cul-fedora 0.7.0 → 0.8.0
Sign up to get free protection for your applications and to get access to all the features.
- data/VERSION +1 -1
- data/cul-fedora.gemspec +2 -2
- data/lib/cul-fedora/item.rb +65 -19
- data/lib/cul-fedora/solr.rb +15 -3
- metadata +4 -4
data/VERSION
CHANGED
@@ -1 +1 @@
|
|
1
|
-
0.
|
1
|
+
0.8.0
|
data/cul-fedora.gemspec
CHANGED
@@ -5,11 +5,11 @@
|
|
5
5
|
|
6
6
|
Gem::Specification.new do |s|
|
7
7
|
s.name = %q{cul-fedora}
|
8
|
-
s.version = "0.
|
8
|
+
s.version = "0.8.0"
|
9
9
|
|
10
10
|
s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
|
11
11
|
s.authors = ["James Stuart"]
|
12
|
-
s.date = %q{2011-
|
12
|
+
s.date = %q{2011-05-02}
|
13
13
|
s.description = %q{Columbia-specific Fedora libraries}
|
14
14
|
s.email = %q{tastyhat@jamesstuart.org}
|
15
15
|
s.extra_rdoc_files = [
|
data/lib/cul-fedora/item.rb
CHANGED
@@ -103,9 +103,12 @@ module Cul
|
|
103
103
|
|
104
104
|
get_fullname = lambda { |node| node.nil? ? nil : (node.css("namePart[@type='family']").collect(&:content) | node.css("namePart[@type='given']").collect(&:content)).join(", ") }
|
105
105
|
|
106
|
-
roles = ["Author","
|
107
|
-
|
106
|
+
roles = ["Author","Creator","Thesis Advisor","Collector","Owner","Speaker","Seminar Chairman","Secretary","Rapporteur","Committee Member","Degree Grantor","Moderator","Editor","Interviewee","Interviewer","Organizer of Meeting","Originator","Teacher"]
|
107
|
+
roles = roles.map { |role| role.downcase }
|
108
108
|
|
109
|
+
organizations = []
|
110
|
+
departments = []
|
111
|
+
|
109
112
|
begin
|
110
113
|
collections = self.belongsTo
|
111
114
|
meta = describedBy.first
|
@@ -124,29 +127,57 @@ module Cul
|
|
124
127
|
|
125
128
|
|
126
129
|
|
127
|
-
title =
|
130
|
+
title = mods.css("titleInfo>title").first.text
|
131
|
+
title_search = normalize_space.call(mods.css("titleInfo>nonSort,title").collect(&:content).join(" "))
|
132
|
+
record_creation_date = mods.at_css("recordInfo>recordCreationDate")
|
133
|
+
if(record_creation_date.nil?)
|
134
|
+
record_creation_date = mods.at_css("recordInfo>recordChangeDate")
|
135
|
+
end
|
136
|
+
if(!record_creation_date.nil? || !record_creation_date.empty?)
|
137
|
+
record_creation_date = DateTime.parse(record_creation_date.text.gsub("UTC", "").strip)
|
138
|
+
add_field.call("record_creation_date", record_creation_date.strftime("%Y-%m-%dT%H:%M:%SZ"))
|
139
|
+
end
|
128
140
|
add_field.call("title_display", title)
|
129
|
-
add_field.call("title_search",
|
141
|
+
add_field.call("title_search", title_search)
|
130
142
|
|
131
143
|
all_names = []
|
132
144
|
mods.css("name[@type='personal']").each do |name_node|
|
133
|
-
if name_node.css("role>roleTerm
|
145
|
+
if name_node.css("role>roleTerm").collect(&:content).any? { |role| roles.include?(role) }
|
134
146
|
|
135
147
|
fullname = get_fullname.call(name_node)
|
136
148
|
|
137
149
|
all_names << fullname
|
138
|
-
|
139
|
-
|
140
|
-
|
150
|
+
if(!name_node["ID"].nil?)
|
151
|
+
add_field.call("author_id_uni", name_node["ID"])
|
152
|
+
end
|
141
153
|
add_field.call("author_search", fullname.downcase)
|
142
154
|
add_field.call("author_facet", fullname)
|
155
|
+
|
156
|
+
name_node.css("affiliation").each do |affiliation_node|
|
157
|
+
affiliation_text = affiliation_node.text
|
158
|
+
if(affiliation_text.include?(". "))
|
159
|
+
affiliation_split = affiliation_text.split(". ")
|
160
|
+
organizations.push(affiliation_split[0].strip)
|
161
|
+
departments.push(affiliation_split[1].strip)
|
162
|
+
end
|
163
|
+
end
|
143
164
|
|
144
165
|
end
|
145
|
-
|
166
|
+
end
|
167
|
+
|
168
|
+
mods.css("name[@type='corporate']").each do |corp_name_node|
|
169
|
+
if(!corp_name_node["ID"].nil? && corp_name_node["ID"].include?("originator"))
|
170
|
+
name_part = corp_name_node.at_css("namePart").text
|
171
|
+
if(name_part.include?(". "))
|
172
|
+
name_part_split = name_part.split(". ")
|
173
|
+
organizations.push(name_part_split[0].strip)
|
174
|
+
departments.push(name_part_split[1].strip)
|
175
|
+
end
|
176
|
+
end
|
146
177
|
end
|
147
178
|
|
148
179
|
add_field.call("authors_display",all_names.join("; "))
|
149
|
-
add_field.call("
|
180
|
+
add_field.call("pub_date", mods.at_css("*[@keyDate='yes']"))
|
150
181
|
|
151
182
|
mods.css("genre").each do |genre_node|
|
152
183
|
add_field.call("genre_facet", genre_node)
|
@@ -158,14 +189,14 @@ module Cul
|
|
158
189
|
add_field.call("abstract", mods.at_css("abstract"))
|
159
190
|
add_field.call("handle", mods.at_css("identifier[@type='hdl']"))
|
160
191
|
|
161
|
-
mods.css("subject
|
162
|
-
|
163
|
-
|
164
|
-
|
165
|
-
|
166
|
-
|
167
|
-
|
168
|
-
|
192
|
+
mods.css("subject").each do |subject_node|
|
193
|
+
if(subject_node.attributes.count == 0)
|
194
|
+
subject_node.css("topic").each do |topic_node|
|
195
|
+
add_field.call("keyword_search", topic_node.content.downcase)
|
196
|
+
add_field.call("subject", topic_node)
|
197
|
+
add_field.call("subject_search", topic_node)
|
198
|
+
end
|
199
|
+
end
|
169
200
|
end
|
170
201
|
|
171
202
|
|
@@ -203,7 +234,22 @@ module Cul
|
|
203
234
|
add_field.call("geographic_area_search", geo)
|
204
235
|
end
|
205
236
|
|
206
|
-
add_field.call("export_as_mla_citation_txt","
|
237
|
+
add_field.call("export_as_mla_citation_txt","")
|
238
|
+
|
239
|
+
if(organizations.count > 0)
|
240
|
+
organizations = organizations.uniq
|
241
|
+
organizations.each do |organization|
|
242
|
+
add_field.call("affiliation_organization", organization)
|
243
|
+
end
|
244
|
+
end
|
245
|
+
|
246
|
+
if(departments.count > 0)
|
247
|
+
departments = departments.uniq
|
248
|
+
departments.each do |department|
|
249
|
+
add_field.call("affiliation_department", department.to_s.sub(", Department of", "").strip)
|
250
|
+
end
|
251
|
+
end
|
252
|
+
|
207
253
|
end
|
208
254
|
|
209
255
|
|
data/lib/cul-fedora/solr.rb
CHANGED
@@ -15,6 +15,11 @@ module Cul
|
|
15
15
|
@rsolr ||= RSolr.connect(:url => @url)
|
16
16
|
end
|
17
17
|
|
18
|
+
def delete_index
|
19
|
+
rsolr.delete_by_query("*:*")
|
20
|
+
rsolr.commit
|
21
|
+
end
|
22
|
+
|
18
23
|
def ingest(options = {})
|
19
24
|
format = options.delete(:format) || raise(ArgumentError, "needs format")
|
20
25
|
|
@@ -22,13 +27,14 @@ module Cul
|
|
22
27
|
items = [items] unless items.kind_of?(Array)
|
23
28
|
collections = options.delete(:collections) || []
|
24
29
|
collections = [collections] unless collections.kind_of?(Array)
|
30
|
+
ignore = options.delete(:ignore) || []
|
31
|
+
ignore = [ignore] unless ignore.kind_of?(Array)
|
25
32
|
|
26
33
|
overwrite = options.delete(:overwrite) || false
|
27
34
|
process = options.delete(:process) || nil
|
28
35
|
skip = options.delete(:skip) || nil
|
29
36
|
|
30
|
-
|
31
|
-
|
37
|
+
|
32
38
|
collections.each do |collection|
|
33
39
|
items |= collection.listMembers
|
34
40
|
end
|
@@ -40,6 +46,12 @@ module Cul
|
|
40
46
|
errors = {}
|
41
47
|
|
42
48
|
items.each do |i|
|
49
|
+
|
50
|
+
if(ignore.index(i.pid).nil? == false)
|
51
|
+
$stdout.puts "Ignoring " + i.pid + "..."
|
52
|
+
next
|
53
|
+
end
|
54
|
+
|
43
55
|
if process && skip && skip > 0
|
44
56
|
skip -= 1
|
45
57
|
next
|
@@ -55,7 +67,7 @@ module Cul
|
|
55
67
|
end
|
56
68
|
|
57
69
|
|
58
|
-
|
70
|
+
$stdout.puts "Indexing " + i.pid + "..."
|
59
71
|
|
60
72
|
result_hash = i.send("index_for_#{format}", options)
|
61
73
|
|
metadata
CHANGED
@@ -1,13 +1,13 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: cul-fedora
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
hash:
|
4
|
+
hash: 63
|
5
5
|
prerelease: false
|
6
6
|
segments:
|
7
7
|
- 0
|
8
|
-
-
|
8
|
+
- 8
|
9
9
|
- 0
|
10
|
-
version: 0.
|
10
|
+
version: 0.8.0
|
11
11
|
platform: ruby
|
12
12
|
authors:
|
13
13
|
- James Stuart
|
@@ -15,7 +15,7 @@ autorequire:
|
|
15
15
|
bindir: bin
|
16
16
|
cert_chain: []
|
17
17
|
|
18
|
-
date: 2011-
|
18
|
+
date: 2011-05-02 00:00:00 -04:00
|
19
19
|
default_executable:
|
20
20
|
dependencies:
|
21
21
|
- !ruby/object:Gem::Dependency
|