Module: RelatonIec::Scrapper

Defined in:
lib/relaton_iec/scrapper.rb

Overview

Scrapper. rubocop:disable Metrics/ModuleLength

Constant Summary collapse

DOMAIN = 
"https://webstore.iec.ch"
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_data) ⇒ Hash

Parse page. rubocop:disable Metrics/AbcSize, Metrics/MethodLength

Parameters:

  • hit (Hash)

Returns:

  • (Hash) 


51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
 
# File 'lib/relaton_iec/scrapper.rb', line 51

def parse_page(hit_data)
  doc = get_page hit_data[:url]

  # Fetch edition.
  edition = doc.at("//th[contains(., 'Edition')]/following-sibling::td/span").text

  status, relations = fetch_status_relations hit_data[:url]

  RelatonIsoBib::IsoBibliographicItem.new(
    docid: [RelatonBib::DocumentIdentifier.new(id: hit_data[:code], type: "IEC")],
    structuredidentifier: fetch_structuredidentifier(doc),
    edition: edition,
    language: ["en"],
    script: ["Latn"],
    titles: fetch_titles(hit_data),
    type: fetch_type(doc),
    docstatus: status,
    ics: fetch_ics(doc),
    dates: fetch_dates(doc),
    contributors: fetch_contributors(hit_data[:code]),
    editorialgroup: fetch_workgroup(doc),
    abstract: fetch_abstract(doc),
    copyright: fetch_copyright(hit_data[:code], doc),
    link: fetch_link(doc, hit_data[:url]),
    relations: relations,
  )
end