geohydra 0.3.1 → 0.3.3

Sign up to get free protection for your applications and to get access to all the features.
Files changed (194) hide show
  1. checksums.yaml +7 -0
  2. data/.gitignore +2 -1
  3. data/.travis.yml +1 -2
  4. data/Gemfile +1 -8
  5. data/Gemfile.lock +87 -102
  6. data/README.md +2 -2
  7. data/VERSION +1 -1
  8. data/bin/accession.rb +99 -89
  9. data/bin/assemble.rb +288 -247
  10. data/bin/assemble_data.rb +54 -51
  11. data/bin/assemble_placenames.rb +85 -85
  12. data/bin/build_stage_options.rb +24 -18
  13. data/bin/derive_wgs84.rb +65 -66
  14. data/bin/extract_thumbnail.rb +38 -37
  15. data/bin/geo2mods.rb +78 -0
  16. data/bin/geohydra +14 -5
  17. data/bin/ingest_arcgis.rb +80 -60
  18. data/bin/iso2geo.rb +64 -0
  19. data/bin/loader_postgis.rb +121 -227
  20. data/bin/run_task.rb +23 -0
  21. data/bin/sync_geoserver_metadata.rb +132 -127
  22. data/bin/xsltproc-saxon +6 -0
  23. data/geohydra.gemspec +6 -4
  24. data/lib/geohydra.rb +5 -0
  25. data/lib/geohydra/accession.rb +24 -13
  26. data/lib/geohydra/{arcgis_to_iso19139_fc.xsl → arcgis_to_iso19110.xsl} +0 -0
  27. data/lib/geohydra/gazetteer.csv +842 -36
  28. data/lib/geohydra/gazetteer.rb +48 -24
  29. data/lib/geohydra/mods2geoblacklight.xsl +248 -0
  30. data/lib/geohydra/mods2ogp.xsl +5 -8
  31. data/lib/geohydra/transform.rb +8 -2
  32. data/lib/geohydra/utils.rb +6 -0
  33. data/lib/geohydra/workflow/gisAssemblyWF.rb +109 -0
  34. data/lib/geohydra/workflow/gisAssemblyWF.xml +85 -0
  35. data/lib/geohydra/workflow/gisDeliveryWF.rb +33 -0
  36. data/lib/geohydra/workflow/gisDeliveryWF.xml +36 -0
  37. data/lib/geohydra/workflow/gisDiscoveryWF.rb +55 -0
  38. data/lib/geohydra/workflow/gisDiscoveryWF.xml +28 -0
  39. data/lib/geohydra/workflow/task.rb +82 -0
  40. data/ogp/README.md +350 -0
  41. data/ogp/download.rb +92 -0
  42. data/ogp/fgdc2mods.sh +9 -0
  43. data/ogp/fgdc2mods.xsl +884 -0
  44. data/ogp/ingest.rb +48 -0
  45. data/ogp/select.rb +20 -0
  46. data/ogp/transform.rb +354 -0
  47. data/ogp/validate.rb +182 -0
  48. data/{bin → scripts}/ingest_tufts.rb +0 -0
  49. data/scripts/iso2html/doit.sh +15 -0
  50. data/scripts/iso2html/main.css +66 -0
  51. data/scripts/iso2html/pacioos-iso-html.xsl +1749 -0
  52. data/scripts/iso2html/utils/replace-newlines.xsl +97 -0
  53. data/scripts/iso2html/utils/replace-string.xsl +80 -0
  54. data/scripts/iso2html/utils/strip-digits.xsl +60 -0
  55. data/{bin → scripts}/loader.rb +0 -0
  56. data/scripts/rename_shapefiles.rb +5 -0
  57. data/scripts/render_gazetteer.rb +36 -0
  58. data/{bin → scripts}/seed.rb +0 -0
  59. data/{bin → scripts}/solr_indexer.rb +0 -0
  60. data/scripts/status.csv +253 -0
  61. data/scripts/status.rb +32 -0
  62. data/{bin → scripts}/validate_data.rb +1 -1
  63. data/solr/kurma-app-dev/conf/lang/stopwords_en.txt +34 -0
  64. data/solr/kurma-app-dev/conf/protwords.txt +21 -0
  65. data/solr/kurma-app-dev/conf/schema.xml +156 -0
  66. data/solr/kurma-app-dev/conf/solrconfig.xml +161 -0
  67. data/solr/kurma-app-dev/conf/synonyms.txt +29 -0
  68. data/solr/kurma-app-dev/purge.sh +8 -0
  69. data/solr/kurma-app-test/conf/lang/stopwords_en.txt +34 -0
  70. data/solr/kurma-app-test/conf/protwords.txt +21 -0
  71. data/solr/kurma-app-test/conf/schema.xml +158 -0
  72. data/solr/kurma-app-test/conf/solrconfig.xml +161 -0
  73. data/solr/kurma-app-test/conf/synonyms.txt +29 -0
  74. data/solr/kurma-app-test/deploy.sh +15 -0
  75. data/solr/kurma-app-test/purge.sh +8 -0
  76. data/solr/ogp-dev/purge.sh +1 -2
  77. data/spec/fixtures/bw938nk9584/temp/{TRIPURA-iso19139-fc.xml → TRIPURA-iso19110.xml} +0 -0
  78. data/spec/fixtures/cc142xj8436/temp/{HARYANA-iso19139-fc.xml → HARYANA-iso19110.xml} +0 -0
  79. data/spec/fixtures/cg716wc7949/temp/{metadata.iso19139-fc.xml → metadata.iso19110.xml} +0 -0
  80. data/spec/fixtures/cm007pv9601/temp/{MEGHALAYA-iso19139-fc.xml → MEGHALAYA-iso19110.xml} +0 -0
  81. data/spec/fixtures/cp055nb0189/temp/{metadata.iso19139-fc.xml → metadata.iso19110.xml} +0 -0
  82. data/spec/fixtures/cs838pw3418/temp/{OIL_GAS_FIELDS-iso19139-fc.xml → OIL_GAS_FIELDS-iso19110.xml} +0 -0
  83. data/spec/fixtures/dd308sy5843/temp/{ORISSA-iso19139-fc.xml → ORISSA-iso19110.xml} +0 -0
  84. data/spec/fixtures/dd452vk1873/temp/{metadata.iso19139-fc.xml → metadata.iso19110.xml} +0 -0
  85. data/spec/fixtures/dg850pt1796/temp/{STATE1951-iso19139-fc.xml → STATE1951-iso19110.xml} +0 -0
  86. data/spec/fixtures/dn744tf5427/temp/{DISTRICT1991-iso19139-fc.xml → DISTRICT1991-iso19110.xml} +0 -0
  87. data/spec/fixtures/dq603nz8402/temp/{STATE2001-iso19139-fc.xml → STATE2001-iso19110.xml} +0 -0
  88. data/spec/fixtures/dv609zt4699/temp/{ASSAM-iso19139-fc.xml → ASSAM-iso19110.xml} +0 -0
  89. data/spec/fixtures/dz222hw0585/temp/{PUNJAB-iso19139-fc.xml → PUNJAB-iso19110.xml} +0 -0
  90. data/spec/fixtures/fd673qb9705/temp/{STATE1971-iso19139-fc.xml → STATE1971-iso19110.xml} +0 -0
  91. data/spec/fixtures/fg451wp8917/temp/{SIKKIM-iso19139-fc.xml → SIKKIM-iso19110.xml} +0 -0
  92. data/spec/fixtures/fh247yz0156/temp/{RAJASTHAN-iso19139-fc.xml → RAJASTHAN-iso19110.xml} +0 -0
  93. data/spec/fixtures/fs487vd1465/temp/{CHHATTISGARH-iso19139-fc.xml → CHHATTISGARH-iso19110.xml} +0 -0
  94. data/spec/fixtures/fs591bn3317/temp/{HIMACHAL_PRADESH-iso19139-fc.xml → HIMACHAL_PRADESH-iso19110.xml} +0 -0
  95. data/spec/fixtures/fw920bc5473/temp/{PLSS_TWN-iso19139-fc.xml → PLSS_TWN-iso19110.xml} +0 -0
  96. data/spec/fixtures/gj831wj3625/temp/{metadata.iso19139-fc.xml → metadata.iso19110.xml} +0 -0
  97. data/spec/fixtures/gp075nv3265/temp/{PONDICHERRY-iso19139-fc.xml → PONDICHERRY-iso19110.xml} +0 -0
  98. data/spec/fixtures/gv800hj8141/temp/{BIHAR-iso19139-fc.xml → BIHAR-iso19110.xml} +0 -0
  99. data/spec/fixtures/gw520gz6339/temp/{DADRA_NAGAR_HAVELI_PT-iso19139-fc.xml → DADRA_NAGAR_HAVELI_PT-iso19110.xml} +0 -0
  100. data/spec/fixtures/gy054hz1045/temp/{HARYANA-iso19139-fc.xml → HARYANA-iso19110.xml} +0 -0
  101. data/spec/fixtures/gz352mw6982/temp/{metadata.iso19139-fc.xml → metadata.iso19110.xml} +0 -0
  102. data/spec/fixtures/hb489vm9892/temp/{DISTRICT1981-iso19139-fc.xml → DISTRICT1981-iso19110.xml} +0 -0
  103. data/spec/fixtures/hw125dq0418/temp/{DELHI-iso19139-fc.xml → DELHI-iso19110.xml} +0 -0
  104. data/spec/fixtures/hw892mn4587/temp/{KERALA-iso19139-fc.xml → KERALA-iso19110.xml} +0 -0
  105. data/spec/fixtures/jb371hz3868/temp/{INCOME-iso19139-fc.xml → INCOME-iso19110.xml} +0 -0
  106. data/spec/fixtures/jc017yk9928/temp/{KARNATAKA-iso19139-fc.xml → KARNATAKA-iso19110.xml} +0 -0
  107. data/spec/fixtures/jf841ys4828/temp/{ANDHRA_PRADESH-iso19139-fc.xml → ANDHRA_PRADESH-iso19110.xml} +0 -0
  108. data/spec/fixtures/jh802mp2160/temp/{DELHI_PT-iso19139-fc.xml → DELHI_PT-iso19110.xml} +0 -0
  109. data/spec/fixtures/jj806fc3801/temp/{metadata.iso19139-fc.xml → metadata.iso19110.xml} +0 -0
  110. data/spec/fixtures/jq835yn7161/temp/{HIMACHAL_PRADESH-iso19139-fc.xml → HIMACHAL_PRADESH-iso19110.xml} +0 -0
  111. data/spec/fixtures/jr455pt6676/temp/{TAMILNADU-iso19139-fc.xml → TAMILNADU-iso19110.xml} +0 -0
  112. data/spec/fixtures/js637zp2537/temp/{BIHAR-iso19139-fc.xml → BIHAR-iso19110.xml} +0 -0
  113. data/spec/fixtures/jv502wg9611/temp/{GOA-iso19139-fc.xml → GOA-iso19110.xml} +0 -0
  114. data/spec/fixtures/jw462ck6560/temp/{JAMMU_KASHMIR-iso19139-fc.xml → JAMMU_KASHMIR-iso19110.xml} +0 -0
  115. data/spec/fixtures/kj800fb6273/temp/{STATE2011-iso19139-fc.xml → STATE2011-iso19110.xml} +0 -0
  116. data/spec/fixtures/km504zq3948/temp/{HIMACHAL_PRADESH-iso19139-fc.xml → HIMACHAL_PRADESH-iso19110.xml} +0 -0
  117. data/spec/fixtures/ks297fy1411/temp/{OFFSH_BLOCKS-iso19139-fc.xml → OFFSH_BLOCKS-iso19110.xml} +0 -0
  118. data/spec/fixtures/md358hy5049/temp/{MIZORAM-iso19139-fc.xml → MIZORAM-iso19110.xml} +0 -0
  119. data/spec/fixtures/mg745bq0193/temp/{MADHYA_PRADESH-iso19139-fc.xml → MADHYA_PRADESH-iso19110.xml} +0 -0
  120. data/spec/fixtures/mh187yx3536/temp/{WEST_BENGAL-iso19139-fc.xml → WEST_BENGAL-iso19110.xml} +0 -0
  121. data/spec/fixtures/mk488yn6694/temp/{GUJARAT-iso19139-fc.xml → GUJARAT-iso19110.xml} +0 -0
  122. data/spec/fixtures/my216kp3008/temp/{DELHI-iso19139-fc.xml → DELHI-iso19110.xml} +0 -0
  123. data/spec/fixtures/my504nz9827/temp/{JAMMU_KASHMIR-iso19139-fc.xml → JAMMU_KASHMIR-iso19110.xml} +0 -0
  124. data/spec/fixtures/ng819jm8700/temp/{MANIPUR-iso19139-fc.xml → MANIPUR-iso19110.xml} +0 -0
  125. data/spec/fixtures/np020jq2139/temp/{metadata.iso19139-fc.xml → metadata.iso19110.xml} +0 -0
  126. data/spec/fixtures/ns377mt1608/temp/{STATE1991-iso19139-fc.xml → STATE1991-iso19110.xml} +0 -0
  127. data/spec/fixtures/nw926np8508/temp/{metadata.iso19139-fc.xml → metadata.iso19110.xml} +0 -0
  128. data/spec/fixtures/ny358rm8559/temp/{TRIPURA-iso19139-fc.xml → TRIPURA-iso19110.xml} +0 -0
  129. data/spec/fixtures/nz176rm8192/temp/{DISTRICT2011-iso19139-fc.xml → DISTRICT2011-iso19110.xml} +0 -0
  130. data/spec/fixtures/nz252rq2252/temp/{UTTAR_PRADESH-iso19139-fc.xml → UTTAR_PRADESH-iso19110.xml} +0 -0
  131. data/spec/fixtures/pd902kb3348/temp/{MADHYA_PRADESH-iso19139-fc.xml → MADHYA_PRADESH-iso19110.xml} +0 -0
  132. data/spec/fixtures/pz792fz1776/temp/{MAHARASHTRA-iso19139-fc.xml → MAHARASHTRA-iso19110.xml} +0 -0
  133. data/spec/fixtures/qb767ss4042/temp/{UTTAR_PRADESH-iso19139-fc.xml → UTTAR_PRADESH-iso19110.xml} +0 -0
  134. data/spec/fixtures/qc091qw0570/temp/{GUJARAT-iso19139-fc.xml → GUJARAT-iso19110.xml} +0 -0
  135. data/spec/fixtures/qc652vr7204/temp/{ANDHRA_PRADESH_PT-iso19139-fc.xml → ANDHRA_PRADESH_PT-iso19110.xml} +0 -0
  136. data/spec/fixtures/qk786js7484/temp/{DISTRICT1961-iso19139-fc.xml → DISTRICT1961-iso19110.xml} +0 -0
  137. data/spec/fixtures/qn676pg6767/temp/{GOA-iso19139-fc.xml → GOA-iso19110.xml} +0 -0
  138. data/spec/fixtures/qr255jh4074/temp/{LOKSABHA_14-iso19139-fc.xml → LOKSABHA_14-iso19110.xml} +0 -0
  139. data/spec/fixtures/qr374kj4827/temp/{ASSAM-iso19139-fc.xml → ASSAM-iso19110.xml} +0 -0
  140. data/spec/fixtures/qy162js1748/temp/{CHHATTISGARH-iso19139-fc.xml → CHHATTISGARH-iso19110.xml} +0 -0
  141. data/spec/fixtures/rd446vf2633/temp/{NAGALAND-iso19139-fc.xml → NAGALAND-iso19110.xml} +0 -0
  142. data/spec/fixtures/rf389hf2983/temp/{CHHATTISGARH_PT-iso19139-fc.xml → CHHATTISGARH_PT-iso19110.xml} +0 -0
  143. data/spec/fixtures/rf859ff4582/temp/{JHARKHAND-iso19139-fc.xml → JHARKHAND-iso19110.xml} +0 -0
  144. data/spec/fixtures/rh343ds8931/temp/{BIHAR-iso19139-fc.xml → BIHAR-iso19110.xml} +0 -0
  145. data/spec/fixtures/rn815xk8157/temp/{SIKKIM-iso19139-fc.xml → SIKKIM-iso19110.xml} +0 -0
  146. data/spec/fixtures/rq653sz4470/temp/{CHHATTISGARH-iso19139-fc.xml → CHHATTISGARH-iso19110.xml} +0 -0
  147. data/spec/fixtures/rt625ws6022/temp/{GULF_FAIRWAYS-iso19139-fc.xml → GULF_FAIRWAYS-iso19110.xml} +0 -0
  148. data/spec/fixtures/sc330vf4259/temp/{JHARKHAND-iso19139-fc.xml → JHARKHAND-iso19110.xml} +0 -0
  149. data/spec/fixtures/sq479mx3086/temp/{OFFSH_PLATF-iso19139-fc.xml → OFFSH_PLATF-iso19110.xml} +0 -0
  150. data/spec/fixtures/sr686bm4098/temp/{DAMAN_DIU_PT-iso19139-fc.xml → DAMAN_DIU_PT-iso19110.xml} +0 -0
  151. data/spec/fixtures/sv303sh5583/temp/{ARUNACHAL_PRADESH-iso19139-fc.xml → ARUNACHAL_PRADESH-iso19110.xml} +0 -0
  152. data/spec/fixtures/sy319nh8520/temp/{GUJARAT-iso19139-fc.xml → GUJARAT-iso19110.xml} +0 -0
  153. data/spec/fixtures/td363vx2792/temp/{HIMACHAL_PRADESH_PT-iso19139-fc.xml → HIMACHAL_PRADESH_PT-iso19110.xml} +0 -0
  154. data/spec/fixtures/tf374bd2484/temp/{DISTRICT1951-iso19139-fc.xml → DISTRICT1951-iso19110.xml} +0 -0
  155. data/spec/fixtures/tj797mj7877/temp/{LOKSABHA_15-iso19139-fc.xml → LOKSABHA_15-iso19110.xml} +0 -0
  156. data/spec/fixtures/tv060wq5179/temp/{ASSAM-iso19139-fc.xml → ASSAM-iso19110.xml} +0 -0
  157. data/spec/fixtures/tv536bn1915/temp/{ARUNACHAL_PRADESH-iso19139-fc.xml → ARUNACHAL_PRADESH-iso19110.xml} +0 -0
  158. data/spec/fixtures/tz359cc2977/temp/{MANIPUR-iso19139-fc.xml → MANIPUR-iso19110.xml} +0 -0
  159. data/spec/fixtures/vb525my6511/temp/{UTTARAKHAND-iso19139-fc.xml → UTTARAKHAND-iso19110.xml} +0 -0
  160. data/spec/fixtures/vh802fs4240/temp/{PONDICHERRY-iso19139-fc.xml → PONDICHERRY-iso19110.xml} +0 -0
  161. data/spec/fixtures/vk120xn2474/temp/{PLSS_SEC-iso19139-fc.xml → PLSS_SEC-iso19110.xml} +0 -0
  162. data/spec/fixtures/vn439bc7316/temp/{KERALA-iso19139-fc.xml → KERALA-iso19110.xml} +0 -0
  163. data/spec/fixtures/vq745jk0695/temp/{MEGHALAYA-iso19139-fc.xml → MEGHALAYA-iso19110.xml} +0 -0
  164. data/spec/fixtures/vr593vj7147/temp/{ANDHRA_PRADESH-iso19139-fc.xml → ANDHRA_PRADESH-iso19110.xml} +0 -0
  165. data/spec/fixtures/vw911qb5271/temp/{DISTRICT2001-iso19139-fc.xml → DISTRICT2001-iso19110.xml} +0 -0
  166. data/spec/fixtures/wg680pz0365/temp/{ANDHRA_PRADESH-iso19139-fc.xml → ANDHRA_PRADESH-iso19110.xml} +0 -0
  167. data/spec/fixtures/wg761xn1926/temp/{HARYANA-iso19139-fc.xml → HARYANA-iso19110.xml} +0 -0
  168. data/spec/fixtures/wh870qw1934/temp/{PUNJAB-iso19139-fc.xml → PUNJAB-iso19110.xml} +0 -0
  169. data/spec/fixtures/wk775mm4673/temp/{MAHARASHTRA-iso19139-fc.xml → MAHARASHTRA-iso19110.xml} +0 -0
  170. data/spec/fixtures/ws171yz2165/temp/{ARUNACHAL_PRADESH_PT-iso19139-fc.xml → ARUNACHAL_PRADESH_PT-iso19110.xml} +0 -0
  171. data/spec/fixtures/wt473hz7153/temp/{CHANDIGARH_PT-iso19139-fc.xml → CHANDIGARH_PT-iso19110.xml} +0 -0
  172. data/spec/fixtures/ww217dj0457/temp/{CO2_PIPE-iso19139-fc.xml → CO2_PIPE-iso19110.xml} +0 -0
  173. data/spec/fixtures/wy875pk9849/temp/{STATE1961-iso19139-fc.xml → STATE1961-iso19110.xml} +0 -0
  174. data/spec/fixtures/xb018tk2042/temp/{STATE1981-iso19139-fc.xml → STATE1981-iso19110.xml} +0 -0
  175. data/spec/fixtures/xg539vw8586/temp/{ORISSA-iso19139-fc.xml → ORISSA-iso19110.xml} +0 -0
  176. data/spec/fixtures/xv475kp4644/temp/{ASSAM_PT-iso19139-fc.xml → ASSAM_PT-iso19110.xml} +0 -0
  177. data/spec/fixtures/xy096gc2959/temp/{GOA-iso19139-fc.xml → GOA-iso19110.xml} +0 -0
  178. data/spec/fixtures/xz518gz3362/temp/{UTTARAKHAND-iso19139-fc.xml → UTTARAKHAND-iso19110.xml} +0 -0
  179. data/spec/fixtures/yh986wy4737/temp/{NAGALAND-iso19139-fc.xml → NAGALAND-iso19110.xml} +0 -0
  180. data/spec/fixtures/yn187fq4474/temp/{KARNATAKA-iso19139-fc.xml → KARNATAKA-iso19110.xml} +0 -0
  181. data/spec/fixtures/yn236mw3250/temp/{TAMILNADU-iso19139-fc.xml → TAMILNADU-iso19110.xml} +0 -0
  182. data/spec/fixtures/yz596nz0112/temp/{WEST_BENGAL-iso19139-fc.xml → WEST_BENGAL-iso19110.xml} +0 -0
  183. data/spec/fixtures/zk596gy7380/temp/{DISTRICT1971-iso19139-fc.xml → DISTRICT1971-iso19110.xml} +0 -0
  184. data/spec/fixtures/zn452hh7431/temp/{RAJASTHAN-iso19139-fc.xml → RAJASTHAN-iso19110.xml} +0 -0
  185. data/spec/fixtures/zt093fw6519/temp/{MIZORAM-iso19139-fc.xml → MIZORAM-iso19110.xml} +0 -0
  186. data/spec/fixtures/zv925hd6723/temp/{OGWELLS-iso19139-fc.xml → OGWELLS-iso19110.xml} +0 -0
  187. data/spec/fixtures/zy658cr1728/temp/{ANDAMAAN_NICOBAR_PT-iso19139-fc.xml → ANDAMAAN_NICOBAR_PT-iso19110.xml} +0 -0
  188. data/spec/fixtures/zz943vx1492/temp/{BASINS-iso19139-fc.xml → BASINS-iso19110.xml} +0 -0
  189. data/spec/unit/gazetteer_spec.rb +100 -35
  190. data/spec/unit/task_spec.rb +68 -0
  191. data/spec/unit/transform_spec.rb +1 -1
  192. data/spec/unit/utils_spec.rb +17 -3
  193. data/workflow.rb +35 -0
  194. metadata +323 -316
@@ -0,0 +1,8 @@
1
+ #!/bin/bash -x
2
+ c=kurma-app-dev
3
+ h=localhost:8983
4
+ s=http
5
+
6
+ curl "$s://${h}/solr/${c}/update" --data '<delete><query>*:*</query></delete>' -H 'Content-type:text/xml; charset=utf-8'
7
+ curl "$s://${h}/solr/${c}/update?commit=true"
8
+
@@ -0,0 +1,34 @@
1
+ # Standard english stop words taken from Lucene's StopAnalyzer
2
+ a
3
+ an
4
+ and
5
+ are
6
+ as
7
+ at
8
+ be
9
+ but
10
+ by
11
+ for
12
+ if
13
+ in
14
+ into
15
+ is
16
+ it
17
+ no
18
+ not
19
+ of
20
+ on
21
+ or
22
+ such
23
+ that
24
+ the
25
+ their
26
+ then
27
+ there
28
+ these
29
+ they
30
+ this
31
+ to
32
+ was
33
+ will
34
+ with
@@ -0,0 +1,21 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ #-----------------------------------------------------------------------
14
+ # Use a protected word file to protect against the stemmer reducing two
15
+ # unrelated words to the same base word.
16
+
17
+ # Some non-words that normally won't be encountered,
18
+ # just to test that they won't be stemmed.
19
+ dontstems
20
+ zwhacky
21
+
@@ -0,0 +1,158 @@
1
+ <?xml version="1.0" encoding="UTF-8"?>
2
+ <schema name="kurma-app-dev" version="1.5">
3
+ <uniqueKey>uuid</uniqueKey>
4
+ <fields>
5
+ <field name="_version_" type="long" stored="true" indexed="true"/>
6
+ <field name="timestamp" type="date" stored="true" indexed="true" default="NOW"/>
7
+ <field name="uuid" type="string" stored="true" indexed="true" required="true"/>
8
+
9
+ <!-- core generated fields -->
10
+ <field name="text" type="text_en" stored="false" indexed="true" multiValued="true"
11
+ termVectors="true" termPositions="true" termOffsets="true" />
12
+
13
+ <!-- dynamic field with simple types by suffix -->
14
+ <dynamicField name="*_b" type="boolean" stored="true" indexed="true"/>
15
+ <dynamicField name="*_d" type="double" stored="true" indexed="true"/>
16
+ <dynamicField name="*_dt" type="date" stored="true" indexed="true"/>
17
+ <dynamicField name="*_f" type="float" stored="true" indexed="true"/>
18
+ <dynamicField name="*_i" type="int" stored="true" indexed="true"/>
19
+ <dynamicField name="*_l" type="long" stored="true" indexed="true"/>
20
+ <dynamicField name="*_s" type="string" stored="true" indexed="true"/>
21
+ <dynamicField name="*_ss" type="string" stored="true" indexed="false"/>
22
+ <dynamicField name="*_si" type="string" stored="false" indexed="true"/>
23
+ <dynamicField name="*_sim" type="string" stored="false" indexed="true" multiValued="true" />
24
+ <dynamicField name="*_sm" type="string" stored="true" indexed="true" multiValued="true" />
25
+ <dynamicField name="*_url" type="string" stored="true" indexed="false"/>
26
+ <dynamicField name="*_blob" type="binary" stored="true" indexed="false"/>
27
+
28
+ <!-- dynamic Text fields by suffix without storage -->
29
+ <dynamicField name="*_t" type="text_en" stored="false" indexed="true"
30
+ termVectors="true" termPositions="true" termOffsets="true"/>
31
+ <dynamicField name="*_tm" type="text_en" stored="false" indexed="true" multiValued="true"
32
+ termVectors="true" termPositions="true" termOffsets="true"/>
33
+ <dynamicField name="*_ti" type="text_en" stored="false" indexed="true"
34
+ termVectors="true" termPositions="true" termOffsets="true"/>
35
+ <dynamicField name="*_tmi" type="text_en" stored="false" indexed="true" multiValued="true"
36
+ termVectors="true" termPositions="true" termOffsets="true"/>
37
+ <dynamicField name="*_sort" type="text_sort" stored="false" indexed="true" multiValued="false"/>
38
+
39
+ <!-- Spatial field types:
40
+
41
+ Solr3:
42
+ <field name="my_pt">83.1,-117.312</field>
43
+ as (y,x)
44
+
45
+ Solr4:
46
+
47
+ <field name="my_bbox">-117.312 83.1 -115.39 84.31</field>
48
+ as (W S E N)
49
+
50
+ <field name="my_geom">POLYGON((1 8, 1 9, 2 9, 2 8, 1 8))</field>
51
+ as WKT for point, linestring, polygon
52
+
53
+ -->
54
+ <dynamicField name="*_pt" type="location" stored="true" indexed="true"/>
55
+ <dynamicField name="*_bbox" type="location_rpt" stored="true" indexed="true"/>
56
+ <dynamicField name="*_geom" type="string" stored="true" indexed="true"/> <!-- XXX: no JTS -->
57
+ </fields>
58
+
59
+ <types>
60
+ <fieldType name="string" class="solr.StrField" sortMissingLast="true" />
61
+ <fieldType name="boolean" class="solr.BoolField" sortMissingLast="true"/>
62
+
63
+ <fieldType name="int" class="solr.TrieIntField" precisionStep="8" positionIncrementGap="0"/>
64
+ <fieldType name="float" class="solr.TrieFloatField" precisionStep="8" positionIncrementGap="0"/>
65
+ <fieldType name="long" class="solr.TrieLongField" precisionStep="8" positionIncrementGap="0"/>
66
+ <fieldType name="double" class="solr.TrieDoubleField" precisionStep="8" positionIncrementGap="0"/>
67
+
68
+ <!-- The format for this date field is of the form 1995-12-31T23:59:59Z.
69
+ The trailing "Z" designates UTC time and is mandatory.
70
+ A Trie based date field for faster date range queries and date faceting. -->
71
+ <fieldType name="date" class="solr.TrieDateField" precisionStep="6" positionIncrementGap="0"/>
72
+
73
+ <!--Binary data type. The data should be sent/retrieved in as Base64 encoded Strings -->
74
+ <fieldtype name="binary" class="solr.BinaryField"/>
75
+
76
+ <!-- A text field with defaults appropriate for English: it
77
+ tokenizes with StandardTokenizer, removes English stop words
78
+ (lang/stopwords_en.txt), down cases, protects words from protwords.txt, and
79
+ finally applies Porter's stemming. The query time analyzer
80
+ also applies synonyms from synonyms.txt. -->
81
+ <fieldType name="text_en" class="solr.TextField" positionIncrementGap="100">
82
+ <analyzer type="index">
83
+ <tokenizer class="solr.StandardTokenizerFactory"/>
84
+ <filter class="solr.StopFilterFactory" ignoreCase="true" words="lang/stopwords_en.txt"/>
85
+ <filter class="solr.LowerCaseFilterFactory"/>
86
+ <filter class="solr.EnglishPossessiveFilterFactory"/>
87
+ <filter class="solr.KeywordMarkerFilterFactory" protected="protwords.txt"/>
88
+ <filter class="solr.PorterStemFilterFactory"/>
89
+ </analyzer>
90
+ <analyzer type="query">
91
+ <tokenizer class="solr.StandardTokenizerFactory"/>
92
+ <filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
93
+ <filter class="solr.StopFilterFactory" ignoreCase="true" words="lang/stopwords_en.txt"/>
94
+ <filter class="solr.LowerCaseFilterFactory"/>
95
+ <filter class="solr.EnglishPossessiveFilterFactory"/>
96
+ <filter class="solr.KeywordMarkerFilterFactory" protected="protwords.txt"/>
97
+ <filter class="solr.PorterStemFilterFactory"/>
98
+ </analyzer>
99
+ </fieldType>
100
+
101
+ <!-- for alpha sorting as a single token -->
102
+ <fieldType name="text_sort" class="solr.TextField" sortMissingLast="true" omitNorms="true">
103
+ <analyzer>
104
+ <tokenizer class="solr.KeywordTokenizerFactory"/>
105
+ <filter class="solr.LowerCaseFilterFactory" />
106
+ <filter class="solr.TrimFilterFactory" />
107
+ <filter class="solr.PatternReplaceFilterFactory" pattern="([^a-z0-9 ])" replacement="" replace="all"/>
108
+ </analyzer>
109
+ </fieldType>
110
+
111
+ <!-- Spatial field types -->
112
+ <fieldType name="location" class="solr.LatLonType" subFieldSuffix="_d"/>
113
+
114
+ <fieldType name="location_rpt" class="solr.SpatialRecursivePrefixTreeFieldType"
115
+ distErrPct="0.025"
116
+ maxDistErr="0.000009"
117
+ units="degrees"
118
+ />
119
+
120
+ <!-- JTS-enabled spatial predicates; requires JTS installation -->
121
+ <!-- XXX: no JTS
122
+ <fieldType name="location_jts" class="solr.SpatialRecursivePrefixTreeFieldType"
123
+ spatialContextFactory="com.spatial4j.core.context.jts.JtsSpatialContextFactory"
124
+ distErrPct="0.025"
125
+ maxDistErr="0.000009"
126
+ units="degrees"
127
+ />
128
+ -->
129
+ </types>
130
+
131
+ <!-- for scoring formula -->
132
+ <copyField source="dc_coverage_spatial_sm" dest="dc_coverage_spatial_tmi" maxChars="10000"/>
133
+ <copyField source="dc_coverage_temporal_sm" dest="dc_coverage_temporal_tmi" maxChars="10000"/>
134
+ <copyField source="dc_creator_s" dest="dc_creator_ti" maxChars="1000"/>
135
+ <copyField source="dc_description_s" dest="dc_description_ti" maxChars="10000"/>
136
+ <copyField source="dc_format_s" dest="dc_format_ti" maxChars="100"/>
137
+ <copyField source="dc_identifier_s" dest="dc_identifier_ti" maxChars="100"/>
138
+ <copyField source="dc_publisher_s" dest="dc_publisher_ti" maxChars="1000"/>
139
+ <copyField source="dc_rights_s" dest="dc_rights_ti" maxChars="100"/>
140
+ <copyField source="dc_source_s" dest="dc_source_ti" maxChars="100"/>
141
+ <copyField source="dc_subject_sm" dest="dc_subject_tmi" maxChars="10000"/>
142
+ <copyField source="dc_title_s" dest="dc_title_ti" maxChars="1000"/>
143
+ <copyField source="layer_collection_s" dest="layer_collection_ti" maxChars="1000"/>
144
+ <copyField source="layer_geom_type_s" dest="layer_geom_type_ti" maxChars="100"/>
145
+ <copyField source="layer_slug_s" dest="layer_slug_ti" maxChars="100"/>
146
+
147
+ <!-- core text search -->
148
+ <copyField source="*_ti" dest="text" />
149
+ <copyField source="*_tmi" dest="text" />
150
+
151
+ <!-- for sorting text fields -->
152
+ <copyField source="dc_title_s" dest="dc_title_sort"/>
153
+ <copyField source="dc_creator_s" dest="dc_creator_sort"/>
154
+ <copyField source="dc_publisher_s" dest="dc_publisher_sort"/>
155
+ <copyField source="layer_collection_s" dest="layer_collection_sort"/>
156
+
157
+ </schema>
158
+
@@ -0,0 +1,161 @@
1
+ <?xml version="1.0" encoding="UTF-8"?>
2
+ <!--
3
+ Licensed to the Apache Software Foundation (ASF) under one or more
4
+ contributor license agreements. See the NOTICE file distributed with
5
+ this work for additional information regarding copyright ownership.
6
+ The ASF licenses this file to You under the Apache License, Version 2.0
7
+ (the "License"); you may not use this file except in compliance with
8
+ the License. You may obtain a copy of the License at
9
+
10
+ http://www.apache.org/licenses/LICENSE-2.0
11
+
12
+ Unless required by applicable law or agreed to in writing, software
13
+ distributed under the License is distributed on an "AS IS" BASIS,
14
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15
+ See the License for the specific language governing permissions and
16
+ limitations under the License.
17
+ -->
18
+ <!--
19
+ For more details about configurations options that may appear in
20
+ this file, see http://wiki.apache.org/solr/SolrConfigXml.
21
+ -->
22
+ <config>
23
+ <luceneMatchVersion>LUCENE_CURRENT</luceneMatchVersion>
24
+ <dataDir>${solr.data.dir:}</dataDir>
25
+ <directoryFactory name="DirectoryFactory" class="${solr.directoryFactory:solr.NRTCachingDirectoryFactory}"/>
26
+ <codecFactory class="solr.SchemaCodecFactory"/>
27
+ <schemaFactory class="ClassicIndexSchemaFactory"/>
28
+ <indexConfig>
29
+ <lockType>${solr.lock.type:native}</lockType>
30
+ </indexConfig>
31
+
32
+ <!-- The default high-performance update handler -->
33
+ <updateHandler class="solr.DirectUpdateHandler2">
34
+ <updateLog>
35
+ <str name="dir">${solr.ulog.dir:}</str>
36
+ </updateLog>
37
+ <autoCommit>
38
+ <maxTime>15000</maxTime>
39
+ <openSearcher>false</openSearcher>
40
+ </autoCommit>
41
+ </updateHandler>
42
+
43
+ <!-- realtime get handler, guaranteed to return the latest stored fields
44
+ of any document, without the need to commit or open a new searcher. The current
45
+ implementation relies on the updateLog feature being enabled. -->
46
+ <requestHandler name="/get" class="solr.RealTimeGetHandler">
47
+ <lst name="defaults">
48
+ <str name="omitHeader">true</str>
49
+ </lst>
50
+ </requestHandler>
51
+
52
+ <requestHandler name="/replication" class="solr.ReplicationHandler" startup="lazy" />
53
+
54
+ <!-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
55
+ Query section - these settings control query time things like caches
56
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -->
57
+ <query>
58
+ <maxBooleanClauses>1024</maxBooleanClauses>
59
+ <filterCache class="solr.FastLRUCache" size="512" initialSize="512" autowarmCount="0"/>
60
+ <queryResultCache class="solr.LRUCache" size="512" initialSize="512" autowarmCount="0"/>
61
+ <documentCache class="solr.LRUCache" size="512" initialSize="512" autowarmCount="0"/>
62
+ <enableLazyFieldLoading>true</enableLazyFieldLoading>
63
+ <queryResultWindowSize>20</queryResultWindowSize>
64
+ <queryResultMaxDocsCached>200</queryResultMaxDocsCached>
65
+ <listener event="newSearcher" class="solr.QuerySenderListener">
66
+ <arr name="queries">
67
+ <lst><str name="q">stanford</str></lst>
68
+ <lst><str name="q">polygon</str></lst>
69
+ </arr>
70
+ </listener>
71
+ <listener event="firstSearcher" class="solr.QuerySenderListener">
72
+ <arr name="queries">
73
+ <lst>
74
+ <str name="q">static firstSearcher warming in solrconfig.xml</str>
75
+ </lst>
76
+ </arr>
77
+ </listener>
78
+ <useColdSearcher>false</useColdSearcher>
79
+ <maxWarmingSearchers>2</maxWarmingSearchers>
80
+ </query>
81
+
82
+ <requestDispatcher handleSelect="false">
83
+ <requestParsers enableRemoteStreaming="true" multipartUploadLimitInKB="2048000" formdataUploadLimitInKB="2048"/>
84
+ <httpCaching never304="true"/>
85
+ </requestDispatcher>
86
+
87
+ <requestHandler name="/search" class="solr.SearchHandler"/>
88
+
89
+ <requestHandler name="/select" class="solr.SearchHandler">
90
+ <lst name="defaults">
91
+ <int name="start">0</int>
92
+ <int name="rows">10</int>
93
+ <str name="wt">json</str>
94
+ <int name="indent">2</int>
95
+ <str name="defType">edismax</str>
96
+ <str name="echoParams">all</str>
97
+ <str name="fl">*,score</str>
98
+ <str name="sort">score desc, dc_title_sort asc</str>
99
+ <str name="q.alt">*:*</str>
100
+ <str name="qf">
101
+ text^1
102
+ dc_description_ti^2
103
+ dc_creator_ti^3
104
+ dc_publisher_ti^3
105
+ layer_collection_ti^4
106
+ dc_subject_tmi^5
107
+ dc_coverage_spatial_tmi^5
108
+ dc_coverage_temporal_tmi^5
109
+ dc_title_ti^6
110
+ dc_rights_ti^7
111
+ dc_source_ti^8
112
+ layer_geom_type_ti^9
113
+ layer_slug_ti^10
114
+ dc_identifier_ti^10
115
+ </str>
116
+ <bool name="facet">true</bool>
117
+ <int name="facet.mincount">1</int>
118
+ <int name="facet.limit">10</int>
119
+ <str name="facet.field">dc_coverage_spatial_sm</str>
120
+ <str name="facet.field">dc_creator_s</str>
121
+ <str name="facet.field">dc_format_s</str>
122
+ <str name="facet.field">dc_language_s</str>
123
+ <str name="facet.field">dc_publisher_s</str>
124
+ <str name="facet.field">dc_rights_s</str>
125
+ <str name="facet.field">dc_source_s</str>
126
+ <str name="facet.field">dc_subject_sm</str>
127
+ <str name="facet.field">layer_collection_s</str>
128
+ <str name="facet.field">layer_geom_type_s</str>
129
+ <str name="facet.field">layer_srs_s</str>
130
+ <str name="facet.field">layer_year_i</str>
131
+ </lst>
132
+ </requestHandler>
133
+
134
+ <requestHandler name="/update" class="solr.UpdateRequestHandler"/>
135
+ <requestHandler name="/admin/" class="solr.admin.AdminHandlers"/>
136
+
137
+ <!-- ping/healthcheck -->
138
+ <requestHandler name="/admin/ping" class="solr.PingRequestHandler">
139
+ <lst name="invariants">
140
+ <str name="q">solrpingquery</str>
141
+ </lst>
142
+ <lst name="defaults">
143
+ <str name="echoParams">all</str>
144
+ </lst>
145
+ <!-- An optional feature of the PingRequestHandler is to configure the
146
+ handler with a "healthcheckFile" which can be used to enable/disable
147
+ the PingRequestHandler.
148
+ relative paths are resolved against the data dir
149
+ -->
150
+ <str name="healthcheckFile">server-enabled.txt</str>
151
+ </requestHandler>
152
+
153
+ <requestHandler name="/analysis/field"
154
+ startup="lazy"
155
+ class="solr.FieldAnalysisRequestHandler" />
156
+
157
+ <!-- Legacy config for the admin interface -->
158
+ <admin>
159
+ <defaultQuery>*:*</defaultQuery>
160
+ </admin>
161
+ </config>
@@ -0,0 +1,29 @@
1
+ # The ASF licenses this file to You under the Apache License, Version 2.0
2
+ # (the "License"); you may not use this file except in compliance with
3
+ # the License. You may obtain a copy of the License at
4
+ #
5
+ # http://www.apache.org/licenses/LICENSE-2.0
6
+ #
7
+ # Unless required by applicable law or agreed to in writing, software
8
+ # distributed under the License is distributed on an "AS IS" BASIS,
9
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
+ # See the License for the specific language governing permissions and
11
+ # limitations under the License.
12
+
13
+ #-----------------------------------------------------------------------
14
+ #some test synonym mappings unlikely to appear in real input text
15
+ aaafoo => aaabar
16
+ bbbfoo => bbbfoo bbbbar
17
+ cccfoo => cccbar cccbaz
18
+ fooaaa,baraaa,bazaaa
19
+
20
+ # Some synonym groups specific to this example
21
+ GB,gib,gigabyte,gigabytes
22
+ MB,mib,megabyte,megabytes
23
+ Television, Televisions, TV, TVs
24
+ #notice we use "gib" instead of "GiB" so any WordDelimiterFilter coming
25
+ #after us won't split it into two words.
26
+
27
+ # Synonym mappings can be used for spelling correction too
28
+ pixima => pixma
29
+
@@ -0,0 +1,15 @@
1
+ #!/bin/bash -x
2
+ export PATH=/usr/bin:/bin
3
+ c=kurma-app-test
4
+
5
+ if [ "$(hostname)" != 'sul-solr-a.stanford.edu' ]; then
6
+ echo "Wrong host"
7
+ exit -1
8
+ fi
9
+
10
+ cd /home/drh/geohydra/solr/$c/conf || exit -1
11
+ java -cp "/usr/share/tomcat6/webapps/solr/WEB-INF/lib/*:/usr/share/tomcat6/lib/*" \
12
+ org.apache.solr.cloud.ZkCLI -cmd upconfig -zkhost 127.0.0.1:2181 \
13
+ -confdir . -confname $c \
14
+ -solrhome /home/lyberadmin/solr-home
15
+
@@ -0,0 +1,8 @@
1
+ #!/bin/bash -x
2
+ c=kurma-app-test
3
+ h=localhost:8080
4
+ s=http
5
+
6
+ curl "$s://${h}/solr/${c}/update" --data '<delete><query>*:*</query></delete>' -H 'Content-type:text/xml; charset=utf-8'
7
+ curl "$s://${h}/solr/${c}/update?commit=true"
8
+
@@ -1,7 +1,6 @@
1
1
  #!/bin/bash -x
2
2
  c=ogp-dev
3
- d=/var/geomdtk/current/workspace
4
- h=localhost:28080
3
+ h=localhost:18080
5
4
  s=http
6
5
 
7
6
  curl "$s://${h}/solr/${c}/update" --data '<delete><query>*:*</query></delete>' -H 'Content-type:text/xml; charset=utf-8'