Module: PaperMetadata

Defined in:
lib/paper_metadata.rb,
lib/paper_metadata/version.rb

Constant Summary collapse

VERSION = 
"0.0.3"

Class Attribute Summary collapse

Class Method Summary collapse

Class Attribute Details

.doi_username ⇒ Object

Returns the value of attribute doi_username.



9
10
11
 
# File 'lib/paper_metadata.rb', line 9

def doi_username
  @doi_username
end

Class Method Details

.metadata_for(identifier) ⇒ Object 



16
17
18
19
20
21
22
 
# File 'lib/paper_metadata.rb', line 16

def metadata_for(identifier)
  if identifier =~ /^arxiv\:/i
    metadata_for_arxiv(identifier)
  elsif identifier =~ /^doi\:/i
    metadata_for_doi(identifier)
  end
end

.metadata_for_arxiv(identifier) ⇒ Object 



50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
 
# File 'lib/paper_metadata.rb', line 50

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")
  res = Net::HTTP.get_response(url)
  doc = Nokogiri::XML(res.body)
  doc.remove_namespaces!
  paper = Hash.new
  if entry = doc.xpath("//entry").first
    paper[:title] = entry.xpath('title').text
    paper[:author] = entry.xpath('author').text.strip
    paper[:id] = entry.xpath('id').text
    paper[:updated] = entry.xpath('updated').text
    paper[:summary] = entry.xpath('summary').text
    paper[:published] = entry.xpath('published').text
  end
  paper
end

.metadata_for_doi(doi) ⇒ Object 



24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
 
# File 'lib/paper_metadata.rb', line 24

def metadata_for_doi(doi)
  doc = Nokogiri::XML(open("http://www.crossref.org/openurl/?id=#{CGI.escape(doi)}&noredirect=true&pid=#{PaperMetadata.doi_username}&format=unixref"))
  paper = Hash.new

  if doc.xpath("//titles/title").first
    paper[:volume] = doc.xpath("//journal_issue/journal_volume/volume").inner_html.to_s
    paper[:isssue] = doc.xpath("//journal_issue/issue").inner_html.to_s
    paper[:first_page] = doc.xpath("//pages/first_page").inner_html.to_s
    paper[:last_page] = doc.xpath("//pages/last_page").inner_html.to_s
    paper[:title] = doc.xpath("//titles/title").inner_html.to_s
    paper[:authors] = doc.xpath("//contributors/person_name").
      map{ |author| author.content.strip}

    paper[:authors] = paper[:authors].map do |author|
      author.gsub(/\s+/, ' ')
    end.join(', ')

    paper[:journal] = doc.xpath("//abbrev_title").inner_html + " " +
      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
  end
  paper
end