paper_metadata 0.0.4 → 0.0.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: f404b9b484329f732c2b71bd2e4abdf84ca97cee
-  data.tar.gz: 8f4b35221f91a12d49a445028462d9561ea8424a
+  metadata.gz: 70781715c1398e4f11015d9e03ccec207db08c3a
+  data.tar.gz: c8338bcad90badd674fa2d227710dabb669554a9
 SHA512:
-  metadata.gz: 9baba7b45129a21cf0ad08a6db0469d0f395bd848a9e619bce67d77cfb9d9ab2f08423d5444ad4a87ef7c44d5b9d4b3cccc4a3b124bbac98d157f50a66fb4d26
-  data.tar.gz: bd61fc685c0419d584040a7fe01b13d5630769378b72ce48f8c369982851cd02def110229c245c5601b38ae3ea8b2da37ad71ef5da5a12b07fbebd9a8f9ab113
+  metadata.gz: 6bf5144e34d59eca67c71d8a8a59959afc9ff322cee8d9371f37d5ab893837c52f9f90556575f431633bef3670335f42d85ad7814a6e5e6f7cb88fadd3233be9
+  data.tar.gz: 9f3a9c84e07ed5c0f5666b8649abd87e4cf626ff2825bd03649cec9a1d94835f2f513174a0046ec761939ce7774910a43a7160b4949c92caaf1e846e08c0109a

data/lib/paper_metadata.rb

@@ -4,6 +4,7 @@ require 'net/http'
 require 'open-uri'
 require 'cgi'
 require 'uri'
+require 'json'
 
 module PaperMetadata
   class << self
@@ -17,6 +18,8 @@ module PaperMetadata
     def metadata_for(identifier)
       if identifier =~ /^arxiv\:(.*)/i
         metadata_for_arxiv($1.strip)
+      elsif identifier =~ /^doi\:\s*(10\.6084\/.*)/i
+        metadata_for_doi_from_figshare($1.strip)
       elsif identifier =~ /^doi\:(.*)/i
         metadata_for_doi($1.strip)
       end
@@ -45,11 +48,44 @@ module PaperMetadata
           doc.xpath("//journal_issue/publication_date/year").first.inner_html
         paper[:resource] = doc.xpath("//journal_article/doi_data/resource").inner_html
       else
-        paper[:status] = :NODOI
+        paper = metadata_for_doi_from_datacite(doi)
       end
       paper
     end
 
+    def metadata_for_doi_from_datacite(doi)
+      paper = {}
+      response = JSON.parse(open("http://search.datacite.org/api?q=#{CGI.escape(doi)}&fl=doi,creator,title,publisher,publicationYear,datacentre&wt=json").read)
+      if response && !response['response']['docs'].empty? && response['response']['docs'].first['title']
+        result = response['response']['docs'].first
+        paper[:title] = result['title'].first
+        paper[:authors] = result['creator'].map{|c| c.split(', ').reverse.join(' ')}.join(', ')
+        paper[:published] = result['publicationYear']
+        paper[:publisher] = result['publisher']
+        paper[:datacentre] = result['datacentre']
+        paper[:journal] = 'DataCite'
+        paper
+      else
+        {status: :NODOI}
+      end
+    end
+
+    def metadata_for_doi_from_figshare(doi)
+      paper = {}
+      response = JSON.parse(open("http://api.figshare.com/articles/#{doi}").read)
+
+      if response && response['items'] && response['items'].first
+        result = response['items'].first
+        paper[:title] = result['title']
+        paper[:authors] = result['authors'].map{|a| a['full_name']}.join(', ')
+        paper[:published] = result['published_date']
+        paper[:journal] = 'FigShare'
+        paper
+      else
+        {status: :NODOI}
+      end
+    end
+
     def metadata_for_arxiv(identifier)
       identifier.gsub!(/^arXiv\:/i, '')
       url = URI.parse("http://export.arxiv.org/api/query?search_query=#{CGI.escape(identifier)}&start=0&max_results=1")
@@ -59,11 +95,12 @@ module PaperMetadata
       paper = Hash.new
       if entry = doc.xpath("//entry").first
         paper[:title] = entry.xpath('title').text
-        paper[:author] = entry.xpath('author').text.strip
+        paper[:authors] = entry.xpath('author').text.split("\n").map{|a| a.strip if a.strip != ""}.compact.join(', ')
         paper[:id] = entry.xpath('id').text
         paper[:updated] = entry.xpath('updated').text
         paper[:summary] = entry.xpath('summary').text
         paper[:published] = entry.xpath('published').text
+        paper[:journal] = 'arXiv'
       end
       paper
     end

data/lib/paper_metadata/version.rb

@@ -1,3 +1,3 @@
 module PaperMetadata
-  VERSION = "0.0.4"
+  VERSION = "0.0.5"
 end

data/test/test_paper_metadata.rb

@@ -9,33 +9,62 @@ class PaperMetadataTest < Test::Unit::TestCase
     stub_request(:any, /www.crossref.org\/.*/).
       to_return(:body => doi_response, :status => 200,  :headers => { 'Content-Length' => doi_response.length } )
 
-    PaperMetadata.doi_username = 'test@example.com'
-
     assert_equal "Basic Modeling Approach To Optimize Elemental Imaging by Laser Ablation ICPMS",
       PaperMetadata.metadata_for('doi:10.1021/ac1014832')[:title]
   end
 
   def test_doi_parsing_live
     WebMock.allow_net_connect!
-    doi_response = File.read(File.join(File.dirname(__FILE__), 'doi.xml'))
-    PaperMetadata.doi_username = 'test@crossref.org'
     assert_equal "Basic Modeling Approach To Optimize Elemental Imaging by Laser Ablation ICPMS",
       PaperMetadata.metadata_for('doi:10.1021/ac1014832')[:title]
     WebMock.disable_net_connect!
   end
 
+  def test_datacite_doi_parsing_live
+    WebMock.allow_net_connect!
+
+    metadata = PaperMetadata.metadata_for('doi:10.6092/issn.1973-9494/3396')
+    assert_equal "AlmaDL Journals: Quality Services for Open Access Scientific Publications at the University of Bologna",
+      metadata[:title]
+
+    assert_equal "Marialaura Vignocchi, Roberta Lauriola, Andrea Zanni, Antonio Puglisi, Raffaele Messuti",
+      metadata[:authors]
+
+    WebMock.disable_net_connect!
+  end
+
+  def test_figshare_doi_parsing_live
+    WebMock.allow_net_connect!
+
+    metadata = PaperMetadata.metadata_for('doi:10.6084/m9.figshare.736442')
+    assert_equal "XML4NGS : A XML-based description of a Next-Generation sequencing project allowing the generation of a ’Makefile’-driven workflow.",
+      metadata[:title]
+
+    assert_equal "Pierre Lindenbaum, Raluca Teusan, Richard Redon, Audrey Bihouée, Solena LeScouarnec",
+      metadata[:authors]
+
+    WebMock.disable_net_connect!
+
+  end
+
   def test_arxiv_parsing
     arxiv_response = File.read(File.join(File.dirname(__FILE__), 'arxiv.xml'))
     stub_request(:any, /.*arxiv.org\/.*/).
       to_return(:body => arxiv_response, :status => 200,  :headers => { 'Content-Length' => arxiv_response.length } )
     assert_equal "Thomas Vojta",
-      PaperMetadata.metadata_for('arXiv:1301.7746')[:author]
+      PaperMetadata.metadata_for('arXiv:1301.7746')[:authors]
   end
 
   def test_arxiv_parsing_live
     WebMock.allow_net_connect!
     assert_equal "Thomas Vojta",
-      PaperMetadata.metadata_for('arXiv:1301.7746')[:author]
+      PaperMetadata.metadata_for('arXiv:1301.7746')[:authors]
+    WebMock.disable_net_connect!
+  end
+
+  def test_nonexistent_doi
+    WebMock.allow_net_connect!
+    assert_equal({status: :NODOI}, PaperMetadata.metadata_for('doi:thisdoi/doesntexist'))
     WebMock.disable_net_connect!
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: paper_metadata
 version: !ruby/object:Gem::Version
-  version: 0.0.4
+  version: 0.0.5
 platform: ruby
 authors:
 - Jure Triglav
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-07-07 00:00:00.000000000 Z
+date: 2013-08-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: pry