Class: Relaton::Nist::DataFetcher

Inherits:
Core::DataFetcher
  • Object
show all
Defined in:
lib/relaton/nist/data_fetcher.rb

Constant Summary collapse

URL = 
"https://github.com/usnistgov/NIST-Tech-Pubs/releases/download/Nov2024/allrecords-MODS.xml"

Instance Method Summary collapse

Instance Method Details

#fetch(_source = nil) ⇒ Object 



14
15
16
17
18
19
20
 
# File 'lib/relaton/nist/data_fetcher.rb', line 14

def fetch(_source = nil)
  FileUtils.rm Dir[File.join(@output, "*.#{@ext}")]
  fetch_tech_pubs
  # add_static_files
  index.save
  report_errors
end

#fetch_tech_pubs ⇒ Object 



22
23
24
25
26
 
# File 'lib/relaton/nist/data_fetcher.rb', line 22

def fetch_tech_pubs
  xml_data = Mechanize.new.get(URL).body
  docs = LocMods::Collection.from_xml xml_data
  docs.mods.each { |doc| write_file ModsParser.new(doc, series, @errors).parse }
end

#index ⇒ Object 



62
63
64
 
# File 'lib/relaton/nist/data_fetcher.rb', line 62

def index
  @index ||= Relaton::Index.find_or_create :nist, file: "#{INDEXFILE}.yaml"
end

#log_error(msg) ⇒ Object 



58
59
60
 
# File 'lib/relaton/nist/data_fetcher.rb', line 58

def log_error(msg)
  Util.error msg
end

#series ⇒ Object 



66
67
68
 
# File 'lib/relaton/nist/data_fetcher.rb', line 66

def series
  @series ||= YAML.load_file File.expand_path("series.yaml", __dir__)
end

#to_bibxml(bib) ⇒ Object 



54
55
56
 
# File 'lib/relaton/nist/data_fetcher.rb', line 54

def to_bibxml(bib)
  bib.to_rfcxml
end

#to_xml(bib) ⇒ Object 



50
51
52
 
# File 'lib/relaton/nist/data_fetcher.rb', line 50

def to_xml(bib)
  Bibdata.to_xml(bib)
end

#to_yaml(bib) ⇒ Object

def add_static_files

Dir["./static/*.yaml"].each do |file|
  bib = Item.from_yaml(File.read(file, encoding: "UTF-8"))
  index.add_or_update bib.docidentifier[0].content, file
end

end



46
47
48
 
# File 'lib/relaton/nist/data_fetcher.rb', line 46

def to_yaml(bib)
  Item.to_yaml(bib)
end

#write_file(bib) ⇒ Object 



28
29
30
31
32
33
34
35
36
37
 
# File 'lib/relaton/nist/data_fetcher.rb', line 28

def write_file(bib)
  id = bib.docidentifier.find(&:primary) || bib.docidentifier.first
  file = output_file id.content.sub(/^NIST IR/, "NISTIR")
  if @files.include? file
    Util.warn "File #{file} exists. Docid: #{bib.docidentifier[0].content}"
  else @files << file
  end
  index.add_or_update bib.docidentifier[0].content, file
  File.write file, serialize(bib), encoding: "UTF-8"
end