
|
59
|
+
<field name="title_245a_display" type="string" indexed="false" stored="true" />
|
|
60
|
+
<field name="vern_title_245a_display" type="string" indexed="false" stored="true" />
|
|
61
|
+
<field name="title_245c_display" type="string" indexed="false" stored="true" />
|
|
62
|
+
<field name="vern_title_245c_display" type="string" indexed="false" stored="true" />
|
|
63
|
+
<field name="title_full_display" type="string" indexed="false" stored="true" />
|
|
64
|
+
<field name="vern_title_full_display" type="string" indexed="false" stored="true" />
|
|
65
|
+
<field name="title_uniform_display" type="string" indexed="false" stored="true" />
|
|
66
|
+
<field name="vern_title_uniform_display" type="string" indexed="false" stored="true" />
|
|
67
|
+
<field name="title_variant_display" type="string" indexed="false" stored="true" multiValued="true" />
|
|
68
|
+
|
|
69
|
+
<field name="title_sort" type="alphaSort" indexed="true" stored="true" />
|
|
70
|
+
|
|
71
|
+
<!-- Series Search Fields -->
|
|
72
|
+
<field name="series_exact_search" type="text_anchored" indexed="true" stored="true" multiValued="true" />
|
|
73
|
+
<field name="series_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
74
|
+
<field name="vern_series_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
75
|
+
<field name="series_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
76
|
+
|
|
77
|
+
<!-- Author Title Search Fields -->
|
|
78
|
+
<field name="author_title_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
79
|
+
|
|
80
|
+
<!-- Author Search Fields -->
|
|
81
|
+
<field name="author_1xx_search" type="text" indexed="true" stored="true" />
|
|
82
|
+
<field name="vern_author_1xx_search" type="textNoStem" indexed="true" stored="true" />
|
|
83
|
+
<field name="author_1xx_unstem_search" type="textNoStem" indexed="true" stored="true" />
|
|
84
|
+
<field name="author_7xx_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
85
|
+
<field name="vern_author_7xx_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
86
|
+
<field name="author_7xx_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
87
|
+
<field name="author_8xx_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
88
|
+
<field name="vern_author_8xx_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
89
|
+
<field name="author_8xx_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
90
|
+
|
|
91
|
+
<!-- Author Facet Fields -->
|
|
92
|
+
<field name="author_person_facet" type="string" indexed="true" stored="true" multiValued="true" />
|
|
93
|
+
<field name="author_other_facet" type="string" indexed="true" stored="true" multiValued="true" />
|
|
94
|
+
|
|
95
|
+
<field name="author_sort" type="alphaSort" indexed="true" stored="true" />
|
|
96
|
+
|
|
97
|
+
<!-- Author Display Fields -->
|
|
98
|
+
<field name="author_person_display" type="string" indexed="false" stored="true" multiValued="true" />
|
|
99
|
+
<field name="vern_author_person_display" type="string" indexed="false" stored="true" />
|
|
100
|
+
<field name="author_person_full_display" type="string" indexed="false" stored="true" multiValued="true" />
|
|
101
|
+
<field name="vern_author_person_full_display" type="string" indexed="false" stored="true" />
|
|
102
|
+
<field name="author_corp_display" type="string" indexed="false" stored="true" multiValued="true"/>
|
|
103
|
+
<field name="vern_author_corp_display" type="string" indexed="false" stored="true" />
|
|
104
|
+
<field name="author_meeting_display" type="string" indexed="false" stored="true" multiValued="true"/>
|
|
105
|
+
<field name="vern_author_meeting_display" type="string" indexed="false" stored="true" />
|
|
106
|
+
|
|
107
|
+
<!-- Subject Search Fields -->
|
|
108
|
+
<field name="topic_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
109
|
+
<field name="vern_topic_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
110
|
+
<field name="topic_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
111
|
+
<field name="topic_subx_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
112
|
+
<field name="vern_topic_subx_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
113
|
+
<field name="topic_subx_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
114
|
+
<field name="geographic_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
115
|
+
<field name="vern_geographic_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
116
|
+
<field name="geographic_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
117
|
+
<field name="geographic_subz_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
118
|
+
<field name="vern_geographic_subz_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
119
|
+
<field name="geographic_subz_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
120
|
+
<field name="subject_other_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
121
|
+
<field name="vern_subject_other_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
122
|
+
<field name="subject_other_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
123
|
+
<field name="subject_other_subvy_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
124
|
+
<field name="vern_subject_other_subvy_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
125
|
+
<field name="subject_other_subvy_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
126
|
+
<field name="subject_all_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
127
|
+
<field name="vern_subject_all_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
128
|
+
<field name="subject_all_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
129
|
+
|
|
130
|
+
<field name="db_az_subject" type="string" indexed="true" stored="true" multiValued="true" />
|
|
131
|
+
<field name="db_az_subject_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
132
|
+
|
|
133
|
+
<!-- Subject Facet Fields -->
|
|
134
|
+
<field name="topic_facet" type="string" indexed="true" stored="true" multiValued="true" />
|
|
135
|
+
<field name="geographic_facet" type="string" indexed="true" stored="true" multiValued="true" />
|
|
136
|
+
<field name="era_facet" type="string" indexed="true" stored="true" multiValued="true" />
|
|
137
|
+
|
|
138
|
+
<!-- Subject Display Fields -->
|
|
139
|
+
<field name="topic_display" type="string" indexed="false" stored="true" multiValued="true" />
|
|
140
|
+
<field name="subject_other_display" type="string" indexed="false" stored="true" multiValued="true" />
|
|
141
|
+
|
|
142
|
+
<!-- Publishing Fields -->
|
|
143
|
+
<field name="pub_search" type="text" indexed="true" stored="true" omitNorms="true" multiValued="true"/>
|
|
144
|
+
<field name="vern_pub_search" type="textNoStem" indexed="true" stored="true" omitNorms="true" multiValued="true"/>
|
|
145
|
+
<field name="pub_country" type="text" indexed="true" stored="true" omitNorms="true"/>
|
|
146
|
+
<!-- TODO: should pub_date_search be a date or a text field? -->
|
|
147
|
+
<field name="pub_date_search" type="text" indexed="true" stored="true" omitNorms="true"/>
|
|
148
|
+
<field name="pub_date_sort" type="alphaSort" indexed="true" stored="true" />
|
|
149
|
+
<!-- Pub Date Facet Fields -->
|
|
150
|
+
<!-- pub_date was facet and display, now deprecated -->
|
|
151
|
+
<field name="pub_date" type="string" indexed="true" stored="true" />
|
|
152
|
+
<field name="pub_year_tisim" type="tint" indexed="true" stored="true" multiValued="true" />
|
|
153
|
+
<field name="pub_display" type="string" indexed="false" stored="true" multiValued="true"/>
|
|
154
|
+
<field name="pub_date_display" type="string" indexed="false" stored="true"/>
|
|
155
|
+
<field name="imprint_display" type="string" indexed="false" stored="true" multiValued="true"/>
|
|
156
|
+
|
|
157
|
+
<!-- URL Fields -->
|
|
158
|
+
<field name="url_fulltext" type="string" indexed="false" stored="true" multiValued="true"/>
|
|
159
|
+
<field name="url_suppl" type="string" indexed="false" stored="true" multiValued="true"/>
|
|
160
|
+
<!-- sfx urls should rarely occur more than once in a marc bib record -->
|
|
161
|
+
<field name="url_sfx" type="string" indexed="false" stored="true" multiValued="true" />
|
|
162
|
+
<field name="url_restricted" type="string" indexed="false" stored="true" multiValued="true" />
|
|
163
|
+
|
|
164
|
+
<!-- Physical Fields -->
|
|
165
|
+
<field name="physical" type="text" indexed="true" stored="true" multiValued="true" />
|
|
166
|
+
<field name="vern_physical" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
167
|
+
|
|
168
|
+
<!-- Table of Contents -->
|
|
169
|
+
<field name="toc_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
170
|
+
<field name="vern_toc_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
171
|
+
<field name="toc_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
172
|
+
<!-- Context -->
|
|
173
|
+
<field name="context_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
174
|
+
<field name="vern_context_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
175
|
+
<field name="context_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
176
|
+
<!-- Summary -->
|
|
177
|
+
<field name="summary_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
178
|
+
<field name="vern_summary_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
179
|
+
<field name="summary_unstem_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
180
|
+
<field name="summary_display" type="string" indexed="false" stored="true" multiValued="true" />
|
|
181
|
+
<!-- Award -->
|
|
182
|
+
<field name="award_search" type="text" indexed="true" stored="true" multiValued="true" />
|
|
183
|
+
|
|
184
|
+
<!-- Item Info Fields (derived from 999) -->
|
|
185
|
+
<!-- Call Number Fields -->
|
|
186
|
+
<field name="callnum_search" type="callnum_ws" indexed="true" stored="true" multiValued="true"/>
|
|
187
|
+
|
|
188
|
+
<!-- for nearby on shelf: term lookups to get next X alpha sorted terms -->
|
|
189
|
+
<field name="shelfkey" type="alphaSort" indexed="true" stored="true" multiValued="true"/>
|
|
190
|
+
<field name="reverse_shelfkey" type="alphaSort" indexed="true" stored="true" multiValued="true"/>
|
|
191
|
+
|
|
192
|
+
<field name="barcode_search" type="string_punct_stop" indexed="true" stored="true" multiValued="true" />
|
|
193
|
+
<field name="preferred_barcode" type="string" indexed="false" stored="true" />
|
|
194
|
+
<field name="access_facet" type="string" indexed="true" stored="true" multiValued="true" />
|
|
195
|
+
<field name="building_facet" type="string" indexed="true" stored="true" multiValued="true" />
|
|
196
|
+
<!-- barcode -|- lib -|- location -|- lopped_callnum -|- shelfkey -|- reverse_shelfkey -|- full_callnum -|- callnum_show_sort -->
|
|
197
|
+
<field name="item_display" type="string" indexed="false" stored="true" multiValued="true" />
|
|
198
|
+
|
|
199
|
+
<!-- lib -|- location -|- note -|- holdings summary -|- last received -->
|
|
200
|
+
<field name="mhld_display" type="string" indexed="false" stored="true" multiValued="true" />
|
|
201
|
+
|
|
202
|
+
<!-- Course Reserve Fields -->
|
|
203
|
+
<field name="crez_course_id_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
204
|
+
<field name="crez_course_name_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
205
|
+
<field name="crez_instructor_search" type="textNoStem" indexed="true" stored="true" multiValued="true" />
|
|
206
|
+
<!-- instructor, course facet field names are exposed to end users in SW -->
|
|
207
|
+
<field name="instructor" type="string" indexed="true" stored="true" multiValued="true" />
|
|
208
|
+
<field name="course" type="string" indexed="true" stored="true" multiValued="true" />
|
|
209
|
+
<!-- crez_course_info is a facet and display field at this time (2012-03-21) -->
|
|
210
|
+
<field name="crez_course_info" type="string" indexed="true" stored="true" multiValued="true" />
|
|
211
|
+
<!-- the next two facets are not used in SW at this time -->
|
|
212
|
+
<field name="crez_dept_facet" type="string" indexed="true" stored="true" multiValued="true" />
|
|
213
|
+
<field name="crez_desk_facet" type="string" indexed="true" stored="true" multiValued="true" />
|
|
214
|
+
|
|
215
|
+
<!-- *************** additional fields for DOR objects ****************** -->
|
|
216
|
+
<field name="druid" type="string_punct_stop" indexed="true" stored="true" />
|
|
217
|
+
<field name="modsxml" type="string" indexed="false" stored="true" />
|
|
218
|
+
<!-- collection (facet and display): "sirsi" or, for DOR items, the id of their parent coll -->
|
|
219
|
+
<field name="collection" type="string" indexed="true" stored="true" multiValued="true" omitNorms="true" />
|
|
220
|
+
<!-- collection_with_title: easy way to indicate item's parent coll title in UI (may be deprecated in future) -->
|
|
221
|
+
<field name="collection_with_title" type="string" indexed="false" stored="true" multiValued="true" />
|
|
222
|
+
<!-- collection_search: allow searching within collections for aggr. colls -->
|
|
223
|
+
<field name="collection_search" type="string_punct_stop" indexed="true" stored="true" multiValued="true" omitNorms="true" />
|
|
224
|
+
<!-- display_type: used by UI code, e.g. 'file' or 'image' -->
|
|
225
|
+
<field name="display_type" type="string" indexed="true" stored="true" multiValued="true" omitNorms="true" />
|
|
226
|
+
<!-- used to determine when something is a digital collection -->
|
|
227
|
+
<field name="collection_type" type="string" indexed="true" stored="true" multiValued="true" />
|
|
228
|
+
<!-- file_id: ids of files (including images) in the digital stacks -->
|
|
229
|
+
<field name="file_id" type="string" indexed="false" stored="true" multiValued="true"/>
|
|
230
|
+
|
|
231
|
+
<!-- *************** dynamic field types ****************** -->
|
|
232
|
+
<!--
|
|
233
|
+
<dynamicField name="*_unstem_search" type="textNoStem" stored="true" indexed="true" multiValued="true" />
|
|
234
|
+
<dynamicField name="*_search" type="text" stored="true" indexed="true" multiValued="true" />
|
|
235
|
+
<dynamicField name="*_facet" type="string" stored="true" indexed="true" multiValued="true" />
|
|
236
|
+
<dynamicField name="*_display" type="string" stored="true" indexed="false" multiValued="true"/>
|
|
237
|
+
-->
|
|
238
|
+
<dynamicField name="*_si" type="string" stored="true" indexed="true" omitNorms="true" />
|
|
239
|
+
<dynamicField name="*_sim" type="string" stored="true" indexed="true" multiValued="true" omitNorms="true" />
|
|
240
|
+
<dynamicField name="*_ss" type="string" stored="true" indexed="false" omitNorms="true" />
|
|
241
|
+
<dynamicField name="*_ssm" type="string" stored="true" indexed="false" multiValued="true" omitNorms="true" />
|
|
242
|
+
<dynamicField name="*_ssi" type="string" stored="true" indexed="true" omitNorms="true" />
|
|
243
|
+
<dynamicField name="*_ssim" type="string" stored="true" indexed="true" multiValued="true" omitNorms="true" />
|
|
244
|
+
<dynamicField name="*_isi" type="tint" stored="true" indexed="true" omitNorms="true" />
|
|
245
|
+
<dynamicField name="*_isim" type="tint" stored="true" indexed="true" multiValued="true" omitNorms="true" />
|
|
246
|
+
<dynamicField name="*_sort" type="alphaSort" stored="true" indexed="true"/>
|
|
247
|
+
<dynamicField name="*_xml" type="string" stored="true" indexed="false" omitNorms="true" />
|
|
248
|
+
<dynamicField name="cjk_*" type="text_cjk" stored="true" indexed="true" multiValued="true" />
|
|
249
|
+
<dynamicField name="*_hsim" type="string_hierarch" stored="true" indexed="true" multiValued="true" />
|
|
250
|
+
<dynamicField name="*_tesim" type="text" stored="true" indexed="true" multiValued="true" omitNorms="true" />
|
|
251
|
+
<dynamicField name="*_bsi" type="boolean" stored="true" indexed="true" multiValued="true" omitNorms="true" />
|
|
252
|
+
<dynamicField name="*_ng" type="text_en_ng" stored="false" indexed="true" multiValued="true"/>
|
|
253
|
+
<dynamicField name="*_pt" type="location" stored="true" indexed="true"/>
|
|
254
|
+
<dynamicField name="*_bbox" type="location_rpt" stored="true" indexed="true" multiValued="true"/>
|
|
255
|
+
</fields>
|
|
256
|
+
|
|
257
|
+
<!-- copy fields -->
|
|
258
|
+
<copyField source="collection" dest="collection_search" />
|
|
259
|
+
<copyField source="pub_date" dest="pub_date_search" />
|
|
260
|
+
<copyField source="db_az_subject" dest="db_az_subject_search" />
|
|
261
|
+
|
|
262
|
+
<!-- unstemmed and anchored search fields: title -->
|
|
263
|
+
<copyField source="title_245a_search" dest="title_245a_exact_search" />
|
|
264
|
+
<copyField source="title_245a_search" dest="title_245a_unstem_search" />
|
|
265
|
+
<copyField source="title_245_search" dest="title_245_unstem_search" />
|
|
266
|
+
<copyField source="title_uniform_search" dest="title_uniform_unstem_search" />
|
|
267
|
+
<copyField source="title_variant_search" dest="title_variant_unstem_search" />
|
|
268
|
+
<copyField source="title_related_search" dest="title_related_unstem_search" />
|
|
269
|
+
<!-- unstemmed search fields: author -->
|
|
270
|
+
<copyField source="author_1xx_search" dest="author_1xx_unstem_search" />
|
|
271
|
+
<copyField source="author_7xx_search" dest="author_7xx_unstem_search" />
|
|
272
|
+
<copyField source="author_8xx_search" dest="author_8xx_unstem_search" />
|
|
273
|
+
<!-- unstemmed search fields: subject -->
|
|
274
|
+
<copyField source="topic_search" dest="topic_unstem_search" />
|
|
275
|
+
<copyField source="topic_subx_search" dest="topic_subx_unstem_search" />
|
|
276
|
+
<copyField source="geographic_search" dest="geographic_unstem_search" />
|
|
277
|
+
<copyField source="geographic_subz_search" dest="geographic_subz_unstem_search" />
|
|
278
|
+
<copyField source="subject_other_search" dest="subject_other_unstem_search" />
|
|
279
|
+
<copyField source="subject_other_subvy_search" dest="subject_other_subvy_unstem_search" />
|
|
280
|
+
<copyField source="subject_all_search" dest="subject_all_unstem_search" />
|
|
281
|
+
<!-- unstemmed search fields: toc/summary -->
|
|
282
|
+
<copyField source="toc_search" dest="toc_unstem_search" />
|
|
283
|
+
<copyField source="context_search" dest="context_unstem_search" />
|
|
284
|
+
<copyField source="summary_search" dest="summary_unstem_search" />
|
|
285
|
+
<!-- other unstemmed search fields -->
|
|
286
|
+
<copyField source="series_search" dest="series_unstem_search" />
|
|
287
|
+
<copyField source="all_search" dest="all_unstem_search" />
|
|
288
|
+
|
|
289
|
+
<!-- course reserve fields -->
|
|
290
|
+
<copyField source="crez_instructor_search" dest="instructor" />
|
|
291
|
+
<copyField source="crez_course_id_search" dest="course" />
|
|
292
|
+
|
|
293
|
+
<!-- image fields -->
|
|
294
|
+
<copyField source="topic_search" dest="topic_display" />
|
|
295
|
+
<copyField source="subject_other_search" dest="subject_other_display" />
|
|
296
|
+
<copyField source="title_variant_search" dest="title_variant_display" />
|
|
297
|
+
<copyField source="summary_search" dest="summary_display" />
|
|
298
|
+
<copyField source="pub_search" dest="pub_display" />
|
|
299
|
+
|
|
300
|
+
<!-- CJK fields -->
|
|
301
|
+
<!-- both bigrams and unigrams in same field -->
|
|
302
|
+
<copyField source="vern_title_245a_search" dest="cjk_title_245a_search" />
|
|
303
|
+
<copyField source="vern_title_245_search" dest="cjk_title_245_search" />
|
|
304
|
+
<copyField source="vern_title_uniform_search" dest="cjk_title_uniform_search" />
|
|
305
|
+
<copyField source="vern_title_variant_search" dest="cjk_title_variant_search" />
|
|
306
|
+
<copyField source="vern_title_related_search" dest="cjk_title_related_search" />
|
|
307
|
+
<copyField source="vern_series_search" dest="cjk_series_search" />
|
|
308
|
+
<copyField source="vern_author_1xx_search" dest="cjk_author_1xx_search" />
|
|
309
|
+
<copyField source="vern_author_7xx_search" dest="cjk_author_7xx_search" />
|
|
310
|
+
<copyField source="vern_author_8xx_search" dest="cjk_author_8xx_search" />
|
|
311
|
+
<copyField source="vern_topic_search" dest="cjk_topic_search" />
|
|
312
|
+
<copyField source="vern_topic_subx_search" dest="cjk_topic_subx_search" />
|
|
313
|
+
<copyField source="vern_geographic_search" dest="cjk_geographic_search" />
|
|
314
|
+
<copyField source="vern_geographic_subz_search" dest="cjk_geographic_subz_search" />
|
|
315
|
+
<copyField source="vern_subject_other_search" dest="cjk_subject_other_search" />
|
|
316
|
+
<copyField source="vern_subject_other_subvy_search" dest="cjk_subject_other_subvy_search" />
|
|
317
|
+
<copyField source="vern_subject_all_search" dest="cjk_subject_all_search" />
|
|
318
|
+
<copyField source="vern_pub_search" dest="cjk_pub_search" />
|
|
319
|
+
<copyField source="vern_physical" dest="cjk_physical_search" />
|
|
320
|
+
<copyField source="toc_search" dest="cjk_toc_search" /> <!-- we find CJK text in regular 505s -->
|
|
321
|
+
<copyField source="vern_toc_search" dest="cjk_toc_search" />
|
|
322
|
+
<copyField source="vern_context_search" dest="cjk_context_search" />
|
|
323
|
+
<copyField source="summary_search" dest="cjk_summary_search" /> <!-- we find CJK text in regular 520s -->
|
|
324
|
+
<copyField source="vern_summary_search" dest="cjk_summary_search" />
|
|
325
|
+
<copyField source="vern_all_search" dest="cjk_all_search" />
|
|
326
|
+
|
|
327
|
+
<!-- Spotlight fields -->
|
|
328
|
+
<copyField source="id" dest="id_ng" maxChars="3000"/>
|
|
329
|
+
<copyField source="title_full_display" dest="full_title_ng" maxChars="3000"/>
|
|
330
|
+
<copyField source="*_tesim" dest="all_search" />
|
|
331
|
+
<copyField source="*_tesim" dest="all_unstem_search" />
|
|
332
|
+
|
|
333
|
+
<types>
|
|
334
|
+
<fieldType name="string" class="solr.StrField" sortMissingLast="true" />
|
|
335
|
+
<fieldtype name="boolean" class="solr.BoolField" sortMissingLast="true" />
|
|
336
|
+
<fieldtype name="binary" class="solr.BinaryField"/>
|
|
337
|
+
<fieldType name="int" class="solr.TrieIntField" precisionStep="0" positionIncrementGap="0"/>
|
|
338
|
+
<fieldType name="float" class="solr.TrieFloatField" precisionStep="0" positionIncrementGap="0"/>
|
|
339
|
+
<fieldType name="long" class="solr.TrieLongField" precisionStep="0" positionIncrementGap="0"/>
|
|
340
|
+
<fieldType name="double" class="solr.TrieDoubleField" precisionStep="0" positionIncrementGap="0"/>
|
|
341
|
+
<fieldType name="date" class="solr.TrieDateField" precisionStep="0" positionIncrementGap="0"/>
|
|
342
|
+
<!-- t fields are for accelerating range queries -->
|
|
343
|
+
<fieldType name="tint" class="solr.TrieIntField" precisionStep="4" positionIncrementGap="0"/>
|
|
344
|
+
<fieldType name="tfloat" class="solr.TrieFloatField" precisionStep="8" positionIncrementGap="0"/>
|
|
345
|
+
<fieldType name="tlong" class="solr.TrieLongField" precisionStep="8" positionIncrementGap="0"/>
|
|
346
|
+
<fieldType name="tdouble" class="solr.TrieDoubleField" precisionStep="8" positionIncrementGap="0"/>
|
|
347
|
+
<fieldType name="tdate" class="solr.TrieDateField" precisionStep="6" positionIncrementGap="0"/>
|
|
348
|
+
<fieldType name="random" class="solr.RandomSortField" indexed="true" />
|
|
349
|
+
|
|
350
|
+
<!-- Analyzed Text, general case -->
|
|
351
|
+
<fieldtype name="text" class="solr.TextField" positionIncrementGap="10000" autoGeneratePhraseQueries="true">
|
|
352
|
+
<analyzer>
|
|
353
|
+
<tokenizer class="solr.WhitespaceTokenizerFactory" />
|
|
354
|
+
<filter class="solr.ICUFoldingFilterFactory"/> <!-- NFKC, case folding, diacritics removed -->
|
|
355
|
+
<filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="false"/>
|
|
356
|
+
<filter class="solr.WordDelimiterFilterFactory"
|
|
357
|
+
splitOnCaseChange="1" generateWordParts="1" catenateWords="1"
|
|
358
|
+
splitOnNumerics="0" generateNumberParts="1" catenateNumbers="1"
|
|
359
|
+
catenateAll="0" preserveOriginal="0" stemEnglishPossessive="1" />
|
|
360
|
+
<filter class="solr.SnowballPorterFilterFactory" language="English"/>
|
|
361
|
+
<filter class="solr.RemoveDuplicatesTokenFilterFactory" />
|
|
362
|
+
</analyzer>
|
|
363
|
+
</fieldtype>
|
|
364
|
+
|
|
365
|
+
<!-- Analyzed Text, no Stemming -->
|
|
366
|
+
<fieldtype name="textNoStem" class="solr.TextField" positionIncrementGap="10000" autoGeneratePhraseQueries="true">
|
|
367
|
+
<analyzer>
|
|
368
|
+
<tokenizer class="solr.WhitespaceTokenizerFactory" />
|
|
369
|
+
<filter class="solr.ICUFoldingFilterFactory"/> <!-- NFKC, case folding, diacritics removed -->
|
|
370
|
+
<filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="false"/>
|
|
371
|
+
<filter class="solr.WordDelimiterFilterFactory"
|
|
372
|
+
splitOnCaseChange="1" generateWordParts="1" catenateWords="1"
|
|
373
|
+
splitOnNumerics="0" generateNumberParts="1" catenateNumbers="1"
|
|
374
|
+
catenateAll="0" preserveOriginal="0" stemEnglishPossessive="0" />
|
|
375
|
+
</analyzer>
|
|
376
|
+
</fieldtype>
|
|
377
|
+
|
|
378
|
+
<!-- Left and Right Anchored Analyzed Text, no Stemming -->
|
|
379
|
+
<fieldtype name="text_anchored" class="solr.TextField" positionIncrementGap="10000" autoGeneratePhraseQueries="true">
|
|
380
|
+
<analyzer>
|
|
381
|
+
<!-- put beginning and ending anchors on field value, removing trailing chars -->
|
|
382
|
+
<!-- watch out for query time whitespace separated chars that will be processed as their own token stream, e.g. in 'felines : warm and fuzzy' -->
|
|
383
|
+
<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="^\s*(.*[\S&&[^\.\,:;/=<>\(\)\[\]\&\|]])[\s\.\,:;/=<>\(\)\[\]\&\|]*$" replacement="aaaaaa$1zzzzzz"/>
|
|
384
|
+
<tokenizer class="solr.WhitespaceTokenizerFactory" />
|
|
385
|
+
<filter class="solr.ICUFoldingFilterFactory"/> <!-- NFKC, case folding, diacritics removed -->
|
|
386
|
+
<filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="false"/>
|
|
387
|
+
<filter class="solr.SynonymFilterFactory" synonyms="synonyms_both_anchors.txt" ignoreCase="true" expand="true"/>
|
|
388
|
+
<filter class="solr.SynonymFilterFactory" synonyms="synonyms_left_anchor.txt" ignoreCase="true" expand="true"/>
|
|
389
|
+
<filter class="solr.SynonymFilterFactory" synonyms="synonyms_right_anchor.txt" ignoreCase="true" expand="true"/>
|
|
390
|
+
<filter class="solr.WordDelimiterFilterFactory"
|
|
391
|
+
splitOnCaseChange="1" generateWordParts="1" catenateWords="1"
|
|
392
|
+
splitOnNumerics="0" generateNumberParts="1" catenateNumbers="1"
|
|
393
|
+
catenateAll="0" preserveOriginal="0" stemEnglishPossessive="1" />
|
|
394
|
+
</analyzer>
|
|
395
|
+
</fieldtype>
|
|
396
|
+
|
|
397
|
+
<fieldtype name="text_cjk" class="solr.TextField" positionIncrementGap="10000" autoGeneratePhraseQueries="false">
|
|
398
|
+
<analyzer>
|
|
399
|
+
<!-- remove spaces among hangul and han chars if there is at least one hangul char -->
|
|
400
|
+
<!-- a korean char guaranteed at the start of the pattern: pattern="(\p{Hangul}\p{Han}*)\s+(?=[\p{Hangul}\p{Han}])" -->
|
|
401
|
+
<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="([\p{InHangul_Jamo}\p{InHangul_Compatibility_Jamo}\p{InHangul_Syllables}][\p{InBopomofo}\p{InBopomofo_Extended}\p{InCJK_Compatibility}\p{InCJK_Compatibility_Forms}\p{InCJK_Compatibility_Ideographs}\p{InCJK_Compatibility_Ideographs_Supplement}\p{InCJK_Radicals_Supplement}\p{InCJK_Symbols_And_Punctuation}\p{InCJK_Unified_Ideographs}\p{InCJK_Unified_Ideographs_Extension_A}\p{InCJK_Unified_Ideographs_Extension_B}\p{InKangxi_Radicals}\p{InHalfwidth_And_Fullwidth_Forms}\p{InIdeographic_Description_Characters}]*)\s+(?=[\p{InHangul_Jamo}\p{InHangul_Compatibility_Jamo}\p{InHangul_Syllables}\p{InBopomofo}\p{InBopomofo_Extended}\p{InCJK_Compatibility}\p{InCJK_Compatibility_Forms}\p{InCJK_Compatibility_Ideographs}\p{InCJK_Compatibility_Ideographs_Supplement}\p{InCJK_Radicals_Supplement}\p{InCJK_Symbols_And_Punctuation}\p{InCJK_Unified_Ideographs}\p{InCJK_Unified_Ideographs_Extension_A}\p{InCJK_Unified_Ideographs_Extension_B}\p{InKangxi_Radicals}\p{InHalfwidth_And_Fullwidth_Forms}\p{InIdeographic_Description_Characters}])" replacement="$1"/>
|
|
402
|
+
<!-- a korean char guaranteed at the end of the pattern: pattern="([\p{Hangul}\p{Han}])\s+(?=[\p{Han}\s]*\p{Hangul})" -->
|
|
403
|
+
<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="([\p{InHangul_Jamo}\p{InHangul_Compatibility_Jamo}\p{InHangul_Syllables}\p{InBopomofo}\p{InBopomofo_Extended}\p{InCJK_Compatibility}\p{InCJK_Compatibility_Forms}\p{InCJK_Compatibility_Ideographs}\p{InCJK_Compatibility_Ideographs_Supplement}\p{InCJK_Radicals_Supplement}\p{InCJK_Symbols_And_Punctuation}\p{InCJK_Unified_Ideographs}\p{InCJK_Unified_Ideographs_Extension_A}\p{InCJK_Unified_Ideographs_Extension_B}\p{InKangxi_Radicals}\p{InHalfwidth_And_Fullwidth_Forms}\p{InIdeographic_Description_Characters}])\s+(?=[\p{InBopomofo}\p{InBopomofo_Extended}\p{InCJK_Compatibility}\p{InCJK_Compatibility_Forms}\p{InCJK_Compatibility_Ideographs}\p{InCJK_Compatibility_Ideographs_Supplement}\p{InCJK_Radicals_Supplement}\p{InCJK_Symbols_And_Punctuation}\p{InCJK_Unified_Ideographs}\p{InCJK_Unified_Ideographs_Extension_A}\p{InCJK_Unified_Ideographs_Extension_B}\p{InKangxi_Radicals}\p{InHalfwidth_And_Fullwidth_Forms}\p{InIdeographic_Description_Characters}\s]*[\p{InHangul_Jamo}\p{InHangul_Compatibility_Jamo}\p{InHangul_Syllables}])" replacement="$1"/>
|
|
404
|
+
<tokenizer class="solr.ICUTokenizerFactory" />
|
|
405
|
+
<filter class="solr.CJKWidthFilterFactory"/>
|
|
406
|
+
<!--<filter class="edu.stanford.lucene.analysis.CJKFoldingFilterFactory"/>-->
|
|
407
|
+
<filter class="solr.ICUTransformFilterFactory" id="Traditional-Simplified"/>
|
|
408
|
+
<filter class="solr.ICUTransformFilterFactory" id="Katakana-Hiragana"/>
|
|
409
|
+
<filter class="solr.ICUFoldingFilterFactory"/> <!-- NFKC, case folding, diacritics removed -->
|
|
410
|
+
<filter class="solr.CJKBigramFilterFactory" han="true" hiragana="true" katakana="true" hangul="true" outputUnigrams="true" />
|
|
411
|
+
</analyzer>
|
|
412
|
+
</fieldtype>
|
|
413
|
+
|
|
414
|
+
<!-- single token analyzed text, for sorting. Punctuation is significant. -->
|
|
415
|
+
<fieldtype name="alphaSort" class="solr.TextField" sortMissingLast="true" omitNorms="true">
|
|
416
|
+
<analyzer>
|
|
417
|
+
<tokenizer class="solr.KeywordTokenizerFactory" />
|
|
418
|
+
<filter class="solr.ICUFoldingFilterFactory"/>
|
|
419
|
+
<filter class="solr.TrimFilterFactory" />
|
|
420
|
+
</analyzer>
|
|
421
|
+
</fieldtype>
|
|
422
|
+
|
|
423
|
+
<!-- single token with punctuation terms removed so dismax doesn't look for punctuation terms in these fields -->
|
|
424
|
+
<!-- On client side, Lucene query parser breaks things up by whitespace *before* field analysis for dismax -->
|
|
425
|
+
<!-- so punctuation terms (& : ;) are stopwords to allow results from other fields when these chars are surrounded by spaces in query -->
|
|
426
|
+
<!-- do not lowercase -->
|
|
427
|
+
<fieldType name="string_punct_stop" class="solr.TextField" omitNorms="true">
|
|
428
|
+
<analyzer type="index">
|
|
429
|
+
<tokenizer class="solr.KeywordTokenizerFactory" />
|
|
430
|
+
<filter class="solr.ICUNormalizer2FilterFactory" name="nfkc" mode="compose" />
|
|
431
|
+
</analyzer>
|
|
432
|
+
<analyzer type="query">
|
|
433
|
+
<tokenizer class="solr.KeywordTokenizerFactory" />
|
|
434
|
+
<filter class="solr.ICUNormalizer2FilterFactory" name="nfkc" mode="compose" />
|
|
435
|
+
<!-- removing punctuation for Lucene query parser issues -->
|
|
436
|
+
<filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords_punctuation.txt" />
|
|
437
|
+
</analyzer>
|
|
438
|
+
</fieldType>
|
|
439
|
+
|
|
440
|
+
<!-- field designed for LC call number searching -->
|
|
441
|
+
<fieldType name="callnum_ws" class="solr.TextField" omitNorms="true" positionIncrementGap="100" autoGeneratePhraseQueries="true">
|
|
442
|
+
<analyzer type="index">
|
|
443
|
+
<!-- LC: no space between class letters and digits; normalize to " ." before first cutter, no leading space -->
|
|
444
|
+
<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="^ *([A-Za-z]{1,3}) ?(\d{1,4}(\.\d+)?) ?\.?([A-Za-z]\d+)" replacement="$1$2 .$4"/>
|
|
445
|
+
<!-- LC: add space between first cutter letter and its digits to allow matching on first cutter letter only -->
|
|
446
|
+
<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="^([A-Za-z]{1,3}\d{1,4}(\.\d+)? \.([A-Za-z]))(\d+)" replacement="$1 $4"/>
|
|
447
|
+
<!-- prepend yyyy to string so searches can be left anchored -->
|
|
448
|
+
<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="^\s*(\S{1})" replacement="yyyy$1"/>
|
|
449
|
+
<tokenizer class="solr.WhitespaceTokenizerFactory" />
|
|
450
|
+
<filter class="solr.LowerCaseFilterFactory" />
|
|
451
|
+
</analyzer>
|
|
452
|
+
<!-- Note that the query string could be a partial call number, so we can't combine all patterns -->
|
|
453
|
+
<analyzer type="query">
|
|
454
|
+
<!-- LC: no space between class letters and digits, no leading space -->
|
|
455
|
+
<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="^ *([A-Za-z]{1,3}) *(\d{1,4})" replacement="$1$2" />
|
|
456
|
+
<!-- LC: normalize to " ." before first cutter or first letter of cutter (could be preceded by " ." "." " " or nothing) -->
|
|
457
|
+
<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="^([A-Za-z]{1,3}\d{1,4}(\.\d+)?) *\.?([A-Za-z](\d+)?)" replacement="$1 .$3" />
|
|
458
|
+
<!-- LC: add space between first cutter letter and its digits to allow matching on first cutter letter only -->
|
|
459
|
+
<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="^([A-Za-z]{1,3}\d{1,4}(\.\d+)? \.([A-Za-z]))([^ ])" replacement="$1 $4"/>
|
|
460
|
+
<!-- prepend yyyy to string so searches can be left anchored -->
|
|
461
|
+
<charFilter class="solr.PatternReplaceCharFilterFactory" pattern="^\s*(\S{1})" replacement="yyyy$1"/>
|
|
462
|
+
<tokenizer class="solr.WhitespaceTokenizerFactory" />
|
|
463
|
+
<filter class="solr.LowerCaseFilterFactory" />
|
|
464
|
+
</analyzer>
|
|
465
|
+
</fieldType>
|
|
466
|
+
|
|
467
|
+
<!-- for hierarchical facets
|
|
468
|
+
a query for Books/NonFic will match documents indexed with values like
|
|
469
|
+
Books/NonFic, Books/NonFic/Law, Books/NonFic/Science/Physics, etc.
|
|
470
|
+
But it will not match documents indexed with values like Books, or Books/Fic...
|
|
471
|
+
-->
|
|
472
|
+
<fieldType name="string_hierarch" class="solr.TextField" positionIncrementGap="100">
|
|
473
|
+
<analyzer type="index">
|
|
474
|
+
<tokenizer class="solr.PathHierarchyTokenizerFactory" delimiter="|" />
|
|
475
|
+
</analyzer>
|
|
476
|
+
<analyzer type="query">
|
|
477
|
+
<tokenizer class="solr.KeywordTokenizerFactory" />
|
|
478
|
+
</analyzer>
|
|
479
|
+
</fieldType>
|
|
480
|
+
|
|
481
|
+
|
|
482
|
+
<!-- A text field with defaults appropriate for English an NGrams -->
|
|
483
|
+
<fieldType name="text_en_ng" class="solr.TextField" positionIncrementGap="100">
|
|
484
|
+
<analyzer type="index">
|
|
485
|
+
<tokenizer class="solr.ICUTokenizerFactory"/>
|
|
486
|
+
<filter class="solr.ICUFoldingFilterFactory"/> <!-- NFKC, case folding, diacritics removed -->
|
|
487
|
+
<filter class="solr.EnglishPossessiveFilterFactory"/>
|
|
488
|
+
<!-- EnglishMinimalStemFilterFactory is less aggressive than PorterStemFilterFactory: -->
|
|
489
|
+
<filter class="solr.EnglishMinimalStemFilterFactory"/>
|
|
490
|
+
<filter class="solr.TrimFilterFactory"/>
|
|
491
|
+
<filter class="solr.EdgeNGramFilterFactory" minGramSize="3" maxGramSize="15" />
|
|
492
|
+
</analyzer>
|
|
493
|
+
|
|
494
|
+
<analyzer type="index">
|
|
495
|
+
<tokenizer class="solr.ICUTokenizerFactory"/>
|
|
496
|
+
<filter class="solr.ICUFoldingFilterFactory"/> <!-- NFKC, case folding, diacritics removed -->
|
|
497
|
+
<filter class="solr.EnglishPossessiveFilterFactory"/>
|
|
498
|
+
<!-- EnglishMinimalStemFilterFactory is less aggressive than PorterStemFilterFactory: -->
|
|
499
|
+
<filter class="solr.EnglishMinimalStemFilterFactory"/>
|
|
500
|
+
<filter class="solr.TrimFilterFactory"/>
|
|
501
|
+
</analyzer>
|
|
502
|
+
</fieldType>
|
|
503
|
+
|
|
504
|
+
<!-- This point type indexes the coordinates as separate fields (subFields)
|
|
505
|
+
If subFieldType is defined, it references a type, and a dynamic field
|
|
506
|
+
definition is created matching *___<typename>. Alternately, if
|
|
507
|
+
subFieldSuffix is defined, that is used to create the subFields.
|
|
508
|
+
Example: if subFieldType="double", then the coordinates would be
|
|
509
|
+
indexed in fields myloc_0___double,myloc_1___double.
|
|
510
|
+
Example: if subFieldSuffix="_d" then the coordinates would be indexed
|
|
511
|
+
in fields myloc_0_d,myloc_1_d
|
|
512
|
+
The subFields are an implementation detail of the fieldType, and end
|
|
513
|
+
users normally should not need to know about them.
|
|
514
|
+
-->
|
|
515
|
+
<fieldType name="point" class="solr.PointType" dimension="2" subFieldSuffix="_d"/>
|
|
516
|
+
|
|
517
|
+
<!-- A specialized field for geospatial search. If indexed, this fieldType must not be multivalued. -->
|
|
518
|
+
<fieldType name="location" class="solr.LatLonType" subFieldSuffix="_coordinate"/>
|
|
519
|
+
|
|
520
|
+
<!-- An alternative geospatial field type new to Solr 4. It supports multiValued and polygon shapes.
|
|
521
|
+
For more information about this and other Spatial fields new to Solr 4, see:
|
|
522
|
+
http://wiki.apache.org/solr/SolrAdaptersForLuceneSpatial4
|
|
523
|
+
-->
|
|
524
|
+
<fieldType name="location_rpt" class="solr.SpatialRecursivePrefixTreeFieldType"
|
|
525
|
+
geo="true" distErrPct="0.025" maxDistErr="0.000009" units="degrees" />
|
|
526
|
+
|
|
527
|
+
</types>
|
|
528
|
+
|
|
529
|
+
</schema>
|