Class: Relaton3gpp::DataFetcher

Inherits:
Object
  • Object
show all
Defined in:
lib/relaton_3gpp/data_fetcher.rb

Constant Summary collapse

CURRENT = 
"current.yaml".freeze

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(output, format) ⇒ DataFetcher

Data fetcher initializer

Parameters:

  • output (String) —

    directory to save files

  • format (String) —

    format of output files (xml, yaml, bibxml)



10
11
12
13
14
15
16
17
18
19
 
# File 'lib/relaton_3gpp/data_fetcher.rb', line 10

def initialize(output, format)
  require "fileutils"
  require "net/ftp"
  require "csv"

  @output = output
  @format = format
  @ext = format.sub(/^bib/, "")
  @files = []
end

Class Method Details

.fetch(source, output: "data", format: "yaml") ⇒ Object

Initialize fetcher and run fetch

Parameters:

  • source (Strin) —

    source name

  • output (Strin) (defaults to: "data") —

    directory to save files, default: “data”

  • format (Strin) (defaults to: "yaml") —

    format of output files (xml, yaml, bibxml), default: yaml



32
33
34
35
36
37
38
39
40
 
# File 'lib/relaton_3gpp/data_fetcher.rb', line 32

def self.fetch(source, output: "data", format: "yaml")
  t1 = Time.now
  puts "Started at: #{t1}"
  FileUtils.mkdir_p output
  new(output, format).fetch(source == "status-smg-3GPP-force")
  t2 = Time.now
  puts "Stopped at: #{t2}"
  puts "Done in: #{(t2 - t1).round} sec."
end

Instance Method Details

#fetch(renewal) ⇒ Object

Parse documents

Parameters:

  • renewal (Boolean) —

    force to update all documents



47
48
49
50
51
52
53
54
55
56
57
58
59
60
 
# File 'lib/relaton_3gpp/data_fetcher.rb', line 47

def fetch(renewal) # rubocop:disable Metrics/MethodLength, Metrics/AbcSize
  file = get_file renewal
  return unless file && File.exist?(file) && File.size(file) > 20_000_000

  if renewal
    FileUtils.rm_f File.join(@output, "/*") # if renewal && dbs["2001-04-25_schedule"].any?
    index.remove_all # if renewal
  end
  CSV.open(file, "r:bom|utf-8", headers: true, col_sep: ";").each do |row|
    save_doc Parser.parse(row)
  end
  File.write CURRENT, @current.to_yaml, encoding: "UTF-8"
  index.save
end

#file_name(bib) ⇒ String

Generate file name

Parameters:

  • bib (RelatonW3c::W3cBibliographicItem) —

    bibliographic item

Returns:

  • (String) —

    file name



150
151
152
153
 
# File 'lib/relaton_3gpp/data_fetcher.rb', line 150

def file_name(bib)
  name = bib.docnumber.gsub(/[\s,:\/]/, "_").squeeze("_").upcase
  File.join @output, "#{name}.#{@ext}"
end

#get_file(renewal) ⇒ String?

Get file from FTP. If file does not exist or changed, return nil

Parameters:

  • renewal (Boolean) —

    force to update all documents

Returns:

  • (String, nil) —

    file name



69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
 
# File 'lib/relaton_3gpp/data_fetcher.rb', line 69

def get_file(renewal) # rubocop:disable Metrics/MethodLength, Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
  @current = YAML.load_file CURRENT if File.exist? CURRENT
  @current ||= {}
  n = 0
  begin
    ftp = Net::FTP.new("www.3gpp.org")
    ftp.resume = true
    ftp.
    ftp.chdir "/Information/Databases/"
    file_path = ftp.list("*.csv").first
    return unless file_path

    d, t, _, file = file_path.split
    dt = DateTime.strptime("#{d} #{t}", "%m-%d-%y %I:%M%p")
    if !renewal && file == @current["file"] && !@current["date"].empty? && dt == DateTime.parse(@current["date"])
      return
    end

    tmp_file = File.join Dir.tmpdir, "3gpp.csv"
    ftp.get(file, tmp_file)
  rescue Net::ReadTimeout => e
    n += 1
    retry if n < 5
    raise e
  end
  @current["file"] = file
  @current["date"] = dt.to_s
  tmp_file
end

#index ⇒ Object 



21
22
23
 
# File 'lib/relaton_3gpp/data_fetcher.rb', line 21

def index
  @index ||= Relaton::Index.find_or_create "3gpp", file: "index-v1.yaml"
end

#save_doc(bib) ⇒ Object

Save document to file

Parameters:

  • bib (RelatonW3c::W3cBibliographicItem, nil) —

    bibliographic item



125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
 
# File 'lib/relaton_3gpp/data_fetcher.rb', line 125

def save_doc(bib) # rubocop:disable Metrics/MethodLength
  return unless bib

  c = case @format
      when "xml" then bib.to_xml(bibdata: true)
      when "yaml" then bib.to_hash.to_yaml
      else bib.send("to_#{@format}")
      end
  file = file_name(bib)
  if @files.include? file
    Util.warn "File #{file} already exists. Document: #{bib.docnumber}"
  else
    @files << file
  end
  index.add_or_update bib.docnumber, file
  File.write file, c, encoding: "UTF-8"
end