harvestdor 0.0.13 → 0.0.14

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: 3036c2f23661946012d909179b67effa433c2dd9
4
- data.tar.gz: 149e27d97f4c9d48ca46772d550537aca8a55362
3
+ metadata.gz: dedc072f7e987057ecc62cd6fb174c9733ce414b
4
+ data.tar.gz: 452697c49d6328ffbb334666e3d9f8892ddf5e3e
5
5
  SHA512:
6
- metadata.gz: 4406cda66fcb2a726564ba912c60364433b9209ac03218fc2fbf70200fe679f8c964fa416abc0f163eecdabc4db51ad7b57de148f38cf4773b84a4de3fde854e
7
- data.tar.gz: b18c647163a3970ae02cb3ee79ba0df04f4d5920ccbc5740a83b19b2958297ce2b1e0c761128293ae7bbb51633e419a6689b839f33f34f602a5292b2a078a4db
6
+ metadata.gz: b544187441fff729f33c33bcda20f4a9d9f3c827fad46dab343d8fc4a9ac71a60a8f9a838a4c767f6e29ce23ef5e54cec843c239f1c416690fa1a12c0eeab44a
7
+ data.tar.gz: adf1836583a2b617321ff6b75649dd391161e31b004b2bc93bf513785ec11f6cc84570fd6d1e9bb661af379b86a291d2053c326cdac612fa28f57052ed11ffcf
data/.coveralls.yml ADDED
@@ -0,0 +1 @@
1
+ service_name: travis-ci
data/.travis.yml CHANGED
@@ -1,12 +1,10 @@
1
1
  language: ruby
2
2
  script: rake rspec
3
3
  rvm:
4
- - ruby-head
5
4
  - 2.1.1
6
5
  - 2.0.0
7
6
  - 1.9.3
8
7
  - jruby-19mode # JRuby in 1.9 mode
9
- # - jruby-head
10
8
  notifications:
11
9
  email:
12
10
  - ndushay@stanford.edu
data/Gemfile CHANGED
@@ -2,3 +2,7 @@ source 'https://rubygems.org'
2
2
 
3
3
  # See harvestdor.gemspec for this gem's dependencies
4
4
  gemspec
5
+
6
+ group :test do
7
+ gem 'coveralls', require: false
8
+ end
data/README.rdoc CHANGED
@@ -1,5 +1,5 @@
1
1
  = Harvestdor
2
- {<img src="https://travis-ci.org/sul-dlss/harvestdor.svg?branch=master" alt="Build Status" />}[https://travis-ci.org/sul-dlss/harvestdor] {<img src="https://gemnasium.com/sul-dlss/harvestdor.svg" alt="Dependency Status" />}[https://gemnasium.com/sul-dlss/harvestdor]
2
+ {<img src="https://travis-ci.org/sul-dlss/harvestdor.svg?branch=master" alt="Build Status" />}[https://travis-ci.org/sul-dlss/harvestdor] {<img src="https://coveralls.io/repos/sul-dlss/harvestdor/badge.png" alt="Coverage Status" />}[https://coveralls.io/r/sul-dlss/harvestdor] {<img src="https://gemnasium.com/sul-dlss/harvestdor.svg" alt="Dependency Status" />}[https://gemnasium.com/sul-dlss/harvestdor] {<img src="https://badge.fury.io/rb/harvestdor.svg" alt="Gem Version" />}[http://badge.fury.io/rb/harvestdor]
3
3
 
4
4
  A Gem to harvest metadata from DOR.
5
5
 
@@ -148,6 +148,7 @@ persist this information with your app for incremental harvests
148
148
 
149
149
  == Releases
150
150
 
151
+ * <b>0.0.14</b> Bug fix for compatibility with jruby
151
152
  * <b>0.0.13</b> Updated to work with Faraday 0.9, releases via rubygems instead of sul-gems
152
153
  * <b>0.0.11</b> better error handling, and better testing for errors
153
154
  * <b>0.0.10</b> tweak specs to test that unnec fetching isn't done.
data/harvestdor.gemspec CHANGED
@@ -34,8 +34,8 @@ Gem::Specification.new do |gem|
34
34
  gem.add_development_dependency "yard"
35
35
  # tests
36
36
  gem.add_development_dependency 'rspec'
37
- gem.add_development_dependency 'simplecov'
38
- gem.add_development_dependency 'simplecov-rcov'
37
+ gem.add_development_dependency 'vcr'
38
+ gem.add_development_dependency 'webmock'
39
39
  # gem.add_development_dependency 'ruby-debug19'
40
40
 
41
41
  end
@@ -111,7 +111,7 @@ module Harvestdor
111
111
  pub_xml_ng_doc = pub_xml(object, purl_url)
112
112
  begin
113
113
  # preserve namespaces, etc for the node
114
- ng_doc = Nokogiri::XML(pub_xml_ng_doc.root.xpath('/publicObject/dc:dc', {'dc' => Harvestdor::OAI_DC_NAMESPACE}).to_xml)
114
+ ng_doc = Nokogiri::XML(pub_xml_ng_doc.root.xpath('/publicObject/dc:dc', {'dc' => Harvestdor::OAI_DC_NAMESPACE}).to_xml(:encoding => 'utf-8'))
115
115
  raise Harvestdor::Errors::MissingDC.new(object.inspect) if !ng_doc || ng_doc.children.empty?
116
116
  ng_doc
117
117
  rescue
@@ -1,3 +1,3 @@
1
1
  module Harvestdor
2
- VERSION = "0.0.13"
2
+ VERSION = "0.0.14"
3
3
  end
@@ -0,0 +1,98 @@
1
+ ---
2
+ http_interactions:
3
+ - request:
4
+ method: get
5
+ uri: http://purl-test.stanford.edu/bb375wb8869.xml
6
+ body:
7
+ encoding: US-ASCII
8
+ string: ''
9
+ headers:
10
+ Accept-Encoding:
11
+ - gzip;q=1.0,deflate;q=0.6,identity;q=0.3
12
+ Accept:
13
+ - "*/*"
14
+ User-Agent:
15
+ - Ruby
16
+ response:
17
+ status:
18
+ code: 200
19
+ message: ''
20
+ headers:
21
+ Date:
22
+ - Mon, 28 Jul 2014 22:23:37 GMT
23
+ Server:
24
+ - Apache/2.2.15 (Red Hat)
25
+ X-Powered-By:
26
+ - Phusion Passenger (mod_rails/mod_rack) 3.0.19
27
+ X-Ua-Compatible:
28
+ - IE=Edge
29
+ Etag:
30
+ - '"c7e2680ca3a1d497d6e46962eac8bb2d"'
31
+ Cache-Control:
32
+ - max-age=0, private, must-revalidate
33
+ X-Request-Id:
34
+ - 510a314d2ea688c5bfaff1d536204974
35
+ X-Runtime:
36
+ - '0.012861'
37
+ Status:
38
+ - '200'
39
+ Content-Length:
40
+ - '4349'
41
+ Content-Type:
42
+ - application/xml; charset=utf-8
43
+ body:
44
+ encoding: UTF-8
45
+ string: "<publicObject id=\"druid:bb375wb8869\" published=\"2012-01-26T16:39:44-08:00\">\n
46
+ \ <identityMetadata>\n <adminPolicy>druid:xs835jp8197</adminPolicy>\n <objectCreator>DOR</objectCreator>\n
47
+ \ <objectId>druid:bb375wb8869</objectId>\n <objectLabel>Kitai_bb375wb8869</objectLabel>\n
48
+ \ <objectType>item</objectType>\n <otherId name=\"uuid\">15ffc446-25f5-11e1-b6f0-0016034322e7</otherId>\n
49
+ \ <tag>Project : Kitai</tag>\n </identityMetadata>\n <contentMetadata
50
+ type=\"image\" objectId=\"bb375wb8869\">\n <resource type=\"image\" sequence=\"1\"
51
+ id=\"bb375wb8869_1\">\n <label>kitai_4</label>\n <file mimetype=\"image/jp2\"
52
+ format=\"JPEG2000\" size=\"8116154\" id=\"36105005868828_05_0001.jp2\">\n
53
+ \ <imageData height=\"6007\" width=\"7175\"/>\n <attr name=\"representation\">uncropped</attr>\n
54
+ \ </file>\n </resource>\n <resource type=\"image\" sequence=\"2\"
55
+ id=\"bb375wb8869_2\">\n <label>kitai_4</label>\n <file mimetype=\"image/jp2\"
56
+ format=\"JPEG2000\" size=\"8141860\" id=\"36105005868828_05_0002.jp2\">\n
57
+ \ <imageData height=\"6017\" width=\"7174\"/>\n <attr name=\"representation\">uncropped</attr>\n
58
+ \ </file>\n </resource>\n </contentMetadata>\n <rightsMetadata>\n
59
+ \ <access type=\"discover\">\n <machine>\n <world/>\n </machine>\n
60
+ \ </access>\n <access type=\"read\">\n <machine>\n <group>stanford</group>\n
61
+ \ </machine>\n </access>\n </rightsMetadata>\n <rdf:RDF xmlns:fedora-model=\"info:fedora/fedora-system:def/model#\"
62
+ xmlns:rdf=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#\" xmlns:fedora=\"info:fedora/fedora-system:def/relations-external#\"
63
+ xmlns:hydra=\"http://projecthydra.org/ns/relations#\">\n <rdf:Description
64
+ rdf:about=\"info:fedora/druid:bb375wb8869\">\n </rdf:Description>\n </rdf:RDF>\n
65
+ \ <oai_dc:dc xmlns:dc=\"http://purl.org/dc/elements/1.1/\" xmlns:srw_dc=\"info:srw/schema/1/dc-schema\"
66
+ xmlns:oai_dc=\"http://www.openarchives.org/OAI/2.0/oai_dc/\" xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\"
67
+ xsi:schemaLocation=\"http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd\">\n
68
+ \ <dc:title>Kita&#x12D;</dc:title>\n <dc:title>China</dc:title>\n <dc:creator>Soviet
69
+ Union Sovetskai&#x361;a Armii&#x361;a. General&#x2B9;ny&#x12D; shtab.</dc:creator>\n
70
+ \ <dc:type>map</dc:type>\n <dc:date>1968-9999</dc:date>\n <dc:publisher>General&#x2B9;ny&#x12D;
71
+ shtab</dc:publisher>\n <dc:language>rus</dc:language>\n <dc:format>maps
72
+ : col. ; 37 x 53 cm. or smaller.</dc:format>\n <dc:format>map</dc:format>\n
73
+ \ <dc:description>Military topographic maps.</dc:description>\n <dc:description>Relief
74
+ shown by contours and spot heights. Depths shown by contours and soundings
75
+ on some sheets.</dc:description>\n <dc:description>Name of jurisdiction[s]
76
+ covered by sheet quadrangle appears in upper margin at left, e.g.: Provint&#x361;sii&#x361;a
77
+ Guandun -- Avtonomny&#x12D; ra&#x12D;on Vnutrenni&#x361;ai&#x361;a Mongolii&#x361;a,
78
+ Provint&#x361;sii Kh&#x117;b&#x117;&#x12D;, Shan&#x2B9;si. May include name
79
+ and province of neighboring country.</dc:description>\n <dc:description>Common
80
+ title precedes jurisdiction name at upper left on each sheet.</dc:description>\n
81
+ \ <dc:description>Sheets individually subtitled and numbered (with roman-alphabet
82
+ lettering) in upper margin, e.g.: Ch&#x117;nma&#x12D;, E-49-III -- Sich&#x117;n,
83
+ K-50-XXXI.</dc:description>\n <dc:description>Ed./publication date in upper
84
+ border at right, e.g.: Izdanie 1970 g. -- Izdanie 1986 g. -- Izdanie 1991
85
+ g.</dc:description>\n <dc:description>Former security classification in
86
+ upper margin at right: Sekretno.</dc:description>\n <dc:description>Includes
87
+ notes, diagrams, compilation date[s], and sheet compiler's/editor's names
88
+ in lower margin.</dc:description>\n <dc:description>Blue sheet no. in upper
89
+ margin precedes black sheet no., e.g.: 05-49-01 -- 11-50-31.</dc:description>\n
90
+ \ <dc:description>Text and ancillary map, \"Skhema gruntov,\" on verso.</dc:description>\n
91
+ \ <dc:description>In Russian.</dc:description>\n <dc:subject/>\n <dc:coverage>Scale
92
+ 1:200,000. 1 cm. to 2 km.</dc:coverage>\n <dc:coverage>116.0,27.333333
93
+ \n117.0,27.333333 \n117.0,26.666667 \n116.0,26.666667 \n116.0,27.333333\n</dc:coverage>\n
94
+ \ <dc:coverage>Pulkovo 1942</dc:coverage>\n <dc:subject>Military maps</dc:subject>\n
95
+ \ <dc:subject/>\n <dc:coverage>China</dc:coverage>\n </oai_dc:dc>\n</publicObject>\n"
96
+ http_version:
97
+ recorded_at: Mon, 28 Jul 2014 22:23:37 GMT
98
+ recorded_with: VCR 2.9.2
@@ -0,0 +1,98 @@
1
+ ---
2
+ http_interactions:
3
+ - request:
4
+ method: get
5
+ uri: http://purl-test.stanford.edu/bb375wb8869.xml
6
+ body:
7
+ encoding: US-ASCII
8
+ string: ''
9
+ headers:
10
+ Accept-Encoding:
11
+ - gzip;q=1.0,deflate;q=0.6,identity;q=0.3
12
+ Accept:
13
+ - "*/*"
14
+ User-Agent:
15
+ - Ruby
16
+ response:
17
+ status:
18
+ code: 200
19
+ message: ''
20
+ headers:
21
+ Date:
22
+ - Mon, 28 Jul 2014 22:26:22 GMT
23
+ Server:
24
+ - Apache/2.2.15 (Red Hat)
25
+ X-Powered-By:
26
+ - Phusion Passenger (mod_rails/mod_rack) 3.0.19
27
+ X-Ua-Compatible:
28
+ - IE=Edge
29
+ Etag:
30
+ - '"c7e2680ca3a1d497d6e46962eac8bb2d"'
31
+ Cache-Control:
32
+ - max-age=0, private, must-revalidate
33
+ X-Request-Id:
34
+ - 494d94bd96cf627c1eff6e27957cddd6
35
+ X-Runtime:
36
+ - '0.011981'
37
+ Status:
38
+ - '200'
39
+ Content-Length:
40
+ - '4349'
41
+ Content-Type:
42
+ - application/xml; charset=utf-8
43
+ body:
44
+ encoding: UTF-8
45
+ string: "<publicObject id=\"druid:bb375wb8869\" published=\"2012-01-26T16:39:44-08:00\">\n
46
+ \ <identityMetadata>\n <adminPolicy>druid:xs835jp8197</adminPolicy>\n <objectCreator>DOR</objectCreator>\n
47
+ \ <objectId>druid:bb375wb8869</objectId>\n <objectLabel>Kitai_bb375wb8869</objectLabel>\n
48
+ \ <objectType>item</objectType>\n <otherId name=\"uuid\">15ffc446-25f5-11e1-b6f0-0016034322e7</otherId>\n
49
+ \ <tag>Project : Kitai</tag>\n </identityMetadata>\n <contentMetadata
50
+ type=\"image\" objectId=\"bb375wb8869\">\n <resource type=\"image\" sequence=\"1\"
51
+ id=\"bb375wb8869_1\">\n <label>kitai_4</label>\n <file mimetype=\"image/jp2\"
52
+ format=\"JPEG2000\" size=\"8116154\" id=\"36105005868828_05_0001.jp2\">\n
53
+ \ <imageData height=\"6007\" width=\"7175\"/>\n <attr name=\"representation\">uncropped</attr>\n
54
+ \ </file>\n </resource>\n <resource type=\"image\" sequence=\"2\"
55
+ id=\"bb375wb8869_2\">\n <label>kitai_4</label>\n <file mimetype=\"image/jp2\"
56
+ format=\"JPEG2000\" size=\"8141860\" id=\"36105005868828_05_0002.jp2\">\n
57
+ \ <imageData height=\"6017\" width=\"7174\"/>\n <attr name=\"representation\">uncropped</attr>\n
58
+ \ </file>\n </resource>\n </contentMetadata>\n <rightsMetadata>\n
59
+ \ <access type=\"discover\">\n <machine>\n <world/>\n </machine>\n
60
+ \ </access>\n <access type=\"read\">\n <machine>\n <group>stanford</group>\n
61
+ \ </machine>\n </access>\n </rightsMetadata>\n <rdf:RDF xmlns:fedora-model=\"info:fedora/fedora-system:def/model#\"
62
+ xmlns:rdf=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#\" xmlns:fedora=\"info:fedora/fedora-system:def/relations-external#\"
63
+ xmlns:hydra=\"http://projecthydra.org/ns/relations#\">\n <rdf:Description
64
+ rdf:about=\"info:fedora/druid:bb375wb8869\">\n </rdf:Description>\n </rdf:RDF>\n
65
+ \ <oai_dc:dc xmlns:dc=\"http://purl.org/dc/elements/1.1/\" xmlns:srw_dc=\"info:srw/schema/1/dc-schema\"
66
+ xmlns:oai_dc=\"http://www.openarchives.org/OAI/2.0/oai_dc/\" xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\"
67
+ xsi:schemaLocation=\"http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd\">\n
68
+ \ <dc:title>Kita&#x12D;</dc:title>\n <dc:title>China</dc:title>\n <dc:creator>Soviet
69
+ Union Sovetskai&#x361;a Armii&#x361;a. General&#x2B9;ny&#x12D; shtab.</dc:creator>\n
70
+ \ <dc:type>map</dc:type>\n <dc:date>1968-9999</dc:date>\n <dc:publisher>General&#x2B9;ny&#x12D;
71
+ shtab</dc:publisher>\n <dc:language>rus</dc:language>\n <dc:format>maps
72
+ : col. ; 37 x 53 cm. or smaller.</dc:format>\n <dc:format>map</dc:format>\n
73
+ \ <dc:description>Military topographic maps.</dc:description>\n <dc:description>Relief
74
+ shown by contours and spot heights. Depths shown by contours and soundings
75
+ on some sheets.</dc:description>\n <dc:description>Name of jurisdiction[s]
76
+ covered by sheet quadrangle appears in upper margin at left, e.g.: Provint&#x361;sii&#x361;a
77
+ Guandun -- Avtonomny&#x12D; ra&#x12D;on Vnutrenni&#x361;ai&#x361;a Mongolii&#x361;a,
78
+ Provint&#x361;sii Kh&#x117;b&#x117;&#x12D;, Shan&#x2B9;si. May include name
79
+ and province of neighboring country.</dc:description>\n <dc:description>Common
80
+ title precedes jurisdiction name at upper left on each sheet.</dc:description>\n
81
+ \ <dc:description>Sheets individually subtitled and numbered (with roman-alphabet
82
+ lettering) in upper margin, e.g.: Ch&#x117;nma&#x12D;, E-49-III -- Sich&#x117;n,
83
+ K-50-XXXI.</dc:description>\n <dc:description>Ed./publication date in upper
84
+ border at right, e.g.: Izdanie 1970 g. -- Izdanie 1986 g. -- Izdanie 1991
85
+ g.</dc:description>\n <dc:description>Former security classification in
86
+ upper margin at right: Sekretno.</dc:description>\n <dc:description>Includes
87
+ notes, diagrams, compilation date[s], and sheet compiler's/editor's names
88
+ in lower margin.</dc:description>\n <dc:description>Blue sheet no. in upper
89
+ margin precedes black sheet no., e.g.: 05-49-01 -- 11-50-31.</dc:description>\n
90
+ \ <dc:description>Text and ancillary map, \"Skhema gruntov,\" on verso.</dc:description>\n
91
+ \ <dc:description>In Russian.</dc:description>\n <dc:subject/>\n <dc:coverage>Scale
92
+ 1:200,000. 1 cm. to 2 km.</dc:coverage>\n <dc:coverage>116.0,27.333333
93
+ \n117.0,27.333333 \n117.0,26.666667 \n116.0,26.666667 \n116.0,27.333333\n</dc:coverage>\n
94
+ \ <dc:coverage>Pulkovo 1942</dc:coverage>\n <dc:subject>Military maps</dc:subject>\n
95
+ \ <dc:subject/>\n <dc:coverage>China</dc:coverage>\n </oai_dc:dc>\n</publicObject>\n"
96
+ http_version:
97
+ recorded_at: Mon, 28 Jul 2014 22:26:22 GMT
98
+ recorded_with: VCR 2.9.2
@@ -0,0 +1,42 @@
1
+ ---
2
+ http_interactions:
3
+ - request:
4
+ method: get
5
+ uri: https://dor-oaiprovider-test.stanford.edu/oai?metadataPrefix=mods&set=is_governed_by_hy787xj5878&verb=ListIdentifiers
6
+ body:
7
+ encoding: US-ASCII
8
+ string: ''
9
+ headers:
10
+ User-Agent:
11
+ - Faraday v0.9.0
12
+ Accept-Encoding:
13
+ - gzip;q=1.0,deflate;q=0.6,identity;q=0.3
14
+ Accept:
15
+ - "*/*"
16
+ response:
17
+ status:
18
+ code: 200
19
+ message: OK
20
+ headers:
21
+ Date:
22
+ - Mon, 28 Jul 2014 22:02:50 GMT
23
+ Server:
24
+ - Apache/2.2.15 (Red Hat)
25
+ X-Content-Type-Options:
26
+ - nosniff
27
+ X-Powered-By:
28
+ - Phusion Passenger 4.0.10
29
+ Content-Length:
30
+ - '1402'
31
+ Status:
32
+ - 200 OK
33
+ Content-Type:
34
+ - application/xml;charset=utf-8
35
+ body:
36
+ encoding: UTF-8
37
+ string: |-
38
+ <?xml version="1.0" encoding="UTF-8" ?>
39
+ <?xml-stylesheet type="text/xsl" href="https://dor-oaiprovider-test.stanford.edu/oai2.xsl" ?><OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd"><responseDate>2014-07-28T22:02:50Z</responseDate><request metadataPrefix="mods" set="is_governed_by_hy787xj5878" verb="ListIdentifiers">https://dor-oaiprovider-test.stanford.edu/oai</request><ListIdentifiers><header><identifier>oai:searchworks.stanford.edu/druid:jt959wc5586</identifier><datestamp>2012-02-27T23:11:50Z</datestamp><setSpec>is_member_of_jh957jy1101</setSpec><setSpec>is_member_of_collection_jh957jy1101</setSpec><setSpec>is_governed_by_hy787xj5878</setSpec></header><header><identifier>oai:searchworks.stanford.edu/druid:pd817cv1860</identifier><datestamp>2012-03-28T21:25:56Z</datestamp><setSpec>is_member_of_jh957jy1101</setSpec><setSpec>is_member_of_collection_jh957jy1101</setSpec><setSpec>is_governed_by_hy787xj5878</setSpec></header><header><identifier>oai:searchworks.stanford.edu/druid:qk487nd2690</identifier><datestamp>2013-05-22T23:38:57Z</datestamp><setSpec>is_member_of_jh957jy1101</setSpec><setSpec>is_member_of_collection_jh957jy1101</setSpec><setSpec>is_governed_by_hy787xj5878</setSpec></header></ListIdentifiers></OAI-PMH>
40
+ http_version:
41
+ recorded_at: Mon, 28 Jul 2014 22:02:50 GMT
42
+ recorded_with: VCR 2.9.2