stanford-mods 1.3.3 → 1.3.4
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.gitignore +1 -0
- data/.rspec +1 -0
- data/.rubocop.yml +4 -0
- data/Gemfile +1 -0
- data/lib/stanford-mods.rb +5 -5
- data/lib/stanford-mods/date_parsing.rb +245 -0
- data/lib/stanford-mods/origin_info.rb +411 -0
- data/lib/stanford-mods/searchworks.rb +23 -474
- data/lib/stanford-mods/searchworks_subjects.rb +208 -0
- data/lib/stanford-mods/version.rb +1 -1
- data/spec/date_parsing_spec.rb +746 -0
- data/spec/fixtures/spotlight_pub_date_data.rb +316 -0
- data/spec/origin_info_spec.rb +449 -0
- data/spec/searchworks_pub_dates_spec.rb +166 -163
- data/spec/spec_helper.rb +16 -5
- data/stanford-mods.gemspec +2 -0
- metadata +25 -2
@@ -3,209 +3,212 @@ require 'spec_helper'
|
|
3
3
|
|
4
4
|
describe "Date methods (searchworks.rb)" do
|
5
5
|
|
6
|
-
|
7
|
-
@ns_decl = "xmlns='#{Mods::MODS_NS}'"
|
8
|
-
end
|
6
|
+
let(:ns_decl) { "xmlns='#{Mods::MODS_NS}'" }
|
9
7
|
|
10
8
|
context "pub_dates" do
|
11
|
-
it "
|
12
|
-
m = "<mods #{
|
13
|
-
|
14
|
-
|
15
|
-
|
16
|
-
|
17
|
-
|
18
|
-
|
19
|
-
expect(
|
9
|
+
it "puts dateIssued values before dateCreated values" do
|
10
|
+
m = "<mods #{ns_decl}><originInfo>
|
11
|
+
<dateCreated>1904</dateCreated>
|
12
|
+
<dateCreated>1904</dateCreated>
|
13
|
+
<dateIssued>1906</dateIssued>
|
14
|
+
</originInfo></mods>"
|
15
|
+
smods_rec = Stanford::Mods::Record.new
|
16
|
+
smods_rec.from_str(m)
|
17
|
+
expect(smods_rec.pub_dates).to eq(['1906', '1904', '1904'])
|
20
18
|
end
|
21
19
|
end
|
22
20
|
|
23
21
|
context "pub_date" do
|
24
|
-
|
25
|
-
@smods_rec = Stanford::Mods::Record.new
|
26
|
-
end
|
22
|
+
let(:smods_rec) { Stanford::Mods::Record.new }
|
27
23
|
|
28
|
-
it "
|
29
|
-
m = "<mods #{
|
30
|
-
|
31
|
-
|
32
|
-
|
33
|
-
expect(
|
34
|
-
end
|
35
|
-
it "
|
36
|
-
m = "<mods #{
|
37
|
-
|
38
|
-
|
39
|
-
|
40
|
-
expect(
|
41
|
-
end
|
42
|
-
it "
|
43
|
-
m = "<mods #{
|
44
|
-
|
45
|
-
|
46
|
-
|
47
|
-
expect(
|
48
|
-
end
|
49
|
-
it '
|
50
|
-
m = "<mods #{
|
51
|
-
|
52
|
-
|
53
|
-
|
54
|
-
expect(
|
55
|
-
end
|
56
|
-
it '
|
57
|
-
m = "<mods #{
|
58
|
-
|
59
|
-
|
60
|
-
|
61
|
-
|
62
|
-
|
63
|
-
|
64
|
-
|
65
|
-
|
66
|
-
|
67
|
-
m
|
68
|
-
|
69
|
-
|
70
|
-
|
71
|
-
|
72
|
-
|
73
|
-
|
74
|
-
|
75
|
-
|
76
|
-
|
77
|
-
|
78
|
-
|
79
|
-
|
80
|
-
|
81
|
-
|
82
|
-
expect(
|
83
|
-
expect(
|
84
|
-
expect(
|
85
|
-
|
86
|
-
|
87
|
-
|
88
|
-
|
89
|
-
|
90
|
-
|
91
|
-
|
92
|
-
|
93
|
-
|
94
|
-
|
95
|
-
|
96
|
-
|
97
|
-
|
98
|
-
|
99
|
-
|
100
|
-
|
101
|
-
|
102
|
-
|
103
|
-
expect(
|
104
|
-
|
105
|
-
|
106
|
-
|
107
|
-
|
108
|
-
|
109
|
-
|
110
|
-
expect(
|
111
|
-
expect(
|
112
|
-
expect(
|
113
|
-
end
|
114
|
-
it '
|
115
|
-
m = "<mods #{
|
116
|
-
|
117
|
-
|
118
|
-
|
119
|
-
expect(
|
120
|
-
expect(
|
121
|
-
|
122
|
-
|
123
|
-
|
124
|
-
|
125
|
-
|
126
|
-
|
127
|
-
|
128
|
-
expect(
|
24
|
+
it "uses dateCreated if no dateIssued" do
|
25
|
+
m = "<mods #{ns_decl}><originInfo>
|
26
|
+
<dateCreated>1904</dateCreated>
|
27
|
+
</originInfo></mods>"
|
28
|
+
smods_rec.from_str(m)
|
29
|
+
expect(smods_rec.pub_date).to eq('1904')
|
30
|
+
end
|
31
|
+
it "parses date" do
|
32
|
+
m = "<mods #{ns_decl}><originInfo>
|
33
|
+
<dateCreated>Aug. 3rd, 1886</dateCreated>
|
34
|
+
</originInfo></mods>"
|
35
|
+
smods_rec.from_str(m)
|
36
|
+
expect(smods_rec.pub_date).to eq('1886')
|
37
|
+
end
|
38
|
+
it "ignores question marks and square brackets" do
|
39
|
+
m = "<mods #{ns_decl}><originInfo>
|
40
|
+
<dateCreated>Aug. 3rd, [18]86?</dateCreated>
|
41
|
+
</originInfo></mods>"
|
42
|
+
smods_rec.from_str(m)
|
43
|
+
expect(smods_rec.pub_date).to eq('1886')
|
44
|
+
end
|
45
|
+
it 'handles an s after the decade' do
|
46
|
+
m = "<mods #{ns_decl}><originInfo>
|
47
|
+
<dateCreated>early 1890s</dateCreated>
|
48
|
+
</originInfo></mods>"
|
49
|
+
smods_rec.from_str(m)
|
50
|
+
expect(smods_rec.pub_date).to eq('1890')
|
51
|
+
end
|
52
|
+
it 'chooses date ending with CE if there are other choices' do
|
53
|
+
m = "<mods #{ns_decl}><originInfo>
|
54
|
+
<dateIssued>7192 AM (li-Adam) / 1684 CE</dateIssued>
|
55
|
+
</originInfo></mods>"
|
56
|
+
smods_rec.from_str(m)
|
57
|
+
expect(smods_rec.pub_date).to eq('1684')
|
58
|
+
end
|
59
|
+
it 'handles 1865-6 (hyphenated range dates)' do
|
60
|
+
m = "<mods #{ns_decl}><originInfo>
|
61
|
+
<dateIssued>1282 AH / 1865-6 CE</dateIssued>
|
62
|
+
</originInfo></mods>"
|
63
|
+
smods_rec.from_str(m)
|
64
|
+
expect(smods_rec.pub_date).to eq('1865')
|
65
|
+
end
|
66
|
+
it 'takes first occurring 4 digit date in string' do
|
67
|
+
m = "<mods #{ns_decl}><originInfo>
|
68
|
+
<dateCreated>Text dated June 4, 1594; miniatures added by 1596</dateCreated>
|
69
|
+
</originInfo></mods>"
|
70
|
+
smods_rec.from_str(m)
|
71
|
+
expect(smods_rec.pub_date).to eq('1594')
|
72
|
+
end
|
73
|
+
it '3 digit BC dates are "B.C." in pub_date_facet and negative in pub_date_sort, pub_year, and pub_date' do
|
74
|
+
m = "<mods #{ns_decl}><originInfo>
|
75
|
+
<dateCreated>300 B.C.</dateCreated>
|
76
|
+
</originInfo></mods>"
|
77
|
+
smods_rec.from_str(m)
|
78
|
+
expect(smods_rec.pub_year).to eq('-700')
|
79
|
+
expect(smods_rec.pub_date).to eq('-700')
|
80
|
+
expect(smods_rec.pub_date_sort).to eq('-700')
|
81
|
+
expect(smods_rec.pub_date_facet).to eq('300 B.C.')
|
82
|
+
end
|
83
|
+
it '"19th CE" becomes 1800 for pub_date_sort, 19th centruy for pub_date_facet' do
|
84
|
+
m = "<mods #{ns_decl}><originInfo>
|
85
|
+
<dateIssued>13th century AH / 19th CE</dateIssued>
|
86
|
+
</originInfo></mods>"
|
87
|
+
smods_rec.from_str(m)
|
88
|
+
expect(smods_rec.pub_date_facet).to eq('19th century')
|
89
|
+
expect(smods_rec.pub_date_sort).to eq('1800')
|
90
|
+
expect(smods_rec.pub_date).to eq('18--')
|
91
|
+
end
|
92
|
+
it 'takes first of multiple CE dates' do
|
93
|
+
m = "<mods #{ns_decl}><originInfo>
|
94
|
+
<dateIssued>6 Dhu al-Hijjah 923 AH / 1517 CE -- 7 Rabi I 924 AH / 1518 CE</dateIssued>
|
95
|
+
</originInfo></mods>"
|
96
|
+
smods_rec.from_str(m)
|
97
|
+
expect(smods_rec.pub_date).to eq('1517')
|
98
|
+
expect(smods_rec.pub_date_sort).to eq('1517')
|
99
|
+
expect(smods_rec.pub_date_facet).to eq('1517')
|
100
|
+
end
|
101
|
+
it 'handles "Late 14th or early 15th century CE" from walters' do
|
102
|
+
m = "<mods #{ns_decl}><originInfo>
|
103
|
+
<dateIssued>Late 14th or early 15th century CE</dateIssued>
|
104
|
+
</originInfo></mods>"
|
105
|
+
smods_rec.from_str(m)
|
106
|
+
expect(smods_rec.pub_date).to eq('14--')
|
107
|
+
expect(smods_rec.pub_date_sort).to eq('1400')
|
108
|
+
expect(smods_rec.pub_date_facet).to eq('15th century')
|
109
|
+
end
|
110
|
+
it 'explicit 3 digit dates have correct pub_date_sort and pub_date_facet' do
|
111
|
+
m = "<mods #{ns_decl}><originInfo>
|
112
|
+
<dateIssued>966 CE</dateIssued>
|
113
|
+
</originInfo></mods>"
|
114
|
+
smods_rec.from_str(m)
|
115
|
+
expect(smods_rec.pub_date).to eq('966')
|
116
|
+
expect(smods_rec.pub_date_sort).to eq('0966')
|
117
|
+
expect(smods_rec.pub_date_facet).to eq('966')
|
118
|
+
end
|
119
|
+
it 'single digit century dates have correct pub_date_sort and pub_date_facet' do
|
120
|
+
m = "<mods #{ns_decl}><originInfo>
|
121
|
+
<dateIssued>3rd century AH / 9th CE</dateIssued>
|
122
|
+
</originInfo></mods>"
|
123
|
+
smods_rec.from_str(m)
|
124
|
+
expect(smods_rec.pub_date).to eq('8--')
|
125
|
+
expect(smods_rec.pub_date_sort).to eq('0800')
|
126
|
+
expect(smods_rec.pub_date_facet).to eq('9th century')
|
127
|
+
end
|
128
|
+
it 'uses dateIssued without marc encoding for pub_date_display and the one with marc encoding for indexing, sorting and faceting' do
|
129
|
+
m = "<mods #{ns_decl}><originInfo>
|
130
|
+
<dateIssued>[186-?]</dateIssued><dateIssued encoding=\"marc\">1860</dateIssued>
|
131
|
+
</originInfo></mods>"
|
132
|
+
smods_rec.from_str(m)
|
133
|
+
expect(smods_rec.pub_date_display).to eq('[186-?]')
|
134
|
+
expect(smods_rec.pub_date).to eq('1860')
|
135
|
+
expect(smods_rec.pub_date_sort).to eq('1860')
|
136
|
+
expect(smods_rec.pub_date_facet).to eq('1860')
|
137
|
+
end
|
138
|
+
it 'uses dateIssued without marc encoding for pub_date_display and the one with marc encoding for indexing, sorting and faceting' do
|
139
|
+
m = "<mods #{ns_decl}><originInfo>
|
140
|
+
<dateIssued>1860?]</dateIssued><dateIssued encoding=\"marc\">186?</dateIssued>
|
141
|
+
</originInfo></mods>"
|
142
|
+
smods_rec.from_str(m)
|
143
|
+
expect(smods_rec.pub_date_display).to eq('1860?]')
|
144
|
+
expect(smods_rec.pub_date).to eq('1860')
|
145
|
+
expect(smods_rec.pub_date_sort).to eq('1860')
|
146
|
+
expect(smods_rec.pub_date_facet).to eq('1860')
|
129
147
|
end
|
130
148
|
end # pub_date
|
131
149
|
|
132
150
|
context "dates with u notation (198u, 19uu)" do
|
133
151
|
context "single digit u notation (198u)" do
|
134
|
-
|
135
|
-
m = "<mods #{
|
152
|
+
let(:smods_rec) do
|
153
|
+
m = "<mods #{ns_decl}>
|
136
154
|
<originInfo>
|
137
155
|
<dateIssued encoding=\"marc\" point=\"start\" keyDate=\"yes\">198u</dateIssued>
|
138
156
|
<dateIssued encoding=\"marc\" point=\"end\">9999</dateIssued>
|
139
157
|
</originInfo></mods>"
|
140
|
-
|
141
|
-
|
142
|
-
|
143
|
-
it "get_u_year recognizes notation" do
|
144
|
-
dates = ["198u", "9999"]
|
145
|
-
uDate = @smods_rec.get_u_year dates
|
146
|
-
expect(uDate).to eql("1980")
|
158
|
+
smr = Stanford::Mods::Record.new
|
159
|
+
smr.from_str(m)
|
160
|
+
smr
|
147
161
|
end
|
148
162
|
it 'pub_date: 198u = 1980' do
|
149
|
-
expect(
|
163
|
+
expect(smods_rec.pub_date).to eq('1980')
|
150
164
|
end
|
151
165
|
it "pub_date_sort: 198u = 1980" do
|
152
|
-
expect(
|
166
|
+
expect(smods_rec.pub_date_sort).to eq('1980')
|
153
167
|
end
|
154
168
|
it "pub_date_facet: 198u = 1980" do
|
155
|
-
expect(
|
169
|
+
expect(smods_rec.pub_date_facet).to eq('1980')
|
156
170
|
end
|
157
171
|
end
|
158
172
|
context "double digit u notation (19uu)" do
|
159
|
-
|
160
|
-
m = "<mods #{
|
173
|
+
let(:smods_rec) do
|
174
|
+
m = "<mods #{ns_decl}>
|
161
175
|
<originInfo>
|
162
176
|
<dateIssued encoding=\"marc\" point=\"start\" keyDate=\"yes\">19uu</dateIssued>
|
163
177
|
<dateIssued encoding=\"marc\" point=\"end\">9999</dateIssued>
|
164
178
|
</originInfo></mods>"
|
165
|
-
|
166
|
-
|
167
|
-
|
168
|
-
it "get_u_year recognizes notation" do
|
169
|
-
dates = ["19uu", "9999"]
|
170
|
-
uDate = @smods_rec.get_u_year dates
|
171
|
-
expect(uDate).to eql("19--")
|
179
|
+
smr = Stanford::Mods::Record.new
|
180
|
+
smr.from_str(m)
|
181
|
+
smr
|
172
182
|
end
|
173
183
|
it 'pub_date: 19uu = 19--' do
|
174
|
-
expect(
|
184
|
+
expect(smods_rec.pub_date).to eq('19--')
|
175
185
|
end
|
176
186
|
it "pub_date_sort: 19uu = 1900" do
|
177
|
-
expect(
|
187
|
+
expect(smods_rec.pub_date_sort).to eq('1900')
|
178
188
|
end
|
179
189
|
it "pub_date_facet: 19uu = 20th century" do
|
180
|
-
expect(
|
190
|
+
expect(smods_rec.pub_date_facet).to eq('20th century')
|
181
191
|
end
|
182
192
|
end
|
183
193
|
end # u notation
|
184
194
|
|
185
|
-
|
186
195
|
context 'pub_date_sort' do
|
187
|
-
|
188
|
-
|
189
|
-
|
190
|
-
|
191
|
-
|
192
|
-
|
193
|
-
|
194
|
-
|
195
|
-
|
196
|
-
|
197
|
-
|
198
|
-
|
199
|
-
|
200
|
-
|
201
|
-
|
202
|
-
|
203
|
-
allow(@smods_rec).to receive(:pub_date).and_return('16--')
|
204
|
-
expect(@smods_rec.pub_date_sort).to eq('1600')
|
205
|
-
end
|
206
|
-
it 'should work on 3 digit century dates' do
|
207
|
-
allow(@smods_rec).to receive(:pub_date).and_return('9--')
|
208
|
-
expect(@smods_rec.pub_date_sort).to eq('0900')
|
196
|
+
let(:smods_rec) { Stanford::Mods::Record.new }
|
197
|
+
it 'four digits' do
|
198
|
+
allow(smods_rec).to receive(:pub_date).and_return('1945')
|
199
|
+
expect(smods_rec.pub_date_sort).to eq('1945')
|
200
|
+
end
|
201
|
+
it '3 digits' do
|
202
|
+
allow(smods_rec).to receive(:pub_date).and_return('945')
|
203
|
+
expect(smods_rec.pub_date_sort).to eq('0945')
|
204
|
+
end
|
205
|
+
it '16--' do
|
206
|
+
allow(smods_rec).to receive(:pub_date).and_return('16--')
|
207
|
+
expect(smods_rec.pub_date_sort).to eq('1600')
|
208
|
+
end
|
209
|
+
it '9--' do
|
210
|
+
allow(smods_rec).to receive(:pub_date).and_return('9--')
|
211
|
+
expect(smods_rec.pub_date_sort).to eq('0900')
|
209
212
|
end
|
210
213
|
end
|
211
214
|
|
data/spec/spec_helper.rb
CHANGED
@@ -1,10 +1,21 @@
|
|
1
1
|
require 'coveralls'
|
2
2
|
Coveralls.wear!
|
3
3
|
|
4
|
-
$LOAD_PATH.unshift(File.join(File.dirname(__FILE__), '..', 'lib'))
|
5
|
-
$LOAD_PATH.unshift(File.dirname(__FILE__))
|
6
|
-
|
7
4
|
require 'stanford-mods'
|
8
5
|
|
9
|
-
# RSpec
|
10
|
-
|
6
|
+
# See http://rubydoc.info/gems/rspec-core/RSpec/Core/Configuration
|
7
|
+
RSpec.configure do |config|
|
8
|
+
config.expect_with :rspec do |expectations|
|
9
|
+
expectations.include_chain_clauses_in_custom_matcher_descriptions = true
|
10
|
+
end
|
11
|
+
|
12
|
+
config.mock_with :rspec do |mocks|
|
13
|
+
mocks.verify_partial_doubles = true
|
14
|
+
end
|
15
|
+
|
16
|
+
if config.files_to_run.one?
|
17
|
+
config.default_formatter = 'doc'
|
18
|
+
end
|
19
|
+
|
20
|
+
#config.order = :random
|
21
|
+
end
|
data/stanford-mods.gemspec
CHANGED
@@ -17,6 +17,8 @@ Gem::Specification.new do |gem|
|
|
17
17
|
gem.require_paths = ["lib"]
|
18
18
|
|
19
19
|
gem.add_dependency 'mods', '~> 2.0.2'
|
20
|
+
# active_support for .ordinalize, eg. 1 -> 1st, 2 -> 2nd for centuries
|
21
|
+
gem.add_dependency 'activesupport'
|
20
22
|
|
21
23
|
# Runtime dependencies
|
22
24
|
# gem.add_runtime_dependency 'nokogiri'
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: stanford-mods
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 1.3.
|
4
|
+
version: 1.3.4
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Naomi Dushay
|
@@ -9,7 +9,7 @@ authors:
|
|
9
9
|
autorequire:
|
10
10
|
bindir: bin
|
11
11
|
cert_chain: []
|
12
|
-
date:
|
12
|
+
date: 2016-01-05 00:00:00.000000000 Z
|
13
13
|
dependencies:
|
14
14
|
- !ruby/object:Gem::Dependency
|
15
15
|
name: mods
|
@@ -25,6 +25,20 @@ dependencies:
|
|
25
25
|
- - "~>"
|
26
26
|
- !ruby/object:Gem::Version
|
27
27
|
version: 2.0.2
|
28
|
+
- !ruby/object:Gem::Dependency
|
29
|
+
name: activesupport
|
30
|
+
requirement: !ruby/object:Gem::Requirement
|
31
|
+
requirements:
|
32
|
+
- - ">="
|
33
|
+
- !ruby/object:Gem::Version
|
34
|
+
version: '0'
|
35
|
+
type: :runtime
|
36
|
+
prerelease: false
|
37
|
+
version_requirements: !ruby/object:Gem::Requirement
|
38
|
+
requirements:
|
39
|
+
- - ">="
|
40
|
+
- !ruby/object:Gem::Version
|
41
|
+
version: '0'
|
28
42
|
- !ruby/object:Gem::Dependency
|
29
43
|
name: rake
|
30
44
|
requirement: !ruby/object:Gem::Requirement
|
@@ -106,14 +120,20 @@ files:
|
|
106
120
|
- Rakefile
|
107
121
|
- config/mappings_hash.rb
|
108
122
|
- lib/stanford-mods.rb
|
123
|
+
- lib/stanford-mods/date_parsing.rb
|
109
124
|
- lib/stanford-mods/geo_spatial.rb
|
110
125
|
- lib/stanford-mods/name.rb
|
126
|
+
- lib/stanford-mods/origin_info.rb
|
111
127
|
- lib/stanford-mods/physical_location.rb
|
112
128
|
- lib/stanford-mods/searchworks.rb
|
113
129
|
- lib/stanford-mods/searchworks_languages.rb
|
130
|
+
- lib/stanford-mods/searchworks_subjects.rb
|
114
131
|
- lib/stanford-mods/version.rb
|
132
|
+
- spec/date_parsing_spec.rb
|
133
|
+
- spec/fixtures/spotlight_pub_date_data.rb
|
115
134
|
- spec/geo_spatial_spec.rb
|
116
135
|
- spec/name_spec.rb
|
136
|
+
- spec/origin_info_spec.rb
|
117
137
|
- spec/physical_location_spec.rb
|
118
138
|
- spec/searchworks_basic_spec.rb
|
119
139
|
- spec/searchworks_format_spec.rb
|
@@ -148,8 +168,11 @@ signing_key:
|
|
148
168
|
specification_version: 4
|
149
169
|
summary: Stanford specific wrangling of MODS metadata
|
150
170
|
test_files:
|
171
|
+
- spec/date_parsing_spec.rb
|
172
|
+
- spec/fixtures/spotlight_pub_date_data.rb
|
151
173
|
- spec/geo_spatial_spec.rb
|
152
174
|
- spec/name_spec.rb
|
175
|
+
- spec/origin_info_spec.rb
|
153
176
|
- spec/physical_location_spec.rb
|
154
177
|
- spec/searchworks_basic_spec.rb
|
155
178
|
- spec/searchworks_format_spec.rb
|