nist-pubid 0.1.5 → 0.1.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 818a96e99a2e198ef626234c1b175c310104708a84198439e327e7c0bccfec8e
-  data.tar.gz: c2da763a9a97d5735b0d39641417f89f622620b10aa6f871f25acc4c8cf5850d
+  metadata.gz: d6cbdcaa4e1cfa7c75b6709f34cfd19670e8a3b2d3bd31473cbc5a5d439843b7
+  data.tar.gz: 4dcb0a4b147990872fe4f72c8eecac96ba3a865bcaa477a5158404ab04fc8e42
 SHA512:
-  metadata.gz: df7f8caaa70c6ebf7bdd4ff2e4506cac98180fe87037a2959baae6b98b7bc394ff7906a9d37e4d4420026078b263e1a8200e02c42c02e0fc2701f0cf007c027b
-  data.tar.gz: 7b21b01c0fbeef957e28b7a68c541c6cc7cf4e4de2eaae446493705bfa91c7a1c4bcf2ffcf0d7d0d960d0b04aa7ec5295d3f461e81dbf9b2096606d880be62d2
+  metadata.gz: b6331d9887c149a7dbcf49c46a0ce7c121bf973cd239e3467658065820daf8c18f752e68b977d52e08155dfa2d947d9be5fa80bb431f462453ed337e02fc1616
+  data.tar.gz: 35660adeef1ec90862ada64b2150162008d0b25cf01a8e9a7b15f909b6ac64265a04cc2bdf1da7ea9a7d08e99754eb8d49910de9d19119b4548b40f3870876db

data/exe/nist-pubid

@@ -1,25 +1,54 @@
 #!/usr/bin/env ruby
 
-require 'rubygems'
-require 'bundler/setup'
-require_relative '../lib/nist_pubid'
-require 'optparse'
+require "rubygems"
+require "bundler/setup"
+require_relative "../lib/nist_pubid"
+require "thor"
+require "csv"
 
-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(" | ")
+class NistPubidCLI < Thor
+  desc "report", "Create report for NIST Tech Pubs database (fetches from GitHub)"
+  option :csv, type: :boolean, desc: "Export to CSV format"
+  def report
+    if options[:csv]
+      NistPubid::NistTechPubs.status.each do |doc|
+        puts [doc[:finalPubId] == doc[:id] && doc[:mr] == doc[:doi] ? "-" : "+",
+              doc[:finalPubId], doc[:id], doc[:doi], doc[:mr], doc[:title]]
+          .to_csv
+      end
+    else
+      puts "Changed | New PubID | Document ID | DOI | New MR | Title + subtitle"
+      NistPubid::NistTechPubs.status.each do |doc|
+        puts "#{doc[:finalPubId] == doc[:id] && doc[:mr] == doc[:doi] ? ' -' : '✅'}"\
+             " | #{doc[:finalPubId]} | #{doc[:id]} | #{doc[:doi]} | #{doc[:mr]} | #{doc[:title]}"
+      end
     end
   end
 
-  opts.on("-h", "--help", "Prints this help") do
-    puts opts
-    exit
-  end
+  desc "convert", "Convert legacy NIST Tech Pubs ID to NIST PubID"
+  option :style, aliases: "-s", type: :string,
+                 desc: "Convert to PubID style (short|long|mr|abbrev)",
+                 default: "short"
+  option :format, aliases: "-f", type: :string,
+                  desc: "Render in format (JSON, string)",
+                  default: "string"
+  def convert(code)
+    unless %w[mr long short abbrev].include?(options[:style])
+      raise "Invalid PubID style"
+    end
+
+    raise "wrong conversion format" unless %w[string json].include? options[:format]
 
-end.parse!
+    unless code.empty?
+      if options[:format] == "string"
+        puts NistPubid::Document.parse(code).to_s(options[:style].to_sym)
+      else
+        puts NistPubid::Document.parse(code).to_json
+      end
+    end
+  rescue NistPubid::Errors::ParseError
+    puts "[Error] This does not seem to be a valid NIST Tech Pubs legacy identifier"
+  end
+end
 
-code = ARGV.join(" ")
-puts NistPubid::Document.parse(code).to_s(:short) unless code.empty?
+NistPubidCLI.start(ARGV)

data/lib/nist_pubid/document.rb

@@ -1,4 +1,5 @@
 # frozen_string_literal: true
+require "json"
 
 REVISION_DESC = {
   long: ", Revision ",
@@ -59,8 +60,22 @@ APPENDIX_DESC = {
 ERRATA_DESC = {
   long: " Errata ",
   abbrev: " Err. ",
-  short: "-err",
-  mr: "-err",
+  short: "err",
+  mr: "err",
+}.freeze
+
+INDEX_DESC = {
+  long: " Index ",
+  abbrev: " Index. ",
+  short: "indx",
+  mr: "indx",
+}.freeze
+
+INSERT_DESC = {
+  long: " Insert ",
+  abbrev: " Ins. ",
+  short: "ins",
+  mr: "ins",
 }.freeze
 
 module NistPubid
@@ -68,7 +83,7 @@ module NistPubid
     attr_accessor :serie, :code, :revision, :publisher, :version, :volume,
                   :part, :addendum, :stage, :translation, :update_number,
                   :edition, :supplement, :update_year, :section, :appendix,
-                  :errata
+                  :errata, :index, :insert
 
     def initialize(publisher:, serie:, docnumber:, **opts)
       @publisher = Publisher.new(publisher: publisher)
@@ -101,6 +116,8 @@ module NistPubid
         .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("NIST CSWP", "NIST CSRC White Paper")
+        .gsub("NIST.CSWP.01162020pt", "NIST.CSWP.01162020(por)")
         .gsub(/(?<=NBS MP )(\d+)\((\d+)\)/, '\1e\2')
         .gsub(/(?<=\d)es/, "(spa)")
         .gsub(/(?<=\d)chi/, "(zho)")
@@ -123,11 +140,14 @@ module NistPubid
         revision: /[\da](?:r|Rev\.\s|([0-9]+[A-Za-z]*-[0-9]+[A-Za-z]*-))([\da]+)/
           .match(code)&.[](2),
         addendum: match(/(?<=(\.))?(add(?:-\d+)?|Addendum)/, code),
-        edition: /(?<=[^a-z])(?<=(\.))?(?:e(?(1)-)|Ed\.\s)(\d+)/
-          .match(code)&.[](2),
+        edition: /(?<=[^a-z])(?<=\.)?(?:e(?(1)-)|Ed\.\s)(\d+)|
+                  NBS\sFIPS\sPUB\s[0-9]+[A-Za-z]*-[0-9]+[A-Za-z]*-([A-Za-z\d]+)
+          /x.match(code)&.captures&.join,
         section: /(?<=sec)\d+/.match(code)&.to_s,
         appendix: /\d+app/.match(code)&.to_s,
-        errata: /-errata/.match(code)&.to_s
+        errata: /-errata|\d+err(?:ata)?/.match(code)&.to_s,
+        index: /\d+index|\d+indx/.match(code)&.to_s,
+        insert: /\d+ins(?:ert)?/.match(code)&.to_s
       }
       supplement = /(?:(?:supp?)-?(\d*)|Supplement|Suppl.)/
         .match(code)
@@ -163,7 +183,7 @@ module NistPubid
     end
 
     def self.parse_docnumber(serie, code)
-      localities = "[Pp]t\\d+|r(?:\\d+|[A-Za-z]?)|e\\d+|p|v|sec\\d+"
+      localities = "[Pp]t\\d+|r(?:\\d+|[A-Za-z]?)|e\\d+|p|v|sec\\d+|inde?x|err(?:ata)?|ins(?:ert)?"
       excluded_parts = "(?!#{localities}|supp?)"
 
       if ["NBS CSM", "NBS CS"].include?(serie)
@@ -179,7 +199,7 @@ module NistPubid
            ([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
+             (?:-[A-Za]+)? # for NIST SP 1075-NCNR, NIST SP 1113-BFRL, NIST IR 6529-a
              (?:-[0-9.]+)? # second part
              (?:
                (?: # only big letter
@@ -203,7 +223,7 @@ module NistPubid
       regex.match(code)&.to_s
     end
 
-    def to_s(format)
+    def to_s(format = :short)
       result = render_serie(format)
       result += " " unless format == :short || stage.nil?
       result += "#{stage&.to_s(format)}"\
@@ -217,6 +237,23 @@ module NistPubid
       result
     end
 
+    def to_json(*args)
+      result = {
+        styles: {
+          short: to_s(:short),
+          abbrev: to_s(:abbrev),
+          long: to_s(:long),
+          mr: to_s(:mr),
+        }
+      }
+
+      instance_variables.each do |var|
+        val = instance_variable_get(var)
+        result[var.to_s.gsub('@', '')] = val unless val.nil?
+      end
+      result.to_json(*args)
+    end
+
     def render_serie(format)
       if serie.to_s(format).include?(publisher.to_s(format))
         return serie.to_s(format)
@@ -234,27 +271,21 @@ module NistPubid
 
     def render_edition(format)
       result = ""
-      if revision
-        result += if %i[long abbrev].include?(format) ||
-            [volume, part, supplement, version, edition].any?
-                    "#{REVISION_DESC[format]}#{revision}"
-                  else
-                    "-#{revision}"
-                  end
-      end
+
+      result += "#{REVISION_DESC[format]}#{revision.to_s.upcase}" if revision
       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 += INDEX_DESC[format] unless index.nil?
+      result += INSERT_DESC[format] unless insert.nil?
 
       result
     end

data/lib/nist_pubid/nist_tech_pubs.rb

@@ -23,19 +23,18 @@ module NistPubid
 
       def convert(doc)
         id = @converted_id[doc[:id]] ||= NistPubid::Document.parse(doc[:id])
-        return id.to_s(:short) unless doc.key?(:doi)
+        return id unless doc.key?(:doi)
 
         begin
           doi = @converted_doi[doc[:doi]] ||=
             NistPubid::Document.parse(doc[:doi])
         rescue Errors::ParseError
-          return id.to_s(:short)
+          return id
         end
         # return more complete pubid
-        id.merge(doi).to_s(:short)
+        id.merge(doi)
       rescue Errors::ParseError
         @converted_doi[doc[:doi]] ||= NistPubid::Document.parse(doc[:doi])
-          .to_s(:short)
       end
 
       def parse_docid(doc)
@@ -57,7 +56,7 @@ module NistPubid
 
       def comply_with_pubid
         fetch.select do |doc|
-          convert(doc) == doc[:id]
+          convert(doc).to_s == doc[:id]
         rescue Errors::ParseError
           false
         end
@@ -65,7 +64,7 @@ module NistPubid
 
       def different_with_pubid
         fetch.reject do |doc|
-          convert(doc) == doc[:id]
+          convert(doc).to_s == doc[:id]
         rescue Errors::ParseError
           true
         end
@@ -73,7 +72,7 @@ module NistPubid
 
       def parse_fail_with_pubid
         fetch.select do |doc|
-          convert(doc) && false
+          convert(doc).to_s && false
         rescue Errors::ParseError
           true
         end
@@ -82,9 +81,22 @@ module NistPubid
       # returning current document id, doi, title and final PubID
       def status
         fetch.map do |doc|
-          [doc[:id], doc[:doi], doc[:title], convert(doc)]
+          final_doc = convert(doc)
+          {
+            id: doc[:id],
+            doi: doc[:doi],
+            title: doc[:title],
+            finalPubId: final_doc.to_s,
+            mr: final_doc.to_s(:mr),
+          }
         rescue Errors::ParseError
-          [doc[:id], doc[:doi], doc[:title], "parse error"]
+          {
+            id: doc[:id],
+            doi: doc[:doi],
+            title: doc[:title],
+            finalPubId: "parse error",
+            mr: "parse_error"
+          }
         end
       end
     end

data/lib/nist_pubid/publisher.rb

@@ -9,7 +9,7 @@ module NistPubid
       @publisher = publisher
     end
 
-    def to_s(format)
+    def to_s(format = :short)
       return @publisher if %i[short mr].include?(format)
 
       PUBLISHERS[format.to_s][@publisher]

data/lib/nist_pubid/serie.rb

@@ -8,7 +8,7 @@ module NistPubid
       @serie = serie == "NISTIR" ? "NIST IR" : serie
     end
 
-    def to_s(format)
+    def to_s(format = :short)
       return @serie if %i[short mr].include?(format)
 
       result = SERIES[format.to_s][@serie]

data/lib/nist_pubid/version.rb

@@ -1,3 +1,3 @@
 module NistPubid
-  VERSION = "0.1.5".freeze
+  VERSION = "0.1.6".freeze
 end

data/nist_pubid.gemspec

@@ -33,4 +33,5 @@ Gem::Specification.new do |spec|
 
   spec.add_dependency "relaton-nist"
   spec.add_dependency "nokogiri"
+  spec.add_dependency "thor"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: nist-pubid
 version: !ruby/object:Gem::Version
-  version: 0.1.5
+  version: 0.1.6
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2022-01-07 00:00:00.000000000 Z
+date: 2022-01-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
@@ -94,6 +94,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: thor
+  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