cirneco 0.9.13 → 0.9.14

data/spec/file_utils_spec.rb

@@ -0,0 +1,203 @@
+require 'spec_helper'
+
+describe Cirneco::DataCenter, vcr: true, :order => :defined do
+  let(:prefix) { ENV['PREFIX'] }
+  let(:username) { ENV['MDS_USERNAME'] }
+  let(:password) { ENV['MDS_PASSWORD'] }
+  let(:source_dir) { "/spec/fixtures/" }
+  let(:bibliography) { "spec/fixtures/references.yaml" }
+  let(:options) { { username: username,
+                    password: password,
+                    sandbox: true,
+                    build_dir: source_dir,
+                    source_dir: source_dir } }
+
+  subject { Cirneco::DataCenter.new(prefix: prefix,
+                                    username: username,
+                                    password: password) }
+
+  context "accession_number" do
+    it 'should get existing' do
+      filepath = fixture_path + 'cool-dois.html.md'
+      response = subject.get_accession_number(filepath)
+      expect(response).to eq(123)
+    end
+
+    it 'should get existing for folder' do
+      response = subject.get_all_accession_numbers(fixture_path)
+      expect(response).to eq([123, 124])
+    end
+
+    it 'should update for file' do
+      filepath = fixture_path + 'cool-dois.html.md'
+      response = subject.update_accession_number(filepath)
+      expect(response).to eq("Accession number MS-123 not changed for cool-dois.html.md")
+    end
+
+    it 'should update for all files' do
+      response = subject.update_all_accession_numbers(fixture_path, opt_in: true)
+      expect(response).to eq(["Accession number MS-124 not changed for cool-dois-minted.html.md", "File cool-dois-no-accession-number.html.md ignored: no empty accession_number", "Accession number MS-123 not changed for cool-dois.html.md"])
+    end
+  end
+
+  context "mint and hide DOIs" do
+    it 'get urls for works' do
+      filepath = fixture_path + 'index.html'
+      response = subject.get_urls_for_works(filepath)
+      expect(response.length).to eq(2)
+    end
+
+    it 'should mint for url' do
+      filepath = fixture_path + 'cool-dois/index.html'
+      response = subject.mint_doi_for_url(filepath, options)
+      expect(response).to eq("DOI 10.5438/0000-03vc minted for cool-dois.html.md")
+    end
+
+    it 'should hide for url' do
+      filepath = fixture_path + 'cool-dois-minted/index.html'
+      response = subject.hide_doi_for_url(filepath, options)
+      expect(response).to eq("DOI 10.5438/55e5-t5c0 hidden for cool-dois-minted.html.md")
+    end
+
+    it 'should mint and hide for url' do
+      filepath = fixture_path + 'cool-dois/index.html'
+      response = subject.mint_and_hide_doi_for_url(filepath, options)
+      expect(response).to eq("DOI 10.5438/0000-03vc minted and hidden for cool-dois.html.md")
+    end
+
+    it 'should mint for all urls' do
+      filepath = fixture_path + 'index.html'
+      response = subject.mint_dois_for_all_urls(filepath, options)
+      expect(response).to eq("DOI 10.5438/0000-03vc minted for cool-dois.html.md\nDOI 10.5438/0000-00ss minted for index.html.erb")
+    end
+
+    it 'should hide for all urls' do
+      filepath = fixture_path + 'index-minted.html'
+      response = subject.hide_dois_for_all_urls(filepath, options)
+      expect(response).to eq("No DOI for cool-dois.html.md\nErrors for DOI 10.5438/0000-nw90: Not found\n")
+    end
+
+    it 'should mint and hide for all urls' do
+      filepath = fixture_path + 'index.html'
+      response = subject.mint_and_hide_dois_for_all_urls(filepath, options)
+      expect(response).to eq("DOI 10.5438/0000-03vc minted and hidden for cool-dois.html.md\nDOI 10.5438/0000-00ss minted and hidden for index.html.erb")
+    end
+
+    it 'should get_json_ld_from_work' do
+      filepath = fixture_path + 'cool-dois/index.html'
+      json = subject.get_json_ld_from_work(filepath)
+      metadata = JSON.parse(json)
+      expect(metadata["url"]).to eq("https://blog.datacite.org/cool-dois/")
+      expect(metadata["author"]).to eq([{"@type"=>"Person", "@id"=>"http://orcid.org/0000-0003-1419-2405", "givenName"=>"Martin", "familyName"=>"Fenner", "name"=>"Martin Fenner"}])
+      expect(metadata["description"]).to eq("In 1998 Tim Berners-Lee coined the term cool URIs (1998), that is URIs that don’t change. We know that URLs referenced in the scholarly literature are often not cool, leading to link rot (Klein et al., 2014) and making it hard or impossible to find...")
+      expect(metadata["citation"]).to eq([{"@type"=>"CreativeWork",
+                                           "@id"=>"https://www.w3.org/Provider/Style/URI"},
+                                          {"@type"=>"CreativeWork",
+                                           "@id"=>"https://doi.org/10.1371/journal.pone.0115253"}])
+      expect(metadata["isPartOf"]).to eq("@type"=>"Blog", "@id"=>"https://blog.datacite.org", "name"=>"DataCite Blog")
+    end
+
+    it 'should post_metadata_for_work' do
+      filepath = fixture_path + 'cool-dois/index.html'
+      json = subject.get_json_ld_from_work(filepath)
+      response = subject.post_metadata_for_work(json, options)
+      expect(response.body["data"]).to eq("OK")
+      expect(response.status).to eq(201)
+    end
+
+    it 'should hide_metadata_for_work' do
+      filepath = fixture_path + 'cool-dois/index.html'
+      json = subject.get_json_ld_from_work(filepath)
+      response = subject.hide_metadata_for_work(json, options)
+      expect(response.body["data"]).to eq("OK")
+      expect(response.status).to eq(200)
+    end
+  end
+
+  context "jats" do
+    it 'should generate metadata for jats' do
+      filepath = fixture_path + 'cool-dois/index.html'
+      metadata = subject.generate_metadata_for_jats(filepath)
+      expect(metadata["author"]).to eq([{"given_name"=>"Martin", "family_name"=>"Fenner", "orcid"=>"0000-0003-1419-2405"}])
+      expect(metadata["license_url"]).to eq("https://creativecommons.org/licenses/by/4.0/")
+    end
+
+    it 'should generate jats xml' do
+      filepath = fixture_path + 'cool-dois/index.html'
+      expect(subject.generate_jats_for_url(filepath, options.merge(bibliography: bibliography))).to eq("JATS XML written for cool-dois.html.md")
+    end
+
+    it 'should generate jats for all urls' do
+      filepath = fixture_path + 'index.html'
+      response = subject.generate_jats_for_all_urls(filepath, options)
+      expect(response).to eq("JATS XML written for cool-dois.html.md\nJATS XML written for index.html.erb")
+    end
+
+    # it 'should validate jats xml' do
+    #   filepath = fixture_path + 'cool-dois/index.html'
+    #   expect(subject.generate_jats_for_url(filepath, options.merge(bibliography: bibliography))).to eq("JATS XML written for cool-dois.html.md")
+    #   xml = IO.read(fixture_path + 'cool-dois/cool-dois.xml')
+    #   expect(subject.validate_jats(xml).body["errors"]).to be_empty
+    # end
+  end
+
+  context "get_related_identifiers" do
+    it 'isPartOf' do
+      metadata = { "isPartOf" => {
+        "@type" => "Blog",
+        "@id" => "https://blog.datacite.org",
+        "name" => "DataCite Blog" } }
+      expect(subject.get_related_identifiers(metadata)).to eq([{:value=>"https://blog.datacite.org/", :related_identifier_type=>"URL", :relation_type=>"IsPartOf"}])
+    end
+  end
+
+  context "filepath from url" do
+    it 'https://blog.datacite.org/' do
+      url = 'https://blog.datacite.org/'
+      filename, build_path, source_path = subject.filepath_from_url(url, build_dir: source_dir, source_dir: source_dir)
+      expect(filename).to eq("index.html.erb")
+      expect(source_path).to eq(fixture_path + "index.html.erb")
+      expect(build_path).to eq(fixture_path + "index.html")
+    end
+
+    it 'https://blog.datacite.org' do
+      url = 'https://blog.datacite.org'
+      filename, build_path, source_path = subject.filepath_from_url(url, build_dir: source_dir, source_dir: source_dir)
+      expect(filename).to eq("index.html.erb")
+      expect(source_path).to eq(fixture_path + "index.html.erb")
+      expect(build_path).to eq(fixture_path + "index.html")
+    end
+
+    it 'index.html' do
+      url = fixture_path + 'index.html'
+      filename, build_path, source_path = subject.filepath_from_url(url, build_dir: source_dir, source_dir: source_dir)
+      expect(filename).to eq("index.html.erb")
+      expect(source_path).to eq(fixture_path + "index.html.erb")
+      expect(build_path).to eq(fixture_path + "index.html")
+    end
+
+    it 'index.html basename' do
+      url = 'index.html'
+      filename, build_path, source_path = subject.filepath_from_url(url, build_dir: source_dir, source_dir: source_dir)
+      expect(filename).to eq("index.html.erb")
+      expect(source_path).to eq(fixture_path + "index.html.erb")
+      expect(build_path).to eq(fixture_path + "index.html")
+    end
+
+    it 'cool-dois.html' do
+      url = fixture_path + 'cool-dois.html'
+      filename, build_path, source_path = subject.filepath_from_url(url, build_dir: source_dir, source_dir: source_dir)
+      expect(filename).to eq("cool-dois.html.md")
+      expect(source_path).to eq(fixture_path + "cool-dois.html.md")
+      expect(build_path).to eq(fixture_path + "cool-dois/index.html")
+    end
+
+    it 'cool-dois.html basename' do
+      url = 'cool-dois.html'
+      filename, build_path, source_path = subject.filepath_from_url(url, build_dir: source_dir, source_dir: source_dir)
+      expect(filename).to eq("cool-dois.html.md")
+      expect(source_path).to eq(fixture_path + "cool-dois.html.md")
+      expect(build_path).to eq(fixture_path + "cool-dois/index.html")
+    end
+  end
+end

data/spec/fixtures/cool-dois-minted/index.html

@@ -44,7 +44,7 @@
       </script>
 
     <script type="application/ld+json">
-      {"@context":"http://schema.org","@type":"BlogPosting","@id":"https://doi.org/10.5438/55E5-T5C0","name":"Cool DOI's","alternateName":"MS-55-5135-1980","url":"https://blog.datacite.org/cool-dois/","author":[{"@type":"Person","@id":"http://orcid.org/0000-0003-1419-2405","givenName":"Martin","familyName":"Fenner","name":"Martin Fenner"}],"publisher":{"@type":"Organization","name":"DataCite"},"dateCreated":"2016-12-15","datePublished":"2016-12-15","dateModified":"2016-12-15","keywords":"doi, featured","version":"1.0","description":"In 1998 Tim Berners-Lee coined the term cool URIs (1998), that is URIs that don’t change. We know that URLs referenced in the scholarly literature are often not cool, leading to link rot (Klein et al., 2014) and making it hard or impossible to find...","license":"https://creativecommons.org/licenses/by/4.0/","image":"https://blog.datacite.org/images/2016/12/cool-dois.png","encoding":{"@type":"MediaObject","@id":"https://blog.datacite.org/cool-dois/cool-dois.xml","fileFormat":"application/xml"},"isPartOf":{"@type":"Blog","@id":"https://doi.org/10.5438/0000-00SS","name":"DataCite Blog"},"citation":[{"@type":"CreativeWork","@id":"https://www.w3.org/Provider/Style/URI"},{"@type":"CreativeWork","@id":"https://doi.org/10.1371/journal.pone.0115253"}]}
+      {"@context":"http://schema.org","@type":"BlogPosting","@id":"https://doi.org/10.5438/55e5-t5c0","name":"Cool DOI's","alternateName":"MS-55-5135-1980","url":"https://blog.datacite.org/cool-dois/","author":[{"@type":"Person","@id":"http://orcid.org/0000-0003-1419-2405","givenName":"Martin","familyName":"Fenner","name":"Martin Fenner"}],"publisher":{"@type":"Organization","name":"DataCite"},"dateCreated":"2016-12-15","datePublished":"2016-12-15","dateModified":"2016-12-15","keywords":"doi, featured","version":"1.0","description":"In 1998 Tim Berners-Lee coined the term cool URIs (1998), that is URIs that don’t change. We know that URLs referenced in the scholarly literature are often not cool, leading to link rot (Klein et al., 2014) and making it hard or impossible to find...","license":"https://creativecommons.org/licenses/by/4.0/","image":"https://blog.datacite.org/images/2016/12/cool-dois.png","encoding":{"@type":"MediaObject","@id":"https://blog.datacite.org/cool-dois/cool-dois.xml","fileFormat":"application/xml"},"isPartOf":{"@type":"Blog","@id":"https://doi.org/10.5438/0000-00SS","name":"DataCite Blog"},"citation":[{"@type":"CreativeWork","@id":"https://www.w3.org/Provider/Style/URI"},{"@type":"CreativeWork","@id":"https://doi.org/10.1371/journal.pone.0115253"}]}
     </script>
   </head>
   <body>

data/spec/fixtures/cool-dois-minted.html.md

@@ -8,7 +8,7 @@ tags:
 accession_number: MS-124
 image: https://blog.datacite.org/images/2016/12/cool-dois.png
 published: false
-doi: 10.5438/0000-03WD
+doi: 10.5438/0000-03wd
 ---
 In 1998 Tim Berners-Lee coined the term cool URIs [-@https://www.w3.org/Provider/Style/URI], that is URIs that don’t change. We know that URLs referenced in the scholarly literature are often not cool, leading to link rot [@https://doi.org/10.1371/journal.pone.0115253] and making it hard or impossible to find the referenced resource.READMORE
 

data/spec/fixtures/cool-dois.html.md

@@ -8,7 +8,7 @@ tags:
 accession_number: MS-123
 image: https://blog.datacite.org/images/2016/12/cool-dois.png
 published: false
-doi: 10.5438/0000-03VC
+doi: 10.5438/0000-03vc
 date: '2016-12-15'
 ---
 In 1998 Tim Berners-Lee coined the term cool URIs [-@https://www.w3.org/Provider/Style/URI], that is URIs that don’t change. We know that URLs referenced in the scholarly literature are often not cool, leading to link rot [@https://doi.org/10.1371/journal.pone.0115253] and making it hard or impossible to find the referenced resource.READMORE

data/spec/fixtures/index-minted.html

@@ -45,7 +45,7 @@
       </script>
 
     <script type="application/ld+json">
-      {"@context":"http://schema.org","@type":"Blog","@id":"https://doi.org/10.5438/0000-NW90","name":"DataCite Blog","alternateName":"MS-25-1785","url":"https://blog.datacite.org/","author":[{"@type":"Person","@id":"http://orcid.org/0000-0002-9300-5278","givenName":"Patricia","familyName":"Cruse","name":"Trisha Cruse"},{"@type":"Person","@id":"http://orcid.org/0000-0001-5952-7630","givenName":"Laura","familyName":"Rueda","name":"Laura Rueda"},{"@type":"Person","@id":"http://orcid.org/0000-0003-3484-6875","givenName":"Kristian","familyName":"Garza","name":"Kristian Garza"},{"@type":"Person","@id":"http://orcid.org/0000-0003-1419-2405","givenName":"Martin","familyName":"Fenner","name":"Martin Fenner"}],"publisher":{"@type":"Organization","name":"DataCite"},"dateCreated":"2015-08-03","datePublished":"2015-08-03","dateModified":"2015-08-03","version":"1.0","description":"The DataCite blog.","license":"https://creativecommons.org/licenses/by/4.0/","hasPart":[{"@type":"BlogPosting","@id":"https://doi.org/10.5438/55E5-T5C0","name":"Cool DOI's","datePublished":"2016-12-15"}]}
+      {"@context":"http://schema.org","@type":"Blog","@id":"https://doi.org/10.5438/0000-nw90","name":"DataCite Blog","alternateName":"MS-25-1785","url":"https://blog.datacite.org/","author":[{"@type":"Person","@id":"http://orcid.org/0000-0002-9300-5278","givenName":"Patricia","familyName":"Cruse","name":"Trisha Cruse"},{"@type":"Person","@id":"http://orcid.org/0000-0001-5952-7630","givenName":"Laura","familyName":"Rueda","name":"Laura Rueda"},{"@type":"Person","@id":"http://orcid.org/0000-0003-3484-6875","givenName":"Kristian","familyName":"Garza","name":"Kristian Garza"},{"@type":"Person","@id":"http://orcid.org/0000-0003-1419-2405","givenName":"Martin","familyName":"Fenner","name":"Martin Fenner"}],"publisher":{"@type":"Organization","name":"DataCite"},"dateCreated":"2015-08-03","datePublished":"2015-08-03","dateModified":"2015-08-03","version":"1.0","description":"The DataCite blog.","license":"https://creativecommons.org/licenses/by/4.0/","hasPart":[{"@type":"BlogPosting","@id":"https://doi.org/10.5438/55E5-T5C0","name":"Cool DOI's","datePublished":"2016-12-15"}]}
     </script>
   </head>
   <body>

data/spec/fixtures/index.html

@@ -48,7 +48,7 @@
       {
         "@context": "http://schema.org",
         "@type": "Blog",
-        "@id": "https://doi.org/10.5438/0000-00SS",
+        "@id": "https://doi.org/10.5438/0000-00ss",
         "name": "DataCite Blog",
         "alternateName": "MS-25-1785",
         "url": "https://blog.datacite.org/",
@@ -89,7 +89,7 @@
         "license": "https://creativecommons.org/licenses/by/4.0/",
         "hasPart": [{
             "@type": "BlogPosting",
-            "@id": "https://doi.org/10.5438/55E5-T5C0",
+            "@id": "https://doi.org/10.5438/55e5-t5c0",
             "name": "Cool DOI's",
             "dateCreated": "2016-12-15",
             "datePublished": "2016-12-15"

data/spec/fixtures/index.html.erb

@@ -9,7 +9,7 @@ author:
 accession_number: MS-25-1785
 published: false
 date: '2017-07-11'
-doi: 10.5438/0000-00SS
+doi: 10.5438/0000-00ss
 ---
 <div class="container-fluid">
   <div class="row">

data/spec/fixtures/vcr_cassettes/Cirneco_DataCenter/jats/should_generate_jats_for_all_urls.yml

@@ -2,7 +2,7 @@
 http_interactions:
 - request:
     method: head
-    uri: https://doi.org/10.5438/55E5-T5C0
+    uri: https://doi.org/10.5438/55e5-t5c0
     body:
       encoding: US-ASCII
       string: ''
@@ -23,16 +23,16 @@ http_interactions:
       Location:
       - https://blog.datacite.org/cool-dois/
       Expires:
-      - Sat, 22 Jul 2017 07:25:52 GMT
+      - Sun, 23 Jul 2017 07:26:33 GMT
       Content-Type:
       - text/html;charset=utf-8
       Content-Length:
       - '157'
       Date:
-      - Sat, 22 Jul 2017 07:03:54 GMT
+      - Sun, 23 Jul 2017 06:56:16 GMT
     body:
       encoding: UTF-8
       string: ''
     http_version: 
-  recorded_at: Sat, 22 Jul 2017 07:03:53 GMT
+  recorded_at: Sun, 23 Jul 2017 06:56:15 GMT
 recorded_with: VCR 3.0.3

data/spec/fixtures/vcr_cassettes/Cirneco_DataCenter/mint_and_hide_DOIs/should_hide_for_all_urls.yml

@@ -23,57 +23,16 @@ http_interactions:
       Location:
       - https://blog.datacite.org/cool-dois/
       Expires:
-      - Sat, 22 Jul 2017 07:25:52 GMT
+      - Sun, 23 Jul 2017 07:26:33 GMT
       Content-Type:
       - text/html;charset=utf-8
       Content-Length:
       - '157'
       Date:
-      - Sat, 22 Jul 2017 07:04:36 GMT
+      - Sun, 23 Jul 2017 06:56:11 GMT
     body:
       encoding: UTF-8
       string: ''
     http_version: 
-  recorded_at: Sat, 22 Jul 2017 07:04:35 GMT
-- request:
-    method: delete
-    uri: https://mds.test.datacite.org/metadata/10.5438/0000-NW90
-    body:
-      encoding: US-ASCII
-      string: ''
-    headers:
-      User-Agent:
-      - Maremma - https://github.com/datacite/maremma
-      Accept:
-      - text/html,application/json,application/xml;q=0.9, text/plain;q=0.8,image/png,*/*;q=0.5
-      Authorization:
-      - Basic <MDS_TOKEN>
-  response:
-    status:
-      code: 404
-      message: Not Found
-    headers:
-      Access-Control-Allow-Methods:
-      - GET, POST, OPTIONS
-      Access-Control-Allow-Origin:
-      - "*"
-      Cache-Control:
-      - no-cache, no-store
-      Date:
-      - Sat, 22 Jul 2017 07:04:36 GMT
-      Expires:
-      - Thu, 01 Jan 1970 00:00:00 GMT
-      Pragma:
-      - no-cache
-      Server:
-      - openresty/1.11.2.2
-      Content-Length:
-      - '17'
-      Connection:
-      - keep-alive
-    body:
-      encoding: UTF-8
-      string: DOI doesn't exist
-    http_version: 
-  recorded_at: Sat, 22 Jul 2017 07:04:35 GMT
+  recorded_at: Sun, 23 Jul 2017 06:56:10 GMT
 recorded_with: VCR 3.0.3