Class: Pubid::Iso::UrnParser

Inherits:
Object
  • Object
show all
Defined in:
lib/pubid/iso/urn_parser.rb

Overview

Parses RFC 5141-bis compliant URNs into ISO identifiers

URN format: urn:iso:std:publisher:type:number:year:supplements

Examples:

  • urn:iso:std:iso:9001:2019

  • urn:iso:std:iso:tr:9001:2019

  • urn:iso:std:iso:9001:2019:stage-40.00

  • urn:iso:std:iso:9001:2019:amd:1:2020

  • urn:iso:std:iso-iec:27001:2013

Constant Summary collapse

TYPED_STAGE_REVERSE_MAP =

Reverse mappings from URN format to PubID components

{
  "WD" => :wd,
  "WDS" => :wds,
  "CD" => :cd,
  "CDV" => :cdv,
  "DIS" => :dis,
  "FDIS" => :fdis,
  "PDAM" => :pdam,
  "DAM" => :dam,
  "FDAM" => :fdamd,
  "DCOR" => :dcor,
  "FDCOR" => :fdcor,
  "CDTS" => :cdts,
  "DTS" => :dts,
  "FDTS" => :fdts,
}.freeze
SUPPLEMENT_TYPE_MAP =
{
  "amd" => :amd,
  "cor" => :cor,
  "sup" => :sup,
  "add" => :add,
}.freeze
TYPE_CODE_REVERSE_MAP =
{
  "tr" => :tr,
  "ts" => :ts,
  "pas" => :pas,
  "guide" => :guide,
  "dir" => :dir,
  "dir-sup" => :dir_sup,
  "iwa-sup" => :iwa_sup,
  "isp" => :isp,
  "iwa" => :iwa,
  "r" => :r,
  "data" => :data,
}.freeze

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.parse(urn) ⇒ Identifier

Parse ISO URN string

Parameters:

  • urn (String)

    URN string to parse

Returns:



58
59
60
# File 'lib/pubid/iso/urn_parser.rb', line 58

def self.parse(urn)
  new.parse_urn(urn)
end

Instance Method Details

#parse_urn(urn) ⇒ Identifier

Parse URN string into identifier

Parameters:

  • urn (String)

    URN string

Returns:



65
66
67
68
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
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
# File 'lib/pubid/iso/urn_parser.rb', line 65

def parse_urn(urn)
  # Remove urn:iso:std: prefix
  unless urn.start_with?("urn:iso:std:")
    raise ArgumentError,
          "Invalid ISO URN: #{urn}"
  end

  parts = urn.sub("urn:iso:std:", "").split(":")

  # Parse publisher(s) - first part
  publishers = parse_publisher(parts.shift)

  # Parse type - optional (defaults to IS)
  type_code = nil
  type_code = parse_type(parts.first) if parts.first && TYPE_CODE_REVERSE_MAP.key?(parts.first.downcase)
  parts.shift if type_code

  # Parse number
  number_part = parts.shift
  number, part, subpart = parse_number_part(number_part)

  # Parse year if present (4-digit year)
  year = nil
  if parts.first&.match(/^\d{4}$/)
    year = parts.shift
  end

  # Handle URN-style part notation (:-22, :-5-1-1, etc.)
  # These can come before or after year/edition in URN format
  # Note: The parts are split by ':', so :-22 becomes two parts: "10164" and "-22"
  if parts.first&.start_with?("-")
    part_str = parts.shift
    # Re-parse the number part with the URN-style part
    number, part, subpart = parse_number_part("#{number}#{part_str}")
  end

  # Parse year if present (4-digit year) - may come after part
  if year.nil? && parts.first&.match(/^\d{4}$/)
    year = parts.shift
  end

  # Parse edition if present (ed-N format) - comes after year in URN
  edition = nil
  if parts.first&.start_with?("ed-")
    edition = parts.shift.sub("ed-", "").to_i
  end

  # Parse language if present (2-letter codes like "en" or comma-separated)
  languages = nil
  if parts.first && !parts.first.match(/^\d+$/) &&
      !parts.first.match?(/^(amd|cor|sup|add|v\d+|stage-|ed-)/i) &&
      !TYPED_STAGE_REVERSE_MAP.key?(parts.first.upcase) &&
      !parts.first.match?(/^[A-Z]+\.\d+$/i)
    languages = parse_languages(parts.shift)
  end

  # Check for stage (stage-XX.XX or typed stage like WD, CD, etc.)
  # IMPORTANT: This must come AFTER edition/part/language parsing but BEFORE supplements parsing
  stage_code = nil
  stage_iteration = nil
  harmonized_stage_code = nil # Track full harmonized code for lookup
  stage_from_abbr = nil # Track stage code from typed abbreviation
  if parts.first&.start_with?("stage-")
    stage_str = parts.shift
    stage_code, stage_iteration = parse_stage_code(stage_str)
    # Set harmonized_stage_code AFTER parse_stage_code has stripped .vX suffix
    harmonized_stage_code = stage_str.sub("stage-", "").sub(/\.v\d+$/i,
                                                            "")
  elsif TYPED_STAGE_REVERSE_MAP.key?(parts.first&.upcase) ||
      (parts.first&.match?(/^[A-Za-z]+\.\d+$/) &&
       TYPED_STAGE_REVERSE_MAP.key?(parts.first.upcase.split(".").first))
    stage_abbr = parts.shift.upcase
    # Check for iteration (WD.2 format)
    if stage_abbr.include?(".")
      abbr_part, iteration_part = stage_abbr.split(".")
      stage_code = TYPED_STAGE_REVERSE_MAP[abbr_part]
      stage_from_abbr = stage_code
      stage_iteration = iteration_part.to_i
    else
      stage_code = TYPED_STAGE_REVERSE_MAP[stage_abbr]
      stage_from_abbr = stage_code
    end
  end

  # Check for supplements (amd, cor, sup, add)
  supplements = []
  while parts.any?
    supp_type = nil
    supp_number = nil
    supp_year = nil
    supp_stage = nil

    # Check for supplement stage
    if parts.first&.start_with?("stage-")
      supp_stage_data = parts.shift
      supp_stage, = parse_stage_code(supp_stage_data)
    elsif TYPED_STAGE_REVERSE_MAP.key?(parts.first&.upcase)
      supp_stage_abbr = parts.shift.upcase
      supp_stage = TYPED_STAGE_REVERSE_MAP[supp_stage_abbr]
    end

    # Check for supplement type (amd, cor, sup, add)
    if SUPPLEMENT_TYPE_MAP.key?(parts.first&.downcase)
      supp_type = SUPPLEMENT_TYPE_MAP[parts.shift.downcase]
    end

    # Check for year or supplement number
    if parts.first&.match(/^\d+$/)
      if parts.first&.match(/^\d{4}$/)
        # 4 digits = year
        supp_year = parts.shift
      else
        # 1-3 digits = supplement number
        supp_number = parts.shift.to_i
      end
    end

    # Check for version (v1, v2, etc.) after year or number
    if parts.first&.start_with?("v")
      version_str = parts.shift
      supp_number = version_str.sub("v", "").to_i
      # Handle version with iteration (v1.2)
      if supp_number.to_s.include?(".")
        supp_number, = supp_number.to_s.split(".")
        supp_number = supp_number.to_i
      end
    end

    # Next part might be year if not already set
    if supp_year.nil? && parts.first&.match(/^\d{4}$/)
      supp_year = parts.shift
    end

    # Check for language after supplement
    supp_languages = nil
    if parts.first && !parts.first.match(/^\d+$/) && !parts.first.match?(/^(amd|cor|sup|add|v\d+|stage-)/i)
      supp_languages = parse_languages(parts.shift)
    end

    supplements << {
      type: supp_type,
      number: supp_number,
      year: supp_year,
      stage: supp_stage,
      languages: supp_languages,
    }
  end

  # Build the identifier hash
  build_identifier(publishers, number, part, subpart, type_code, stage_code, stage_iteration,
                   harmonized_stage_code, stage_from_abbr, year, edition, languages, supplements)
end