Module: RelatonItu::Scrapper

Defined in:
lib/relaton_itu/scrapper.rb

Overview

Scrapper.

Constant Summary collapse

ROMAN_MONTHS = 
%w[I II III IV V VI VII VIII IX X XI XII].freeze
TYPES = 
{
  "ISO" => "international-standard",
  "TS" => "technicalSpecification",
  "TR" => "technicalReport",
  "PAS" => "publiclyAvailableSpecification",
  "AWI" => "appruvedWorkItem",
  "CD" => "committeeDraft",
  "FDIS" => "finalDraftInternationalStandard",
  "NP" => "newProposal",
  "DIS" => "draftInternationalStandard",
  "WD" => "workingDraft",
  "R" => "recommendation",
  "Guide" => "guide",
}.freeze

Class Method Summary collapse

Class Method Details

.parse_page(hit, imp: false) ⇒ Hash

Parse page.

Parameters:

Returns:

  • (Hash) 


30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
 
# File 'lib/relaton_itu/scrapper.rb', line 30

def parse_page(hit, imp: false) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
  doc = get_page hit
  return unless doc.code == "200"

  if imp
    a = doc.at "//span[contains(@id, 'tab_ig_uc_rec')]/a"
    return unless a

    doc = get_page hit, a[:href].to_s
  end

  # Fetch edition.
  edition = doc.at("//table/tr/td[contains(@style,'color: white')]/span[contains(@id, 'Label8')]/b")&.text
  docid = fetch_docid(doc, hit)

  ItuBibliographicItem.new(
    id: fetch_id(docid),
    fetched: Date.today.to_s,
    type: "standard",
    docid: docid,
    edition: edition,
    language: ["en"],
    script: ["Latn"],
    title: fetch_titles(doc),
    doctype: DocumentType.new(type: hit.hit[:type]),
    docstatus: fetch_status(doc),
    ics: [], # fetch_ics(doc),
    date: fetch_dates(doc),
    contributor: fetch_contributors(hit.hit[:code]),
    editorialgroup: fetch_workgroup(hit.hit[:code], doc),
    abstract: fetch_abstract(doc, hit),
    copyright: fetch_copyright(hit.hit[:code], doc),
    link: fetch_link(doc),
    relation: fetch_relations(doc),
    place: ["Geneva"],
  )
end