dor-services 5.3.4 → 5.4.0

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 CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA1:
3
- metadata.gz: 2dd35a675cb046c99db17934a58e4587b9caa03e
4
- data.tar.gz: fcf7860c66176b880be68945ab882ff761abe815
3
+ metadata.gz: d9a88f0d0bd7ef08c38ddb485d4207398f9c9d82
4
+ data.tar.gz: 0908b1eaf6a8b92e1f3e66c8fb91ae0f38d5fcfe
5
5
  SHA512:
6
- metadata.gz: 4283ad61ea3c41740204253a88e8b45be6546b8420c2048787f3b42eb44b6871acce199bbb5263a138c36b776dcf91e105de0bc97fb7b456b50e65980cb028b7
7
- data.tar.gz: bbb4592f2cdf234ab5b981f38a74e6785574e07a3353325b9bf31437d20368eb3d041a81b6eac0cf7f2d31e4aa04d5cdd611a9315f4ebc956a2de429e2393e64
6
+ metadata.gz: 899c1855ff906efcac21cccf89af369e88faaac71ac6b53e21ee2dd224252331d903135cd61d30ecdaba32ca138baf28b4b2f17f213a858b7235eb7256851bf7
7
+ data.tar.gz: d2a357095da5f7e320844f3f7f069e62d8f616a38d68229a4128d0ba1f1f093aaa8b91d1b9b3f9ba4da15a43f53646905b20997bac67309710090c44d12359eb
@@ -16,12 +16,9 @@
16
16
  :password: ""
17
17
  :port: 61613
18
18
  :client_id: 'dor-services'
19
- :solrizer:
19
+ :solr:
20
20
  :url:
21
21
  :opts: {}
22
- :gsearch:
23
- :rest_url:
24
- :url:
25
22
  :metadata:
26
23
  :catalog:
27
24
  :url:
@@ -37,8 +37,7 @@ Dor::Config.configure do
37
37
  local_document_cache_root '/purl'
38
38
  end
39
39
 
40
- gsearch.url 'https://host/solr/gsearch'
41
- solrizer.url 'https://host/solr'
40
+ solr.url 'https://host/solr'
42
41
  workflow.url 'https://host/workflow/'
43
42
  dor_services.url 'https://host/dor/v1'
44
43
 
data/lib/dor/config.rb CHANGED
@@ -55,8 +55,10 @@ module Dor
55
55
  end
56
56
 
57
57
  def make_solr_connection(add_opts = {})
58
- opts = Config.solrizer.opts.merge(add_opts).merge(
59
- :url => Config.solrizer.url
58
+ c = Dor::Config.solr || Dor::Config.solrizer
59
+
60
+ opts = c.opts.merge(add_opts).merge(
61
+ :url => c.url
60
62
  )
61
63
  ::RSolr::Ext.connect(opts)
62
64
  end
@@ -75,12 +77,11 @@ module Dor
75
77
  end
76
78
  }
77
79
  },
78
- :sdr => {
79
- :rest_client => Confstruct.deferred { |c| config.make_rest_client c.url }
80
+ :dor_services => {
81
+ :rest_client => Confstruct.deferred { |c| config.make_rest_client c.url, c.cert_file, c.key_file, c.key_pass }
80
82
  },
81
- :gsearch => {
82
- :rest_client => Confstruct.deferred { |c| config.make_rest_client c.rest_url },
83
- :client => Confstruct.deferred { |c| config.make_rest_client c.url }
83
+ :sdr => {
84
+ :rest_client => Confstruct.deferred { |c| config.make_rest_client c.url, c.cert_file, c.key_file, c.key_pass }
84
85
  },
85
86
  :stomp => {
86
87
  :connection => Confstruct.deferred { |c| Stomp::Connection.new c.user, c.password, c.host, c.port, true, 5, { 'client-id' => c.client_id }},
@@ -102,26 +103,17 @@ module Dor
102
103
  config.fedora.delete(key)
103
104
  end
104
105
 
105
- if ActiveFedora.respond_to?(:configurator)
106
- if config.solrizer.url.present?
107
- ActiveFedora::SolrService.register
108
- ActiveFedora::SolrService.instance.instance_variable_set :@conn, make_solr_connection
109
- end
110
- else
111
- ActiveFedora::RubydoraConnection.connect fedora_config if fedora.url.present?
112
- if solrizer.url.present?
113
- ActiveFedora::SolrService.register config.solrizer.url, config.solrizer.opts
114
- conn = ActiveFedora::SolrService.instance.conn.connection
115
- if config.ssl.cert_file.present?
116
- conn.use_ssl = true
117
- conn.cert = OpenSSL::X509::Certificate.new(File.read(config.ssl.cert_file))
118
- conn.key = OpenSSL::PKey::RSA.new(File.read(config.ssl.key_file), config.ssl.key_pass) if config.ssl.key_file.present?
119
- conn.verify_mode = OpenSSL::SSL::VERIFY_NONE
120
- end
121
- end
122
- ActiveFedora.init
123
- ActiveFedora.fedora_config_path = File.expand_path('../../../config/dummy.yml', __FILE__)
106
+ if config.solrizer.present?
107
+ stack = Kernel.caller.dup
108
+ stack.shift while stack[0] =~ %r{(active_support/callbacks|dor/config|dor-services)\.rb}
109
+ ActiveSupport::Deprecation.warn "Dor::Config -- solrizer configuration is deprecated. Please use solr instead.", stack
124
110
  end
111
+
112
+ if (config.solr || config.solrizer).url.present?
113
+ ActiveFedora::SolrService.register
114
+ ActiveFedora::SolrService.instance.instance_variable_set :@conn, make_solr_connection
115
+ end
116
+
125
117
  end
126
118
 
127
119
  # Act like an ActiveFedora.configurator
@@ -137,7 +129,7 @@ module Dor
137
129
  end
138
130
 
139
131
  def solr_config
140
- { :url => solrizer.url }
132
+ { :url => (solr || solrizer).url }
141
133
  end
142
134
 
143
135
  def predicate_config
@@ -146,5 +138,5 @@ module Dor
146
138
  end
147
139
 
148
140
  Config = Configuration.new(YAML.load(File.read(File.expand_path('../../../config/config_defaults.yml', __FILE__))))
149
- ActiveFedora.configurator = Config if ActiveFedora.respond_to?(:configurator)
141
+ ActiveFedora.configurator = Config
150
142
  end
@@ -1,7 +1,6 @@
1
1
  module Dor
2
2
  # TODO: class docs
3
3
  class WorkflowDs < ActiveFedora::OmDatastream
4
-
5
4
  set_terminology do |t|
6
5
  t.root(:path => 'workflows')
7
6
  t.workflow {
@@ -9,7 +8,7 @@ module Dor
9
8
  t.process {
10
9
  t.name_( :path => {:attribute => 'name' }, :index_as => [:displayable, :not_searchable] )
11
10
  t.status( :path => {:attribute => 'status' }, :index_as => [:displayable, :not_searchable] )
12
- t.timestamp(:path => {:attribute => 'datetime' }, :index_as => [:displayable, :not_searchable] ) # , :data_type => :date)
11
+ t.timestamp(:path => {:attribute => 'datetime' }, :index_as => [:displayable, :not_searchable] ) #, :data_type => :date)
13
12
  t.elapsed( :path => {:attribute => 'elapsed' }, :index_as => [:displayable, :not_searchable] )
14
13
  t.lifecycle(:path => {:attribute => 'lifecycle'}, :index_as => [:displayable, :not_searchable] )
15
14
  t.attempts( :path => {:attribute => 'attempts' }, :index_as => [:displayable, :not_searchable] )
@@ -24,7 +23,7 @@ module Dor
24
23
  Workflow::Document.new(xml.to_s)
25
24
  end
26
25
 
27
- alias_method :[], :get_workflow
26
+ alias :[] :get_workflow
28
27
 
29
28
  def ng_xml
30
29
  @ng_xml ||= Nokogiri::XML::Document.parse(content)
@@ -233,10 +233,15 @@ module Dor
233
233
 
234
234
  # populate cache if necessary
235
235
  unless title_hash.key?(rel_druid)
236
- related_obj = Dor.find(rel_druid)
237
- related_obj_title = get_related_obj_display_title(related_obj, rel_druid)
238
- is_from_hydrus = (related_obj && related_obj.tags.include?('Project : Hydrus'))
239
- title_hash[rel_druid] = {'related_obj_title' => related_obj_title, 'is_from_hydrus' => is_from_hydrus}
236
+ begin
237
+ related_obj = Dor.find(rel_druid)
238
+ related_obj_title = get_related_obj_display_title(related_obj, rel_druid)
239
+ is_from_hydrus = (related_obj && related_obj.tags.include?('Project : Hydrus'))
240
+ title_hash[rel_druid] = {'related_obj_title' => related_obj_title, 'is_from_hydrus' => is_from_hydrus}
241
+ rescue ActiveFedora::ObjectNotFoundError
242
+ # This may happen if the given APO or Collection does not exist (bad data)
243
+ title_hash[rel_druid] = {'related_obj_title' => rel_druid, 'is_from_hydrus' => false}
244
+ end
240
245
  end
241
246
 
242
247
  # cache should definitely be populated, so just use that to write solr field
@@ -69,9 +69,10 @@ module Dor
69
69
 
70
70
  # Call dor services app to have it publish the metadata
71
71
  def publish_metadata_remotely
72
- dor_services = RestClient::Resource.new(Config.dor_services.url + "/v1/objects/#{pid}/publish")
73
- dor_services.post ''
74
- dor_services.url
72
+ dor_services = Dor::Config.dor_services.rest_client
73
+ endpoint = dor_services["v1/objects/#{pid}/publish"]
74
+ endpoint.post ''
75
+ endpoint.url
75
76
  end
76
77
  end
77
78
  end
@@ -13,15 +13,6 @@ module Dor
13
13
  Dor::VERSION
14
14
  end
15
15
 
16
- def reindex(*pids)
17
- client = Config.gsearch.rest_client
18
- pids.in_groups_of(20, false) do |group|
19
- group.each { |pid| client["?operation=updateIndex&action=fromPid&value=#{pid}"].get }
20
- yield group if block_given?
21
- end
22
- pids
23
- end
24
-
25
16
  def risearch(query, opts = {})
26
17
  client = Config.fedora.client['risearch']
27
18
  client.options[:timeout] = opts.delete(:timeout)
@@ -54,19 +45,6 @@ module Dor
54
45
  end
55
46
  end
56
47
 
57
- def gsearch(params)
58
- client = Config.gsearch.client
59
- query_params = params.merge(:wt => 'json')
60
- query_string = query_params.collect { |k, v|
61
- if v.is_a?(Array)
62
- v.collect { |vv| "#{k}=#{URI.encode(vv.to_s)}" }.join('&')
63
- else
64
- "#{k}=#{URI.encode(v.to_s)}"
65
- end
66
- }.join('&')
67
- JSON.parse(client["select?#{query_string}"].get)
68
- end
69
-
70
48
  def query(query, args = {})
71
49
  params = args.merge({ :q => query })
72
50
  params[:start] ||= 0
data/lib/dor/version.rb CHANGED
@@ -1,3 +1,3 @@
1
1
  module Dor
2
- VERSION = '5.3.4'
2
+ VERSION = '5.4.0'
3
3
  end
metadata CHANGED
@@ -1,7 +1,7 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: dor-services
3
3
  version: !ruby/object:Gem::Version
4
- version: 5.3.4
4
+ version: 5.4.0
5
5
  platform: ruby
6
6
  authors:
7
7
  - Michael Klein
@@ -14,7 +14,7 @@ authors:
14
14
  autorequire:
15
15
  bindir: bin
16
16
  cert_chain: []
17
- date: 2016-02-12 00:00:00.000000000 Z
17
+ date: 2016-02-20 00:00:00.000000000 Z
18
18
  dependencies:
19
19
  - !ruby/object:Gem::Dependency
20
20
  name: active-fedora
@@ -552,6 +552,34 @@ dependencies:
552
552
  - - ">="
553
553
  - !ruby/object:Gem::Version
554
554
  version: '0'
555
+ - !ruby/object:Gem::Dependency
556
+ name: vcr
557
+ requirement: !ruby/object:Gem::Requirement
558
+ requirements:
559
+ - - ">="
560
+ - !ruby/object:Gem::Version
561
+ version: '0'
562
+ type: :development
563
+ prerelease: false
564
+ version_requirements: !ruby/object:Gem::Requirement
565
+ requirements:
566
+ - - ">="
567
+ - !ruby/object:Gem::Version
568
+ version: '0'
569
+ - !ruby/object:Gem::Dependency
570
+ name: webmock
571
+ requirement: !ruby/object:Gem::Requirement
572
+ requirements:
573
+ - - ">="
574
+ - !ruby/object:Gem::Version
575
+ version: '0'
576
+ type: :development
577
+ prerelease: false
578
+ version_requirements: !ruby/object:Gem::Requirement
579
+ requirements:
580
+ - - ">="
581
+ - !ruby/object:Gem::Version
582
+ version: '0'
555
583
  - !ruby/object:Gem::Dependency
556
584
  name: yard
557
585
  requirement: !ruby/object:Gem::Requirement
@@ -651,8 +679,6 @@ files:
651
679
  - lib/dor/workflow/document.rb
652
680
  - lib/dor/workflow/graph.rb
653
681
  - lib/dor/workflow/process.rb
654
- - lib/gsearch/demoFoxmlToSolr.xslt
655
- - lib/gsearch/schema.xml
656
682
  - lib/tasks/rdoc.rake
657
683
  homepage:
658
684
  licenses:
@@ -675,7 +701,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
675
701
  version: 1.3.6
676
702
  requirements: []
677
703
  rubyforge_project:
678
- rubygems_version: 2.4.8
704
+ rubygems_version: 2.4.5.1
679
705
  signing_key:
680
706
  specification_version: 4
681
707
  summary: Ruby implmentation of DOR services used by the SULAIR Digital Library
@@ -1,602 +0,0 @@
1
- <?xml version="1.0" encoding="UTF-8"?>
2
- <!-- $Id: demoFoxmlToLucene.xslt 5734 2006-11-28 11:20:15Z gertsp $ -->
3
- <xsl:stylesheet version="1.0"
4
- exclude-result-prefixes="dc dt ex exts fedora-model fedora-types fn foxml hydra oai_dc rdf rel uvalibadmin uvalibdesc xsl"
5
- xmlns:dc="http://purl.org/dc/elements/1.1/"
6
- xmlns:exts="xalan://dk.defxws.fedoragsearch.server.GenericOperationsImpl"
7
- xmlns:ex="http://exslt.org/common"
8
- xmlns:dt="http://exslt.org/dates-and-times"
9
- xmlns:fedora-model="info:fedora/fedora-system:def/model#"
10
- xmlns:fedora-types="http://www.fedora.info/definitions/1/0/types/"
11
- xmlns:fn="http://www.w3.org/TR/xpath-functions/"
12
- xmlns:foxml="info:fedora/fedora-system:def/foxml#"
13
- xmlns:hydra="http://projecthydra.org/ns/relations#" xmlns:mods="http://www.loc.gov/mods/v3"
14
- xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/"
15
- xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
16
- xmlns:rel="info:fedora/fedora-system:def/relations-external#"
17
- xmlns:uvalibadmin="http://dl.lib.virginia.edu/bin/admin/admin.dtd/"
18
- xmlns:uvalibdesc="http://dl.lib.virginia.edu/bin/dtd/descmeta/descmeta.dtd"
19
- xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:zs="http://www.loc.gov/zing/srw/">
20
- <xsl:output encoding="UTF-8" indent="yes" method="xml"/>
21
- <!--
22
- This xslt stylesheet generates the Solr doc element consisting of field elements
23
- from a FOXML record. The PID field is mandatory.
24
- Options for tailoring:
25
- - generation of fields from other XML metadata streams than DC
26
- - generation of fields from other datastream types than XML
27
- - from datastream by ID, text fetched, if mimetype can be handled
28
- currently the mimetypes text/plain, text/xml, text/html, application/pdf can be handled.
29
- -->
30
- <xsl:variable name="INDEXVERSION">2.2.0</xsl:variable>
31
-
32
- <xsl:param name="INCLUDE_EXTERNALS" select="true()"/>
33
- <xsl:param name="REPOSITORYNAME" select="repositoryName"/>
34
- <xsl:param name="FEDORASOAP" select="repositoryName"/>
35
- <xsl:param name="FEDORAUSER" select="repositoryName"/>
36
- <xsl:param name="FEDORAPASS" select="repositoryName"/>
37
- <xsl:param name="TRUSTSTOREPATH" select="repositoryName"/>
38
- <xsl:param name="TRUSTSTOREPASS" select="repositoryName"/>
39
- <xsl:variable name="PID" select="/foxml:digitalObject/@PID"/>
40
- <xsl:variable name="docBoost" select="1.4*2.5"/>
41
- <xsl:variable name="OBJECTTYPE"
42
- select="//foxml:datastream/foxml:datastreamVersion[last()]//identityMetadata/objectType/text()"/>
43
- <xsl:variable name="first-workflow-url">
44
- <xsl:value-of select="//foxml:contentLocation[@TYPE='URL' and contains(@REF, 'workflow/dor/objects')]/@REF"/>
45
- </xsl:variable>
46
- <xsl:variable name="workflow-stem">
47
- <xsl:value-of select="substring-before($first-workflow-url, $PID)"/>
48
- </xsl:variable>
49
- <xsl:variable name="DATASTREAM_LIST">
50
- <ds name="RELS-EXT"/>
51
- <ds name="DC"/>
52
- <ds name="identityMetadata"/>
53
- <ds name="descMetadata"/>
54
- <ds name="embargoMetadata"/>
55
- <ds name="administrativeMetadata"/>
56
- <ds name="roleMetadata"/>
57
- <ds name="contentMetadata"/>
58
- <ds name="rightsMetadata"/>
59
- <ds name="workflows"/>
60
- <!--<ds match="WF"/>-->
61
- </xsl:variable>
62
-
63
- <!-- or any other calculation, default boost is 1.0 -->
64
- <xsl:template match="/">
65
- <add>
66
- <doc boost="{$docBoost}">
67
- <xsl:apply-templates/>
68
- </doc>
69
- </add>
70
- </xsl:template>
71
-
72
- <xsl:template match="/foxml:digitalObject">
73
- <field name="index_version_field">
74
- <xsl:value-of select="$INDEXVERSION"/>
75
- </field>
76
- <field boost="2.5" name="PID">
77
- <xsl:value-of select="$PID"/>
78
- </field>
79
- <field name="namespace_field">
80
- <xsl:value-of select="substring-before($PID,':')"/>
81
- </field>
82
- <field name="namespace_facet">
83
- <xsl:value-of select="substring-before($PID,':')"/>
84
- </field>
85
- <field name="link_text_display">
86
- <xsl:choose>
87
- <xsl:when test="//dc:title">
88
- <xsl:value-of select="//dc:title/text()"/>
89
- </xsl:when>
90
- <xsl:otherwise>
91
- <xsl:value-of select="$PID"/>
92
- </xsl:otherwise>
93
- </xsl:choose>
94
- </field>
95
- <xsl:call-template name="lifecycle"/>
96
- <xsl:apply-templates select="*"/>
97
- <xsl:call-template name="workflows"/>
98
- </xsl:template>
99
-
100
- <xsl:template match="foxml:objectProperties/foxml:property">
101
- <field>
102
- <xsl:attribute name="name">
103
- <!-- if this is a data field, append with date, otherwise field -->
104
- <xsl:choose>
105
- <xsl:when test="contains(@NAME, 'Date')">
106
- <xsl:value-of select="concat('fgs_', substring-after(@NAME,'#'), '_date')"/>
107
- </xsl:when>
108
- <xsl:otherwise>
109
- <xsl:value-of select="concat('fgs_', substring-after(@NAME,'#'), '_field')"
110
- />
111
- </xsl:otherwise>
112
- </xsl:choose>
113
- </xsl:attribute>
114
- <xsl:value-of select="@VALUE"/>
115
- </field>
116
- </xsl:template>
117
-
118
- <xsl:template match="foxml:datastream">
119
- <field name="fedora_datastream_version_field">
120
- <xsl:value-of select="foxml:datastreamVersion[last()]/@ID"/>
121
- </field>
122
- <xsl:apply-templates select="foxml:datastreamVersion[last()]"/>
123
- </xsl:template>
124
-
125
- <!-- Index inline datastreams -->
126
- <xsl:template match="foxml:datastream[foxml:datastreamVersion/foxml:xmlContent]/foxml:datastreamVersion[last()]">
127
- <xsl:variable name="datastream-name" select="../@ID"/>
128
- <xsl:apply-templates select="foxml:xmlContent/*">
129
- <xsl:with-param name="datastream-name" select="$datastream-name"/>
130
- </xsl:apply-templates>
131
- </xsl:template>
132
-
133
- <!-- Index managed/external datastreams -->
134
- <xsl:template match="foxml:datastream[foxml:datastreamVersion/foxml:contentLocation]/foxml:datastreamVersion[last()]">
135
- <xsl:if test="$INCLUDE_EXTERNALS">
136
- <xsl:variable name="ds" select="."/>
137
- <xsl:call-template name="process-datastream">
138
- <xsl:with-param name="ds" select="$ds"/>
139
- <xsl:with-param name="datastreams" select="ex:node-set($DATASTREAM_LIST)"/>
140
- </xsl:call-template>
141
- </xsl:if>
142
- </xsl:template>
143
-
144
- <xsl:template name="process-datastream">
145
- <xsl:param name="ds"/>
146
- <xsl:param name="datastreams"/>
147
- <xsl:variable name="datastream-name" select="$ds/../@ID"/>
148
- <xsl:variable name="datastream-ts" select="$ds/@CREATED"/>
149
- <xsl:variable name="content-location" select="$ds/foxml:contentLocation/@REF"/>
150
- <xsl:for-each select="$datastreams/*">
151
- <xsl:if test="(@name and (@name = $datastream-name)) or (@match and contains($datastream-name,@match))">
152
- <xsl:variable name="content-uri">
153
- <xsl:choose>
154
- <xsl:when test="contains($content-location, '/fedora/get/')">http://localhost:8080/fedora/<xsl:value-of select="substring-after($content-location,'/fedora/')"/></xsl:when>
155
- <xsl:otherwise><xsl:value-of select="$content-location"/></xsl:otherwise>
156
- </xsl:choose>
157
- </xsl:variable>
158
- <xsl:message>Retrieving <xsl:value-of select="$content-uri"/></xsl:message>
159
- <xsl:apply-templates select="document($content-uri)/*">
160
- <xsl:with-param name="datastream-name" select="$datastream-name"/>
161
- </xsl:apply-templates>
162
- </xsl:if>
163
- </xsl:for-each>
164
- </xsl:template>
165
-
166
- <xsl:template name="workflows">
167
- <xsl:variable name="combined-workflow-stream" select="document(concat($workflow-stem,$PID,'/workflows'))"/>
168
- <xsl:apply-templates select="$combined-workflow-stream/workflows/workflow"/>
169
- </xsl:template>
170
-
171
- <!-- Index RELS-EXT -->
172
- <xsl:template match="rdf:RDF/rdf:Description">
173
- <!-- Grab the cmodel -->
174
- <xsl:for-each select="./fedora-model:hasModel">
175
- <field name="fedora_has_model_field">
176
- <xsl:value-of select="@rdf:resource"/>
177
- </field>
178
- </xsl:for-each>
179
- <xsl:for-each select="*[@rdf:resource]">
180
- <xsl:variable name="doc-pid" select="substring-after(./@rdf:resource,'info:fedora/')"/>
181
- <field name="{local-name(.)}_id_field">
182
- <xsl:value-of select="$doc-pid"/>
183
- </field>
184
- <field name="{local-name(.)}_id_facet">
185
- <xsl:value-of select="$doc-pid"/>
186
- </field>
187
- </xsl:for-each>
188
- </xsl:template>
189
-
190
- <!-- Index DC -->
191
- <xsl:template match="oai_dc:dc">
192
- <xsl:for-each select="dc:title|dc:creator|dc:identifier">
193
- <field name="dc_{local-name(.)}_text">
194
- <xsl:value-of select="text()"/>
195
- </field>
196
- </xsl:for-each>
197
- <xsl:for-each select="./*">
198
- <field name="dc_{local-name(.)}_field">
199
- <xsl:value-of select="text()"/>
200
- </field>
201
- </xsl:for-each>
202
- </xsl:template>
203
-
204
-
205
-
206
- <!-- Index identity metadata -->
207
- <xsl:template match="identityMetadata">
208
- <xsl:for-each select="./objectType">
209
- <field name="object_type_field">
210
- <xsl:value-of select="./text()"/>
211
- </field>
212
- </xsl:for-each>
213
- <xsl:for-each select="./sourceId">
214
- <field name="dor_id_field">
215
- <xsl:value-of select="concat(@source, ':', normalize-space(./text()))"/>
216
- </field>
217
- <field name="source_id_field">
218
- <xsl:value-of select="concat(@source, ':', normalize-space(./text()))"/>
219
- </field>
220
- <field name="identifier_text">
221
- <xsl:value-of select="concat(@source, ':', normalize-space(./text()))"/>
222
- </field>
223
- <field name="identifier_text">
224
- <xsl:value-of select="normalize-space(./text())"/>
225
- </field>
226
- </xsl:for-each>
227
- <xsl:for-each select="./otherId">
228
- <field name="dor_id_field">
229
- <xsl:value-of select="concat(@name, ':', normalize-space(./text()))"/>
230
- </field>
231
- <field name="dor_{@name}_id_field">
232
- <xsl:value-of select="normalize-space(./text())"/>
233
- </field>
234
- </xsl:for-each>
235
- <!-- tags -->
236
- <xsl:for-each select="./tag">
237
- <xsl:variable name="text-value" select="normalize-space(./text())"/>
238
- <field name="tag_field">
239
- <xsl:value-of select="$text-value"/>
240
- </field>
241
- <field name="tag_facet">
242
- <xsl:value-of select="$text-value"/>
243
- </field>
244
- <xsl:variable name="tag-name"
245
- select="normalize-space(substring-before($text-value, ':'))"/>
246
- <xsl:variable name="field-name">
247
- <xsl:call-template name="valid-field-name">
248
- <xsl:with-param name="name" select="$tag-name"/>
249
- </xsl:call-template>
250
- <xsl:text>_tag</xsl:text>
251
- </xsl:variable>
252
- <field name="{$field-name}_field">
253
- <xsl:value-of select="normalize-space(substring-after($text-value, ':'))"/>
254
- </field>
255
- <field name="{$field-name}_facet">
256
- <xsl:value-of select="normalize-space(substring-after($text-value, ':'))"/>
257
- </field>
258
- </xsl:for-each>
259
- </xsl:template>
260
-
261
- <!-- Index MODS descriptive metadata -->
262
- <xsl:template match="mods:mods">
263
- <field name="metadata_format_field">mods</field>
264
- <field name="metadata_format_facet">mods</field>
265
- <!-- Grab the MODS identifiers -->
266
- <xsl:for-each select="./mods:identifier">
267
- <xsl:variable name="identifier-label">
268
- <xsl:call-template name="valid-field-name">
269
- <xsl:with-param name="name" select="@displayLabel"/>
270
- </xsl:call-template>
271
- </xsl:variable>
272
- <field name="mods_identifier_field">
273
- <xsl:value-of select="@displayLabel"/>:<xsl:value-of select="text()"/>
274
- </field>
275
- <field name="mods_{$identifier-label}_identifier_field">
276
- <xsl:value-of select="text()"/>
277
- </field>
278
- <field name="mods_identifier_text">
279
- <xsl:value-of select="@displayLabel"/>:<xsl:value-of select="text()"/>
280
- </field>
281
- <field name="mods_{$identifier-label}_identifier_text">
282
- <xsl:value-of select="text()"/>
283
- </field>
284
- </xsl:for-each>
285
- <xsl:for-each select="mods:titleInfo">
286
- <xsl:variable name="title-info">
287
- <xsl:call-template name="mods-title"/>
288
- </xsl:variable>
289
- <field name="mods_titleInfo_field">
290
- <xsl:value-of select="$title-info"/>
291
- </field>
292
- <field name="mods_title_text">
293
- <xsl:value-of select="$title-info"/>
294
- </field>
295
- </xsl:for-each>
296
- <xsl:for-each select="mods:name">
297
- <xsl:variable name="name-info">
298
- <xsl:call-template name="mods-name"/>
299
- </xsl:variable>
300
- <field name="mods_name_field">
301
- <xsl:value-of select="$name-info"/>
302
- </field>
303
- <field name="mods_name_text">
304
- <xsl:value-of select="$name-info"/>
305
- </field>
306
- <xsl:if test="mods:role/mods:roleTerm[@type='text']">
307
- <xsl:variable name="role" select="mods:role/mods:roleTerm[@type='text']/text()"/>
308
- <field name="mods_{$role}_field">
309
- <xsl:value-of select="mods:namePart/text()"/>
310
- </field>
311
- <field name="mods_{$role}_text">
312
- <xsl:value-of select="mods:namePart/text()"/>
313
- </field>
314
- </xsl:if>
315
- </xsl:for-each>
316
- <xsl:for-each select="mods:originInfo/mods:publisher">
317
- <field name="mods_publisher_field">
318
- <xsl:value-of select="text()"/>
319
- </field>
320
- <field name="mods_publisher_text">
321
- <xsl:value-of select="text()"/>
322
- </field>
323
- </xsl:for-each>
324
- <xsl:for-each select="mods:originInfo/mods:place/mods:placeTerm[@type='text']">
325
- <field name="mods_origininfo_place_field">
326
- <xsl:value-of select="text()"/>
327
- </field>
328
- <field name="mods_origininfo_place_text">
329
- <xsl:value-of select="text()"/>
330
- </field>
331
- </xsl:for-each>
332
- <!-- Index some, but not all, MODS fields -->
333
- <xsl:for-each select="//mods:coordinates|//mods:extent|//mods:scale|//mods:topic">
334
- <xsl:if test="./text()">
335
- <field name="mods_{local-name(.)}_field">
336
- <xsl:value-of select="text()"/>
337
- </field>
338
- </xsl:if>
339
- </xsl:for-each>
340
- <xsl:for-each
341
- select="//mods:*[contains(local-name(),'date') or contains(local-name(), 'Date')]">
342
- <xsl:variable name="date-label">
343
- <xsl:call-template name="valid-field-name">
344
- <xsl:with-param name="name" select="local-name()"/>
345
- </xsl:call-template>
346
- </xsl:variable>
347
- <field name="mods_{$date-label}_field">
348
- <xsl:value-of select="normalize-space(./text())"/>
349
- </field>
350
- </xsl:for-each>
351
- <!--
352
- <xsl:for-each select=".//mods:*[string-length(normalize-space(text())) &gt; 0]">
353
- <field name="mods_{local-name(.)}_field"><xsl:value-of select="normalize-space(./text())"/></field>
354
- </xsl:for-each>
355
- -->
356
- </xsl:template>
357
-
358
- <!-- mods-title and mods-name templates cribbed from mods2dc.xslt -->
359
- <xsl:template name="mods-title">
360
- <xsl:value-of select="mods:nonSort"/>
361
- <xsl:if test="mods:nonSort">
362
-
363
- <xsl:text> </xsl:text>
364
- </xsl:if>
365
- <xsl:value-of select="mods:title"/>
366
- <xsl:if test="mods:subTitle">
367
- <xsl:text>: </xsl:text>
368
- <xsl:value-of select="mods:subTitle"/>
369
- </xsl:if>
370
- <xsl:if test="mods:partNumber">
371
-
372
- <xsl:text>. </xsl:text>
373
- <xsl:value-of select="mods:partNumber"/>
374
- </xsl:if>
375
- <xsl:if test="mods:partName">
376
- <xsl:text>. </xsl:text>
377
- <xsl:value-of select="mods:partName"/>
378
- </xsl:if>
379
- </xsl:template>
380
-
381
- <xsl:template name="mods-name">
382
- <xsl:variable name="name">
383
- <xsl:for-each select="mods:namePart[not(@type)]">
384
- <xsl:value-of select="."/>
385
-
386
- <xsl:text> </xsl:text>
387
- </xsl:for-each>
388
- <xsl:value-of select="mods:namePart[@type='family']"/>
389
- <xsl:if test="mods:namePart[@type='given']">
390
- <xsl:text>, </xsl:text>
391
- <xsl:value-of select="mods:namePart[@type='given']"/>
392
- </xsl:if>
393
- <xsl:if test="mods:namePart[@type='date']">
394
-
395
- <xsl:text>, </xsl:text>
396
- <xsl:value-of select="mods:namePart[@type='date']"/>
397
- <xsl:text/>
398
- </xsl:if>
399
- <xsl:if test="mods:displayForm">
400
- <xsl:text> (</xsl:text>
401
- <xsl:value-of select="mods:displayForm"/>
402
-
403
- <xsl:text>) </xsl:text>
404
- </xsl:if>
405
- <xsl:for-each select="mods:role[mods:roleTerm[@type='text']!='creator']">
406
- <xsl:text> (</xsl:text>
407
- <xsl:value-of select="normalize-space(.)"/>
408
- <xsl:text>) </xsl:text>
409
- </xsl:for-each>
410
-
411
- </xsl:variable>
412
- <xsl:value-of select="normalize-space($name)"/>
413
- </xsl:template>
414
-
415
- <!-- Index content metadata -->
416
- <xsl:template match="contentMetadata">
417
- <field name="content_type_facet">
418
- <xsl:value-of select="@type"/>
419
- </field>
420
- <xsl:apply-templates select="./resource">
421
- <xsl:sort data-type="number" select="@sequence"/>
422
- </xsl:apply-templates>
423
- </xsl:template>
424
- <xsl:template match="contentMetadata/resource/file">
425
- <field name="content_file_field">
426
- <xsl:value-of select="@id"/>
427
- </field>
428
- <xsl:if test="@shelve = 'yes'">
429
- <field name="shelved_content_file_field">
430
- <xsl:value-of select="@id"/>
431
- </field>
432
- </xsl:if>
433
- </xsl:template>
434
-
435
- <!-- Index rights metadata -->
436
- <xsl:template match="rightsMetadata">
437
- <xsl:if test="access[@type='read']/machine/embargoReleaseDate">
438
- <field name="embargo_status_facet">embargoed</field>
439
- <field name="embargo_status_field">embargoed</field>
440
- <field name="embargo_release_date">
441
- <xsl:call-template name="long-enough-date">
442
- <xsl:with-param name="date" select="access[@type='read']/machine/embargoReleaseDate/text()"/>
443
- </xsl:call-template>
444
- </field>
445
- </xsl:if>
446
- </xsl:template>
447
-
448
- <!-- Index embargo metadata -->
449
- <xsl:template match="embargoMetadata">
450
- <xsl:if test="(status != '') and (releaseDate != '')">
451
- <field name="embargo_status_facet">
452
- <xsl:value-of select="status"/>
453
- </field>
454
- <field name="embargo_status_field">
455
- <xsl:value-of select="status"/>
456
- </field>
457
- <field name="embargo_release_date">
458
- <xsl:call-template name="long-enough-date">
459
- <xsl:with-param name="date" select="releaseDate"/>
460
- </xsl:call-template>
461
- </field>
462
- </xsl:if>
463
- </xsl:template>
464
-
465
- <!-- Workflows -->
466
- <xsl:template name="lifecycle">
467
- <xsl:variable name="lifecycle-uri" select="concat($workflow-stem, $PID, '/lifecycle')"/>
468
- <xsl:message>Retrieving <xsl:value-of select="$lifecycle-uri"/></xsl:message>
469
- <xsl:variable name="doc" select="document($lifecycle-uri)"/>
470
- <xsl:apply-templates select="$doc/*"/>
471
- </xsl:template>
472
-
473
- <xsl:template match="lifecycle">
474
- <xsl:for-each select="milestone">
475
- <xsl:variable name="zdate" select="concat(substring(@date,1,19),'Z')"/>
476
- <field name="lifecycle_field">
477
- <xsl:value-of select="text()"/>:<xsl:value-of select="$zdate"/>
478
- </field>
479
- <field>
480
- <xsl:attribute name="name">
481
- <xsl:value-of select="concat(text(),'_date')"/>
482
- </xsl:attribute>
483
- <xsl:value-of select="$zdate"/>
484
- </field>
485
- <xsl:if test="position() = last()">
486
- <field name="lifecycle_facet">
487
- <xsl:value-of select="text()"/>
488
- </field>
489
- </xsl:if>
490
- </xsl:for-each>
491
- </xsl:template>
492
-
493
- <xsl:template match="workflows">
494
- <xsl:apply-templates select="./workflow"/>
495
- </xsl:template>
496
-
497
- <xsl:template match="workflow">
498
- <xsl:param name="datastream-name" select="@id | ancestor::foxml:datastream/@ID"/>
499
- <xsl:variable name="workflow-name" select="$datastream-name"/>
500
- <xsl:variable name="workflow-token">
501
- <xsl:call-template name="valid-field-name">
502
- <xsl:with-param name="name" select="$workflow-name"/>
503
- </xsl:call-template>
504
- </xsl:variable>
505
- <field name="wf_facet">
506
- <xsl:value-of select="$workflow-name"/>
507
- </field>
508
- <field name="wf_wsp_facet">
509
- <xsl:value-of select="$workflow-name"/>
510
- </field>
511
- <field name="wf_wps_facet">
512
- <xsl:value-of select="$workflow-name"/>
513
- </field>
514
- <xsl:for-each select="process">
515
- <xsl:sort select="@datetime" data-type="text"/>
516
- <field name="wf_wsp_facet">
517
- <xsl:value-of select="concat($workflow-name,':',@status)"/>
518
- </field>
519
- <field name="wf_wsp_facet">
520
- <xsl:value-of select="concat($workflow-name,':',@status,':',@name)"/>
521
- </field>
522
- <field name="wf_wps_facet">
523
- <xsl:value-of select="concat($workflow-name,':',@name)"/>
524
- </field>
525
- <field name="wf_wps_facet">
526
- <xsl:value-of select="concat($workflow-name,':',@name,':',@status)"/>
527
- </field>
528
- <field name="wf_swp_facet">
529
- <xsl:value-of select="@status"/>
530
- </field>
531
- <field name="wf_swp_facet">
532
- <xsl:value-of select="concat(@status,':',$workflow-name)"/>
533
- </field>
534
- <field name="wf_swp_facet">
535
- <xsl:value-of select="concat(@status,':',$workflow-name,':',@name)"/>
536
- </field>
537
- </xsl:for-each>
538
- </xsl:template>
539
-
540
- <!-- Admin Policy specific fields -->
541
- <xsl:template match="administrativeMetadata">
542
- <xsl:if test="./descMetadata/format">
543
- <field name="apo_metadata_format_field">
544
- <xsl:value-of select="./descMetadata/format/text()"/>
545
- </field>
546
- </xsl:if>
547
- <xsl:if test="./descMetadata/format">
548
- <field name="apo_metadata_source_field">
549
- <xsl:value-of select="./descMetadata/source/text()"/>
550
- </field>
551
- </xsl:if>
552
- <xsl:for-each select="registration/workflow">
553
- <field name="apo_registration_workflow_field">
554
- <xsl:value-of select="@id"/>
555
- </field>
556
- </xsl:for-each>
557
- </xsl:template>
558
-
559
- <xsl:template match="roleMetadata">
560
- <xsl:for-each select="./role/*">
561
- <xsl:variable name="role_value"><xsl:value-of select="identifier/@type"/>:<xsl:value-of
562
- select="identifier/text()"/></xsl:variable>
563
- <field name="apo_role_{local-name(.)}_{../@type}_field">
564
- <xsl:value-of select="$role_value"/>
565
- </field>
566
- <field name="apo_role_{local-name(.)}_{../@type}_facet">
567
- <xsl:value-of select="$role_value"/>
568
- </field>
569
- <xsl:if test="../@type = 'depositor' or ../@type = 'manager'">
570
- <field name="apo_register_permissions_field">
571
- <xsl:value-of select="$role_value"/>
572
- </field>
573
- <field name="apo_register_permissions_facet">
574
- <xsl:value-of select="$role_value"/>
575
- </field>
576
- </xsl:if>
577
- </xsl:for-each>
578
- </xsl:template>
579
-
580
- <!-- Utility Templates -->
581
- <xsl:template match="text()|@*|processing-instruction()|comment()"/>
582
-
583
- <xsl:template name="long-enough-date">
584
- <xsl:param name="date"/>
585
- <xsl:variable name="date-format" select="'0000-00-00T00:00:00Z'"/>
586
- <xsl:choose>
587
- <xsl:when test="string-length($date) &lt; string-length($date-format)">
588
- <xsl:value-of select="$date"/><xsl:value-of select="substring($date-format,string-length($date)+1)"/>
589
- </xsl:when>
590
- <xsl:otherwise>
591
- <xsl:value-of select="$date"/>
592
- </xsl:otherwise>
593
- </xsl:choose>
594
- </xsl:template>
595
-
596
- <xsl:template name="valid-field-name">
597
- <xsl:param name="name"/>
598
- <xsl:value-of
599
- select="translate($name,' ABCDEFGHIJKLMNOPQRSTUVWXYZ','_abcdefghijklmnopqrstuvwxyz')"/>
600
- </xsl:template>
601
-
602
- </xsl:stylesheet>
@@ -1,229 +0,0 @@
1
- <?xml version="1.0" ?>
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
- <!-- This is the Solr schema file. This file should be named "schema.xml" and
20
- should be in the conf directory under the solr home
21
- (i.e. ./solr/conf/schema.xml by default)
22
- or located where the classloader for the Solr webapp can find it.
23
-
24
- For more information, on how to customize this file, please see
25
- http://wiki.apache.org/solr/SchemaXml
26
- -->
27
-
28
- <schema name="active_fedora" version="0.1">
29
- <!-- attribute "name" is the name of this schema and is only used for display purposes.
30
- Applications should change this to reflect the nature of the search collection.
31
- version="1.1" is Solr's version number for the schema syntax and semantics. It should
32
- not normally be changed by applications.
33
- 1.0: multiValued attribute did not exist, all fields are multiValued by nature
34
- 1.1: multiValued attribute introduced, false by default -->
35
-
36
- <types>
37
- <!-- field type definitions. The "name" attribute is
38
- just a label to be used by field definitions. The "class"
39
- attribute and any other attributes determine the real
40
- behavior of the fieldtype.
41
- Class names starting with "solr" refer to java classes in the
42
- org.apache.solr.analysis package.
43
- -->
44
-
45
- <!-- The StrField type is not analyzed, but indexed/stored verbatim.
46
- - StrField and TextField support an optional compressThreshold which
47
- limits compression (if enabled in the derived fields) to values which
48
- exceed a certain size (in characters).
49
- -->
50
- <fieldtype name="string" class="solr.StrField" sortMissingLast="true" omitNorms="true"/>
51
-
52
- <!-- boolean type: "true" or "false" -->
53
- <fieldtype name="boolean" class="solr.BoolField" sortMissingLast="true" omitNorms="true"/>
54
-
55
- <!-- The optional sortMissingLast and sortMissingFirst attributes are
56
- currently supported on types that are sorted internally as strings.
57
- - If sortMissingLast="true", then a sort on this field will cause documents
58
- without the field to come after documents with the field,
59
- regardless of the requested sort order (asc or desc).
60
- - If sortMissingFirst="true", then a sort on this field will cause documents
61
- without the field to come before documents with the field,
62
- regardless of the requested sort order.
63
- - If sortMissingLast="false" and sortMissingFirst="false" (the default),
64
- then default lucene sorting will be used which places docs without the
65
- field first in an ascending sort and last in a descending sort.
66
- -->
67
-
68
-
69
- <!-- numeric field types that store and index the text
70
- value verbatim (and hence don't support range queries, since the
71
- lexicographic ordering isn't equal to the numeric ordering) -->
72
- <fieldtype name="integer" class="solr.IntField" omitNorms="true"/>
73
- <fieldtype name="long" class="solr.LongField" omitNorms="true"/>
74
- <fieldtype name="float" class="solr.FloatField" omitNorms="true"/>
75
- <fieldtype name="double" class="solr.DoubleField" omitNorms="true"/>
76
-
77
-
78
- <!-- Numeric field types that manipulate the value into
79
- a string value that isn't human-readable in its internal form,
80
- but with a lexicographic ordering the same as the numeric ordering,
81
- so that range queries work correctly. -->
82
- <fieldtype name="sint" class="solr.SortableIntField" sortMissingLast="true" omitNorms="true"/>
83
- <fieldtype name="slong" class="solr.SortableLongField" sortMissingLast="true" omitNorms="true"/>
84
- <fieldtype name="sfloat" class="solr.SortableFloatField" sortMissingLast="true" omitNorms="true"/>
85
- <fieldtype name="sdouble" class="solr.SortableDoubleField" sortMissingLast="true" omitNorms="true"/>
86
-
87
-
88
- <!-- The format for this date field is of the form 1995-12-31T23:59:59Z, and
89
- is a more restricted form of the canonical representation of dateTime
90
- http://www.w3.org/TR/xmlschema-2/#dateTime
91
- The trailing "Z" designates UTC time and is mandatory.
92
- Optional fractional seconds are allowed: 1995-12-31T23:59:59.999Z
93
- All other components are mandatory.
94
-
95
- Expressions can also be used to denote calculations that should be
96
- performed relative to "NOW" to determine the value, ie...
97
-
98
- NOW/HOUR
99
- ... Round to the start of the current hour
100
- NOW-1DAY
101
- ... Exactly 1 day prior to now
102
- NOW/DAY+6MONTHS+3DAYS
103
- ... 6 months and 3 days in the future from the start of
104
- the current day
105
-
106
- Consult the DateField javadocs for more information.
107
- -->
108
- <fieldtype name="date" class="solr.DateField" sortMissingLast="true" omitNorms="true"/>
109
-
110
- <!-- solr.TextField allows the specification of custom text analyzers
111
- specified as a tokenizer and a list of token filters. Different
112
- analyzers may be specified for indexing and querying.
113
-
114
- The optional positionIncrementGap puts space between multiple fields of
115
- this type on the same document, with the purpose of preventing false phrase
116
- matching across fields.
117
-
118
- For more info on customizing your analyzer chain, please see
119
- http://wiki.apache.org/solr/AnalyzersTokenizersTokenFilters
120
- -->
121
-
122
- <!-- One can also specify an existing Analyzer class that has a
123
- default constructor via the class attribute on the analyzer element
124
- <fieldtype name="text_greek" class="solr.TextField">
125
- <analyzer class="org.apache.lucene.analysis.el.GreekAnalyzer"/>
126
- </fieldType>
127
- -->
128
-
129
- <!-- A text field that only splits on whitespace for exact matching of words -->
130
- <fieldtype name="text_ws" class="solr.TextField" positionIncrementGap="100">
131
- <analyzer>
132
- <tokenizer class="solr.WhitespaceTokenizerFactory"/>
133
- </analyzer>
134
- </fieldtype>
135
-
136
- <fieldtype name="text" class="solr.TextField" positionIncrementGap="100">
137
- <analyzer>
138
- <tokenizer class="solr.StandardTokenizerFactory"/>
139
- <filter class="solr.StandardFilterFactory"/>
140
- <filter class="solr.LowerCaseFilterFactory"/>
141
- </analyzer>
142
- </fieldtype>
143
-
144
- <fieldtype name="text_zh" class="solr.TextField">
145
- <analyzer class="org.apache.lucene.analysis.cn.ChineseAnalyzer"/>
146
- </fieldtype>
147
-
148
-
149
- <!-- Less flexible matching, but less false matches. Probably not ideal for product names,
150
- but may be good for SKUs. Can insert dashes in the wrong place and still match. -->
151
- <fieldtype name="textTight" class="solr.TextField" positionIncrementGap="100" >
152
- <analyzer>
153
- <tokenizer class="solr.WhitespaceTokenizerFactory"/>
154
- <filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="false"/>
155
- <filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt"/>
156
- <filter class="solr.WordDelimiterFilterFactory" generateWordParts="0" generateNumberParts="0" catenateWords="1" catenateNumbers="1" catenateAll="0"/>
157
- <filter class="solr.LowerCaseFilterFactory"/>
158
- <filter class="solr.EnglishPorterFilterFactory" protected="protwords.txt"/>
159
- <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>
160
- </analyzer>
161
- </fieldtype>
162
-
163
- </types>
164
-
165
-
166
- <fields>
167
- <!-- Valid attributes for fields:
168
- name: mandatory - the name for the field
169
- type: mandatory - the name of a previously defined type from the <types> section
170
- indexed: true if this field should be indexed (searchable or sortable)
171
- stored: true if this field should be retrievable
172
- compressed: [false] if this field should be stored using gzip compression
173
- (this will only apply if the field type is compressable; among
174
- the standard field types, only TextField and StrField are)
175
- multiValued: true if this field may contain multiple values per document
176
- omitNorms: (expert) set to true to omit the norms associated with
177
- this field (this disables length normalization and index-time
178
- boosting for the field, and saves some memory). Only full-text
179
- fields or fields that need an index-time boost need norms.
180
- -->
181
-
182
- <field name="PID" type="string" indexed="true" stored="true"/>
183
-
184
- <!-- catchall field, containing all other searchable text fields (implemented
185
- via copyField further on in this schema -->
186
- <field name="text" type="text" indexed="true" stored="false" multiValued="true"/>
187
- <!-- catchall field, containing all other searchable text fields (implemented
188
- via copyField further on in this schema -->
189
- <field name="date" type="date" indexed="true" stored="false" multiValued="true"/>
190
-
191
- <!-- Dynamic field definitions. If a field name is not found, dynamicFields
192
- will be used if the name matches any of the patterns.
193
- RESTRICTION: the glob-like pattern in the name attribute must have
194
- a "*" only at the start or the end.
195
- EXAMPLE: name="*_i" will match any field ending in _i (like myid_i, z_i)
196
- Longer patterns will be matched first. if equal size patterns
197
- both match, the first appearing in the schema will be used. -->
198
- <dynamicField name="*_date" type="date" indexed="true" stored="true" multiValued="true"/>
199
- <dynamicField name="*_name" type="date" indexed="true" stored="true" multiValued="true"/>
200
- <dynamicField name="*_field" type="text" indexed="true" stored="true" multiValued="true"/>
201
-
202
- <dynamicField name="*_facet" type="string" indexed="true" stored="true" multiValued="true"/>
203
- <dynamicField name="*_zh_text" type="text_zh" indexed="true" stored="true" multiValued="true"/>
204
- <dynamicField name="*_text" type="text" indexed="true" stored="true" multiValued="true"/>
205
- <dynamicField name="*_display" type="text" indexed="false" stored="true" multiValued="true"/>
206
- </fields>
207
-
208
- <!-- field to use to determine and enforce document uniqueness. -->
209
- <uniqueKey>PID</uniqueKey>
210
-
211
- <!-- field for the QueryParser to use when an explicit fieldname is absent -->
212
- <defaultSearchField>text</defaultSearchField>
213
-
214
- <!-- SolrQueryParser configuration: defaultOperator="AND|OR" -->
215
- <solrQueryParser defaultOperator="AND"/>
216
-
217
- <!-- copyField commands copy one field to another at the time a document
218
- is added to the index. It's used either to index the same field differently,
219
- or to add multiple fields to the same field for easier/faster searching. -->
220
- <copyField source="*_date" dest="date"/>
221
- <copyField source="*_text" dest="text"/>
222
- <copyField source="*_facet" dest="text"/>
223
-
224
- <!-- Similarity is the scoring routine for each document vs. a query.
225
- A custom similarity may be specified here, but the default is fine
226
- for most applications. -->
227
- <!-- <similarity class="org.apache.lucene.search.DefaultSimilarity"/> -->
228
-
229
- </schema>