Class: Pubid::Nist::NistTechPubs

Inherits:
Object
  • Object
show all
Defined in:
lib/pubid/nist/nist_tech_pubs.rb

Constant Summary collapse

URL = 
"https://raw.githubusercontent.com/usnistgov/NIST-Tech-Pubs/nist-pages/xml/allrecords.xml".freeze

Class Attribute Summary collapse

Class Method Summary collapse

Class Attribute Details

.converted_doi ⇒ Object

Returns the value of attribute converted_doi.



15
16
17
 
# File 'lib/pubid/nist/nist_tech_pubs.rb', line 15

def converted_doi
  @converted_doi
end

.converted_id ⇒ Object

Returns the value of attribute converted_id.



15
16
17
 
# File 'lib/pubid/nist/nist_tech_pubs.rb', line 15

def converted_id
  @converted_id
end

.documents ⇒ Object

Returns the value of attribute documents.



15
16
17
 
# File 'lib/pubid/nist/nist_tech_pubs.rb', line 15

def documents
  @documents
end

Class Method Details

.comply_with_pubid ⇒ Object 



68
69
70
71
72
73
74
 
# File 'lib/pubid/nist/nist_tech_pubs.rb', line 68

def comply_with_pubid
  fetch.select do |doc|
    convert(doc).to_s == doc[:id]
  rescue Pubid::Core::Errors::ParseError
    false
  end
end

.convert(doc) ⇒ Object 



29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
 
# File 'lib/pubid/nist/nist_tech_pubs.rb', line 29

def convert(doc)
  id = @converted_id[doc[:id]] ||= Pubid::Nist::Identifier.parse(doc[:id])
  return id unless doc.key?(:doi)

  begin
    doi = @converted_doi[doc[:doi]] ||=
      Pubid::Nist::Identifier.parse(doc[:doi])
  rescue Pubid::Core::Errors::ParseError
    return id
  end
  # return more complete pubid
  id.merge(doi)
rescue Pubid::Core::Errors::ParseError
  @converted_doi[doc[:doi]] ||= Pubid::Nist::Identifier.parse(doc[:doi])
end

.different_with_pubid ⇒ Object 



76
77
78
79
80
81
82
 
# File 'lib/pubid/nist/nist_tech_pubs.rb', line 76

def different_with_pubid
  fetch.reject do |doc|
    convert(doc).to_s == doc[:id]
  rescue Pubid::Core::Errors::ParseError
    true
  end
end

.fetch ⇒ Object 



17
18
19
20
21
22
23
24
25
26
27
 
# File 'lib/pubid/nist/nist_tech_pubs.rb', line 17

def fetch
  Lightly.prune
  @documents ||= Lightly.get "documents" do
    Nokogiri::XML(URI.open(URL))
      .xpath("/body/query/doi_record/report-paper/report-paper_metadata")
      .map { |doc| parse_docid doc }
  end
rescue StandardError => e
  warn e.message
  []
end

.parse_docid(doc) ⇒ Object 



45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
 
# File 'lib/pubid/nist/nist_tech_pubs.rb', line 45

def parse_docid(doc)
  id = doc.at("publisher_item/item_number", "publisher_item/identifier")
         &.text&.sub(%r{^/}, "")
  if id == "NBS BH 10"
    # XXX: "doi" attribute is missing for doi_data
    doi = "NBS.BH.10"
  else
    doi = doc.at("doi_data/doi").text.gsub("10.6028/", "")
  end

  title = doc.at("titles/title").text
  title += " #{doc.at('titles/subtitle').text}" if doc.at("titles/subtitle")
  case doi
  when "10.6028/NBS.CIRC.12e2revjune" then id.sub!("13e", "12e")
  when "10.6028/NBS.CIRC.36e2" then id.sub!("46e", "36e")
  when "10.6028/NBS.HB.67suppJune1967" then id.sub!("1965", "1967")
  when "10.6028/NBS.HB.105-1r1990" then id.sub!("105-1-1990", "105-1r1990")
  when "10.6028/NIST.HB.150-10-1995" then id.sub!(/150-10$/, "150-10-1995")
  end

  { id: id || doi, doi: doi, title: title }
end

.parse_fail_with_pubid ⇒ Object 



84
85
86
87
88
89
90
 
# File 'lib/pubid/nist/nist_tech_pubs.rb', line 84

def parse_fail_with_pubid
  fetch.select do |doc|
    convert(doc).to_s && false
  rescue Pubid::Core::Errors::ParseError
    true
  end
end

.status ⇒ Object

returning current document id, doi, title and final PubID



93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
 
# File 'lib/pubid/nist/nist_tech_pubs.rb', line 93

def status
  fetch.lazy.map do |doc|
    final_doc = convert(doc)
    {
      id: doc[:id],
      doi: doc[:doi],
      title: doc[:title],
      finalPubId: final_doc.to_s,
      mr: final_doc.to_s(:mr),
    }
  rescue Pubid::Core::Errors::ParseError
    {
      id: doc[:id],
      doi: doc[:doi],
      title: doc[:title],
      finalPubId: "parse error",
      mr: "parse_error",
    }
  end
end