nist-pubid 0.1.3 → 0.1.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 48d1d54a6499cb679c33cfa2a7cc2a682b6b41e9c66aab0c301375c5df28e3f3
-  data.tar.gz: 64fa186792b6564f8afc4f2388f4e2859b04dbcdca5f7cd00e40b192c7f9fc4a
+  metadata.gz: c894fedad676cd7d989d771acdceeacfb5b8e55afd9c24777fe127593820ad52
+  data.tar.gz: def901ca2a772e9594b487a4af7d6db8d0ce97c801bb2f3c2324a5b1d4a00f96
 SHA512:
-  metadata.gz: 25197a16bbaa8b4799d0279cf19303de616da1e34452997df576be22cb459aa3b00f538af680d9fa37330d6c7358587d56173ab2bc10f57755774537185c205e
-  data.tar.gz: 3e6ae4db9939f1f415ae56ec5aba095a41706157143740a6412d78b560cf850854a779104b22edbc278234fc657ed3289aaf1ab832ef909d58b9d465ec86170b
+  metadata.gz: dc67a75ba8a03228d68e21d7967806a4b83663cfac77641b1050d807d2b84fe42b12da64f588205e63056cdb3641a72181d28599f076cb17a22bac933cd74c1c
+  data.tar.gz: 3e00e519a9fe89fc2adb276416d31b15669ae4211ae44508d8136b5f3a4e6c49e232508933b777a652ea74bc85883dea8e18c62030a2ae17ed4bed9d308e8ac9

data/.github/workflows/release.yml

@@ -16,7 +16,7 @@ jobs:
   release:
     runs-on: ubuntu-18.04
     steps:
-      - uses: actions/checkout@v1
+      - uses: actions/checkout@v2
 
       - uses: ruby/setup-ruby@v1
         with:

data/README.adoc

@@ -4,7 +4,7 @@
 
 NIST publications are numbered according to an identification scheme.
 
-This gem implements a mechanism to parse and utilize NIST publication identifers
+This gem implements a mechanism to parse and utilize NIST publication identifiers
 provided in the https://github.com/usnistgov/NIST-Tech-Pubs[NIST-Tech-Pubs]
 repository.
 
@@ -296,9 +296,10 @@ A "Part" can also be indicated by an appended alphabetic character to the end.
 | Section | Sec. | `sec`
 | Supplement | Suppl. | `sup`
 | Index | Index | `indx`
-| Addendum | Add. | `add` (TBD with NIST)
+| Addendum | Add. | `add` (TBC with NIST)
 | Insert | Ins. | `ins` (TBC with NIST)
 | Errata | Err. | `err` (TBC with NIST)
+| Appendix | App. | `app` (TBC with NIST)
 
 |===
 

data/exe/nist-pubid

@@ -0,0 +1,25 @@
+#!/usr/bin/env ruby
+
+require 'rubygems'
+require 'bundler/setup'
+require 'nist_pubid'
+require 'optparse'
+
+OptionParser.new do |opts|
+  opts.banner = "Usage: nist_pubid [ -s | PubID ]"
+  opts.on("-s", "--status", "Get a NIST Tech Pubs list with final PubID") do
+    puts "Current document ID | DOI | Title | Final PubID"
+    NistPubid::NistTechPubs.status.each do |row|
+      puts row.join(" | ")
+    end
+  end
+
+  opts.on("-h", "--help", "Prints this help") do
+    puts opts
+    exit
+  end
+
+end.parse!
+
+code = ARGV.join(" ")
+puts NistPubid::Document.parse(code).to_s(:short) unless code.empty?

data/lib/nist_pubid/document.rb

@@ -35,10 +35,40 @@ EDITION_DESC = {
   mr: "e",
 }.freeze
 
+SUPPLEMENT_DESC = {
+  long: " Supplement ",
+  abbrev: " Suppl. ",
+  short: "sup",
+  mr: "sup",
+}.freeze
+
+SECTION_DESC = {
+  long: " Section ",
+  abbrev: " Sec. ",
+  short: "sec",
+  mr: "sec",
+}.freeze
+
+APPENDIX_DESC = {
+  long: " Appendix ",
+  abbrev: " App. ",
+  short: "app",
+  mr: "app",
+}.freeze
+
+ERRATA_DESC = {
+  long: " Errata ",
+  abbrev: " Err. ",
+  short: "-err",
+  mr: "-err",
+}.freeze
+
 module NistPubid
   class Document
     attr_accessor :serie, :code, :revision, :publisher, :version, :volume,
-                  :part, :addendum, :stage, :translation, :update, :edition
+                  :part, :addendum, :stage, :translation, :update_number,
+                  :edition, :supplement, :update_year, :section, :appendix,
+                  :errata
 
     def initialize(publisher:, serie:, docnumber:, **opts)
       @publisher = Publisher.new(publisher: publisher)
@@ -47,28 +77,128 @@ module NistPubid
       opts.each { |key, value| send("#{key}=", value) }
     end
 
+    # returns weight based on amount of defined attributes
+    def weight
+      instance_variables.inject(0) do |sum, var|
+        sum + (instance_variable_get(var).nil? ? 0 : 1)
+      end
+    end
+
+    def merge(document)
+      instance_variables.each do |var|
+        val = document.instance_variable_get(var)
+        instance_variable_set(var, val) unless val.nil?
+      end
+
+      self
+    end
+
+    def self.update_old_code(code)
+      code = code.gsub("FIPS", "FIPS PUB") unless code.include?("FIPS PUB")
+      code.gsub("NBS MONO", "NBS MN").gsub("NIST MONO", "NIST MN")
+        .gsub("NIST MP", "NBS MP")
+        .gsub("NIST SP 304a-2017", "NIST SP 304A-2017")
+        .gsub("NIST SP 260-162 2006ed.", "NIST SP 260-162e2006")
+        .gsub("NBS CIRC 154suprev", "NBS CIRC 154r1sup")
+        .gsub("NIST SP 260-126 rev 2013", "NIST SP 260-126r2013")
+        .gsub(/(?<=NBS MP )(\d+)\((\d+)\)/, '\1e\2')
+        .gsub(/(?<=\d)es/, "(spa)")
+        .gsub(/(?<=\d)chi/, "(zho)")
+        .gsub(/(?<=\d)viet/, "(vie)")
+        .gsub(/(?<=\d)port/, "(por)")
+        # .gsub(/^LCIRC/, "NBS LC")
+    end
+
     def self.parse(code)
+      code = update_old_code(code)
       matches = {
         publisher: match(Publisher.regexp, code) || "NIST",
-        serie: match(Serie.regexp, code)&.gsub(/\./, " "),
+        serie: match(Serie.regexp, code),
         stage: Stage.parse(code),
-        docnumber: match(/(?<=\.|\s)[0-9-]{3,}[A-Z]?/, code),
-        part: /(?<=(\.))?pt(?(1)-)([A-Z\d]+)/.match(code)&.[](2),
-        volume: /(?<=(\.))?v(?(1)-)(\d+)/.match(code)&.[](2),
-        version: match(/(?<=(\.))?ver(?(1)[-\d]|[.\d])+/, code)&.gsub(/-/, "."),
-        revision: /(?<=[^a-z])(?<=(\.))?(?:r(?(1)-)|Rev\.\s)(\d+)/
+        part: /(?<=(\.))?(?<![a-z])+(?:pt|Pt|p)(?(1)-)([A-Z\d]+)/.match(code)
+                &.[](2),
+        version:
+          /(?<=\.)?(?:(?:ver)((?(1)[-\d]|[.\d])+|\d+)|(?:v)(\d+\.[.\d]+))/
+            .match(code).to_a[1..-1]&.compact&.first&.gsub(/-/, "."),
+        revision: /[\da](?:r|Rev\.\s|([0-9]+[A-Za-z]*-[0-9]+[A-Za-z]*-))([\da]+)/
           .match(code)&.[](2),
-        addendum: match(/(?<=(\.))?(add(?(1)-)\d+|Addendum)/, code),
-        update: match(/(?<=Upd\s)([\d:]+)/, code),
+        addendum: match(/(?<=(\.))?(add(?:-\d+)?|Addendum)/, code),
         edition: /(?<=[^a-z])(?<=(\.))?(?:e(?(1)-)|Ed\.\s)(\d+)/
           .match(code)&.[](2),
+        section: /(?<=sec)\d+/.match(code)&.to_s,
+        appendix: /\d+app/.match(code)&.to_s,
+        errata: /-errata/.match(code)&.to_s
       }
-      code = code.gsub(matches[:stage].original_code, "") unless matches[:stage].nil?
+      supplement = /(?:(?:supp?)-?(\d*)|Supplement|Suppl.)/
+        .match(code)
+      unless supplement.nil?
+        matches[:supplement] = supplement[1].nil? ? "" : supplement[1]
+      end
+
+      update = code.scan(/((?<=Upd)\s?[\d:]+|-upd)-?(\d*)/).first
+
+      (matches[:update_number], matches[:update_year]) = update if update
+
+      unless matches[:serie]
+        raise Errors::ParseError.new("failed to parse serie for #{code}")
+      end
+
+      unless matches[:stage].nil?
+        code = code.gsub(matches[:stage].original_code, "")
+      end
+
+      unless ["NBS CSM", "NBS CS"].include?(matches[:serie])
+        matches[:volume] = /(?<=(\.))?v(?(1)-)(\d+)(?!\.\d+)/.match(code)&.[](2)
+      end
+
+      matches[:revision] = nil if matches[:addendum]
+
+      matches[:docnumber] = parse_docnumber(matches[:serie], code)
+
+      matches[:serie] = SERIES["mr"].invert[matches[:serie]] || matches[:serie]
+      # matches[:serie].gsub!(/\./, " ")
       matches[:translation] = match(/(?<=\()\w{3}(?=\))/, code)
 
       new(**matches)
     end
 
+    def self.parse_docnumber(serie, code)
+      localities = "[Pp]t\\d+|r(?:\\d+|[A-Za-z]?)|e\\d+|p|v|sec\\d+"
+      excluded_parts = "(?!#{localities}|supp?)"
+
+      if ["NBS CSM", "NBS CS"].include?(serie)
+        docnumber = /v(\d+)n(\d+)/.match(code).to_a[1..-1]&.join("-")
+      else
+        # match docnumbers with localities in the first part, like NBS CIRC 11e2-1915
+        docnumber =
+          /(?:#{serie.gsub(" ", "\s|\.")})(?:\s|\.)?([0-9]+)(?:#{localities})(-[0-9]+)?/
+            .match(code)&.captures&.join
+
+        docnumber ||=
+          /(?:#{serie.gsub(" ", "\s|\.")})(?:\s|\.)? # match serie
+           ([0-9]+ # first part of report number
+             (?:#{excluded_parts}[A-Za-z]+)? # with letter but without localities
+             (?:-m)? # for NBS CRPL 4-m-5
+             (?:-[A-Z]+)? # for NIST SP 1075-NCNR, NIST SP 1113-BFRL, etc
+             (?:-[0-9.]+)? # second part
+             (?:
+               (?: # only big letter
+                 (#{excluded_parts}[A-Z]|(?![a-z]))+|#{excluded_parts}[a-z]?|#{excluded_parts}[a-z]+
+               )? # or small letter but without localities
+             )
+           )/x
+            .match(code)&.[](1)&.upcase
+      end
+
+      unless docnumber
+        raise Errors::ParseError.new(
+          "failed to parse document identifier for #{code}",
+        )
+      end
+
+      docnumber
+    end
+
     def self.match(regex, code)
       regex.match(code)&.to_s
     end
@@ -78,6 +208,7 @@ module NistPubid
       result += " " unless format == :short || stage.nil?
       result += "#{stage&.to_s(format)}"\
                 " #{code}#{render_part(format)}#{render_edition(format)}"\
+                "#{render_localities(format)}"\
                 "#{render_update(format)}#{render_translation(format)}"
       result = render_addendum(result, format)
 
@@ -87,13 +218,14 @@ module NistPubid
     end
 
     def render_serie(format)
-      return serie.to_s(format) if %i[mr short].include?(format)
+      if serie.to_s(format).include?(publisher.to_s(format))
+        return serie.to_s(format)
+      end
 
       "#{publisher.to_s(format)} #{serie.to_s(format)}"
     end
 
     def render_part(format)
-      # TODO: Section, Supplement, Index, Insert, Errata
       result = ""
       result += "#{VOLUME_DESC[format]}#{volume}" unless volume.nil?
       result += "#{PART_DESC[format]}#{part}" unless part.nil?
@@ -102,24 +234,50 @@ module NistPubid
 
     def render_edition(format)
       result = ""
-      result += "#{REVISION_DESC[format]}#{revision}" unless revision.nil?
+      if revision
+        result += if %i[long abbrev].include?(format) ||
+            [volume, part, supplement, version, edition].any?
+                    "#{REVISION_DESC[format]}#{revision}"
+                  else
+                    "-#{revision}"
+                  end
+      end
       result += "#{VERSION_DESC[format]}#{version}" unless version.nil?
       result += "#{EDITION_DESC[format]}#{edition}" unless edition.nil?
       result
     end
 
+    def render_localities(format)
+      # TODO: Index, Insert
+
+      result = ""
+      result += "#{SUPPLEMENT_DESC[format]}#{supplement}" unless supplement.nil?
+      result += "#{SECTION_DESC[format]}#{section}" unless section.nil?
+      result += "#{APPENDIX_DESC[format]}" unless appendix.nil?
+      result += "#{ERRATA_DESC[format]}" unless errata.nil?
+
+      result
+    end
+
     def render_update(format)
-      return "" if update.nil?
+      return "" if update_year.nil?
+
+      if update_number.match?(/\d+/)
+        update_text = update_number
+        update_text += "-#{update_year}" if update_year
+      else
+        update_text = "1"
+      end
 
       case format
       when :long
-        " Update #{update}"
+        " Update #{update_text}"
       when :abbrev
-        " Upd. #{update}"
+        " Upd. #{update_text}"
       when :short
-        "/Upd #{update}"
+        "/Upd#{update_text}"
       when :mr
-        ".u#{update.gsub(":", "-")}"
+        ".u#{update_text}"
       end
     end
 

data/lib/nist_pubid/errors.rb

@@ -0,0 +1,5 @@
+module NistPubid
+  module Errors
+    class ParseError < StandardError; end
+  end
+end

data/lib/nist_pubid/nist_tech_pubs.rb

@@ -0,0 +1,92 @@
+require "relaton_nist/data_fetcher"
+require "nokogiri"
+require "open-uri"
+
+module NistPubid
+  class NistTechPubs
+    URL = "https://raw.githubusercontent.com/usnistgov/NIST-Tech-Pubs/nist-pages/xml/allrecords.xml".freeze
+
+    @converted_id = @converted_doi = {}
+
+    class << self
+
+      attr_accessor :documents, :converted_id, :converted_doi
+
+      def fetch
+        @documents ||= Nokogiri::XML(URI.open(URL))
+          .xpath("/body/query/doi_record/report-paper/report-paper_metadata")
+          .map { |doc| parse_docid doc }
+      rescue StandardError => e
+        warn e.message
+        []
+      end
+
+      def convert(doc)
+        id = @converted_id[doc[:id]] ||= NistPubid::Document.parse(doc[:id])
+        return id.to_s(:short) unless doc.key?(:doi)
+
+        begin
+          doi = @converted_doi[doc[:doi]] ||=
+            NistPubid::Document.parse(doc[:doi])
+        rescue Errors::ParseError
+          return id.to_s(:short)
+        end
+        # return more complete pubid
+        id.merge(doi).to_s(:short)
+      rescue Errors::ParseError
+        @converted_doi[doc[:doi]] ||= NistPubid::Document.parse(doc[:doi])
+          .to_s(:short)
+      end
+
+      def parse_docid(doc)
+        id = doc.at("publisher_item/item_number", "publisher_item/identifier")
+          .text.sub(%r{^/}, "")
+        doi = doc.at("doi_data/doi").text.gsub("10.6028/", "")
+        title = doc.at("titles/title").text
+        title += " #{doc.at('titles/subtitle').text}" if doc.at("titles/subtitle")
+        case doi
+        when "10.6028/NBS.CIRC.12e2revjune" then id.sub!("13e", "12e")
+        when "10.6028/NBS.CIRC.36e2" then id.sub!("46e", "36e")
+        when "10.6028/NBS.HB.67suppJune1967" then id.sub!("1965", "1967")
+        when "10.6028/NBS.HB.105-1r1990" then id.sub!("105-1-1990", "105-1r1990")
+        when "10.6028/NIST.HB.150-10-1995" then id.sub!(/150-10$/, "150-10-1995")
+        end
+
+        { id: id, doi: doi, title: title }
+      end
+
+      def comply_with_pubid
+        fetch.select do |doc|
+          convert(doc) == doc[:id]
+        rescue Errors::ParseError
+          false
+        end
+      end
+
+      def different_with_pubid
+        fetch.reject do |doc|
+          convert(doc) == doc[:id]
+        rescue Errors::ParseError
+          true
+        end
+      end
+
+      def parse_fail_with_pubid
+        fetch.select do |doc|
+          convert(doc) && false
+        rescue Errors::ParseError
+          true
+        end
+      end
+
+      # returning current document id, doi, title and final PubID
+      def status
+        fetch.map do |doc|
+          [doc[:id], doc[:doi], doc[:title], convert(doc)]
+        rescue Errors::ParseError
+          [doc[:id], doc[:doi], doc[:title], "parse error"]
+        end
+      end
+    end
+  end
+end

data/lib/nist_pubid/serie.rb

@@ -18,7 +18,9 @@ module NistPubid
     end
 
     def self.regexp
-      /(#{(SERIES["long"].keys + SERIES["mr"].values + ["NISTIR"]).join('|')})(?=\.|\s|\()/
+      /(#{(SERIES["long"].keys + SERIES["mr"].values
+            .map { |v| v.gsub(".", '\.') } + ["NISTIR"])
+            .sort_by(&:length).reverse.join('|')})/
     end
   end
 end

data/lib/nist_pubid/version.rb

@@ -1,3 +1,3 @@
 module NistPubid
-  VERSION = "0.1.3".freeze
+  VERSION = "0.1.4".freeze
 end

data/lib/nist_pubid.rb

@@ -5,3 +5,5 @@ require "nist_pubid/document"
 require "nist_pubid/publisher"
 require "nist_pubid/serie"
 require "nist_pubid/stage"
+require "nist_pubid/errors"
+require "nist_pubid/nist_tech_pubs"

data/nist_pubid.gemspec

@@ -28,4 +28,9 @@ Gem::Specification.new do |spec|
 
   spec.add_development_dependency "rake", "~> 13.0"
   spec.add_development_dependency "rspec", "~> 3.0"
+  spec.add_development_dependency "vcr"
+  spec.add_development_dependency "webmock"
+
+  spec.add_dependency "relaton-nist"
+  spec.add_dependency "nokogiri"
 end

data/series.yaml

@@ -87,7 +87,7 @@ mr:
   NBS BH: NBS.BH
   NBS CRPL: NBS.CRPL
   NBS CRPL-F-A: NBS.CRPL-F-A
-  NBS CRPL-F-B: NBS CRPL-F-B245
+  NBS CRPL-F-B: NBS CRPL-F-B
   NBS IP: NBS.IP
   NBS CIRC: NBS.CIRC
   NBS CIS: NBS.CIS

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: nist-pubid
 version: !ruby/object:Gem::Version
-  version: 0.1.3
+  version: 0.1.4
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2021-12-09 00:00:00.000000000 Z
+date: 2022-01-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
@@ -38,10 +38,67 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '3.0'
+- !ruby/object:Gem::Dependency
+  name: vcr
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: webmock
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: relaton-nist
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: nokogiri
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
 description: Library to generate, parse and manipulate NIST PubID.
 email:
 - open.source@ribose.com
-executables: []
+executables:
+- nist-pubid
 extensions: []
 extra_rdoc_files:
 - README.adoc
@@ -55,8 +112,11 @@ files:
 - LICENSE.txt
 - README.adoc
 - Rakefile
+- exe/nist-pubid
 - lib/nist_pubid.rb
 - lib/nist_pubid/document.rb
+- lib/nist_pubid/errors.rb
+- lib/nist_pubid/nist_tech_pubs.rb
 - lib/nist_pubid/publisher.rb
 - lib/nist_pubid/serie.rb
 - lib/nist_pubid/stage.rb
@@ -84,7 +144,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.1.4
+rubygems_version: 3.3.3
 signing_key:
 specification_version: 4
 summary: Library to generate, parse and manipulate NIST PubID.