relaton-nist 1.14.6 → 1.14.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ff158dab9df4bfc2254789771db9ca74bc422b9d6ac78351da0ccbca1d858815
-  data.tar.gz: e7ef8d2cb1ff797095f7c6646194610ffe7bab67f97d909c9d54d75d9f0552c3
+  metadata.gz: bb068b75779f92292d1fef25e9ae392a76df53d5d52bc2a644fb9d16ec0c5998
+  data.tar.gz: f64ba4dd75be3fd93a5a134c4babe39f08d4b6f774e2d488ba8df525f469fa1c
 SHA512:
-  metadata.gz: ef6f0be37367d7e71ac0783be8252219a9a49e7affe27bfcef78c21929d64c10b2f4cba081a4515d75e64bfccffed21dd72848d2653594363f9e4fa753c7c002
-  data.tar.gz: f9b24bb62b115b11f291a9514420426a758d3207062e4bd75fc09169f32338d3891ca802441df405ce9b4ab159f39aea2762c458e811c30e87523e9835a94dbc
+  metadata.gz: b2f68dc9000e02d2b23a611d02d84d76222d24811313b88fa5ccef8fbda2564058a93dd6ef1b0f71a7b36e0f278cb1107046c47b5098d3d0c43a423f235c3c70
+  data.tar.gz: c38c9c1d2faf307843417b2be67d52a1064160b8ad3b2412c120b26cbe730ecef57ea1019d75eddf47d3e9d5155109571cf1ba69c2a3bce17d02960576ce9e3d

data/.github/workflows/release.yml

@@ -7,7 +7,8 @@ on:
     inputs:
       next_version:
         description: |
-          Next release version. Possible values: x.y.z, major, minor, patch or pre|rc|etc
+          Next release version. Possible values: x.y.z, major, minor, patch (or pre|rc|etc).
+          Also, you can pass 'skip' to skip 'git tag' and do 'gem push' for the current version
         required: true
         default: 'skip'
   repository_dispatch:

data/README.adoc

@@ -43,6 +43,14 @@ sources:
 . bibliographic feed from NIST CSRC
 . NIST Library dataset
 
+The NIST CSRC provides:
+
+* NIST SP 800-*
+* NIST SP 500-*
+* NIST SP 1800-*
+* NIST FIPS {31, 39, 41, 46, 46-$$*$$, 48, 65, 73, 74, 81, 83, 87, 102, 112, 113, 139, 140-$$*$$, 141, 171, 180, 180-$$*$$, 181, 186, 186-$$*$$, 188, 190, 191, 196, 197, 198, 198-1, 199, 200, 201, 201-*, 202}
+
+The NIST Library dataset provides documents listed in the https://github.com/relaton/relaton-data-nist/blob/main/index-v1.yaml[index].
 
 == Installation
 

data/lib/relaton_nist/data_fetcher.rb

@@ -36,9 +36,16 @@ module RelatonNist
       ]
     end
 
+    #
+    # Parse document's ID from XML
+    #
+    # @param [Nokogiri::XML::Element] doc XML element
+    #
+    # @return [String] document's ID
+    #
     def pub_id(doc)
       # anchor(doc).gsub(".", " ")
-      fetch_doi(doc).split("/")[1..].join("/").gsub(".", " ")
+      fetch_doi(doc).split("/")[1..].join("/").gsub(".", " ").sub(/^nist\sir/, "NIST IR")
     end
 
     def fetch_doi(doc) # rubocop:disable Metrics/CyclomaticComplexity
@@ -269,8 +276,6 @@ module RelatonNist
     # @return [Array<RelatonBib::Series>] series
     #
     def fetch_series(doc)
-      series_path = File.expand_path("series.yaml", __dir__)
-      series = YAML.load_file series_path
       prf, srs, num = pub_id(doc).split
       sname = series[srs] || srs
       title = RelatonBib::TypedTitleString.new(content: "#{prf} #{sname}")
@@ -278,6 +283,10 @@ module RelatonNist
       [RelatonBib::Series.new(title: title, abbreviation: abbr, number: num)]
     end
 
+    def series
+      @series ||= YAML.load_file File.expand_path("series.yaml", __dir__)
+    end
+
     #
     # Save document
     #

data/lib/relaton_nist/hit.rb

@@ -19,24 +19,24 @@ module RelatonNist
     #
     # @return [Iteger] sorting weigth
     #
-    def sort_value # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
-      @sort_value ||= begin
-        sort_phrase = [hit[:series], hit[:code], hit[:title]].join " "
-        corr = hit_collection&.text&.split&.map do |w|
-          if w =~ /\w+/ &&
-              sort_phrase =~ Regexp.new(Regexp.escape(w), Regexp::IGNORECASE)
-            1
-          else 0
-          end
-        end&.sum.to_i
-        corr + case hit[:status]
-               when "final" then 4
-               when "withdrawn" then 3
-               when "draft" then 2
-               when "draft (obsolete)" then 1
-               else 0
-               end
-      end
-    end
+    # def sort_value # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
+    #   @sort_value ||= begin
+    #     sort_phrase = [hit[:series], hit[:code]].join " "
+    #     corr = hit_collection&.text&.split&.map do |w|
+    #       if w =~ /\w+/ &&
+    #           sort_phrase =~ Regexp.new(Regexp.escape(w), Regexp::IGNORECASE)
+    #         1
+    #       else 0
+    #       end
+    #     end&.sum.to_i
+    #     corr + case hit[:status]
+    #            when "final" then 4
+    #            when "withdrawn" then 3
+    #            when "draft" then 2
+    #            when "draft (obsolete)" then 1
+    #            else 0
+    #            end
+    #   end
+    # end
   end
 end

data/lib/relaton_nist/hit_collection.rb

@@ -56,13 +56,13 @@ module RelatonNist
     def search_filter # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/PerceivedComplexity,Metrics/MethodLength
       @array.select do |item|
         parts = doi_parts(item.hit[:json]) || code_parts(item.hit[:code])
-        (refparts[:code] && [parts[:series], item.hit[:series]].include?(refparts[:series]) &&
+        refparts[:code] && [parts[:series], item.hit[:series]].include?(refparts[:series]) &&
           refparts[:code].casecmp(parts[:code].upcase).zero? &&
-          (refparts[:prt] == parts[:prt]) &&
-          (refparts[:vol] == parts[:vol]) &&
-          (refparts[:ver] == parts[:ver]) &&
-          (refparts[:rev] == parts[:rev]) &&
-          refparts[:draft] == parts[:draft] && refparts[:add] == parts[:add])
+          refparts[:prt] == parts[:prt] &&
+          (refparts[:vol].nil? || refparts[:vol] == parts[:vol]) &&
+          (refparts[:ver].nil? || refparts[:ver] == parts[:ver]) &&
+          (refparts[:rev].nil? || refparts[:rev] == parts[:rev]) &&
+          refparts[:draft] == parts[:draft] && refparts[:add] == parts[:add]
       end
     end
 
@@ -71,8 +71,8 @@ module RelatonNist
     def code_parts(code) # rubocop:disable Metrics/MethodLength
       {
         # prefix: match(/^(?:NIST|NBS)\s?/, code),
-        series: match(/(?<val>(?:SP|FIPS|IR|ITL\sBulletin|White\sPaper))\s/, code),
-        code: match(/(?<val>[0-9-]{3,}[A-Z]?)/, code),
+        series: match(/(?<val>(?:SP|FIPS|CSWP|IR|ITL\sBulletin|White\sPaper))\s/, code),
+        code: match(/(?<val>[0-9-]+(?:(?!(?:ver|r|v|pt)\d|-add\d?)[A-Za-z-])*)/, code),
         prt: match(/(?:pt|\sPart\s)(?<val>\d+)/, code),
         vol: match(/(?:v|\sVol\.\s)(?<val>\d+)/, code),
         ver: match(/(?:ver|\sVer\.\s|Version\s)(?<val>[\d.]+)/, code),
@@ -81,7 +81,7 @@ module RelatonNist
         # (?:\s(?<vol2>Vol\.\s\d+))?
         # (?:\s(?<ver2>(?:Ver\.|Version)\s[\d.]+))?
         # (?:\s(?<rev2>Rev\.\s\d+))?
-        add: match(/\sAdd(?:endum)?(?<val>\d*)/, code),
+        add: match(/(?:-add|\sAdd)(?:endum)?(?<val>\d*)/, code),
         draft: !match(/\((?:Retired\s)?Draft\)/, code).nil?,
       }
     end
@@ -92,8 +92,8 @@ module RelatonNist
       id = json["doi"].split("/").last
       {
         # prefix: match(/^(?:NIST|NBS)\./, id),
-        series: match(/(?:SP|FIPS|IR|ITL\sBulletin|White\sPaper)(?=\.)/, id),
-        code: match(/(?<=\.)\d{3,}(?:-\d+)*(?:[[:alpha:]](?!\d|raft|er|t?\d))?/, id),
+        series: match(/(?:SP|FIPS|CSWP|IR|ITL\sBulletin|White\sPaper)(?=\.)/, id),
+        code: match(/(?<=\.)\d+(?:-\d+)*(?:[[:alpha:]](?!\d|raft|er|t?\d))?/, id),
         prt: match(/pt?(?<val>\d+)/, id),
         vol: match(/v(?<val>\d+)(?!\.\d)/, id),
         ver: match(/v(?:er)?(?<val>[\d.]+)/, id),
@@ -110,9 +110,9 @@ module RelatonNist
     #
     def refparts # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
       @refparts ||= {
-        perfix: match(/^(NIST|NBS)\s?/, text),
-        series: match(/(SP|FIPS|IR|ITL\sBulletin|White\sPaper)(?=\.|\s)/, text),
-        code: match(/(?<=\.|\s)[0-9-]{3,}[A-Z]?/, text),
+        perfix: match(/^(NIST|NBS)/, text),
+        series: match(/(SP|FIPS|CSWP|IR|ITL\sBulletin|White\sPaper)(?=\.|\s)/, text),
+        code: match(/(?<=\.|\s)[0-9-]+(?:(?!(ver|r|v|pt)\d|-add\d?)[A-Za-z-])*/, text),
         prt: match(/(?:(?<dl>\.)?pt(?(<dl>)-)|\sPart\s)(?<val>[A-Z\d]+)/, text),
         vol: match(/(?:(?<dl>\.)?v(?(<dl>)-)|\sVol\.\s)(?<val>\d+)/, text),
         ver: match(/(?:(?<dl>\.)?\s?ver|\sVer\.\s)(?<val>\d(?(<dl>)[-\d]|[.\d])*)/, text)&.gsub(/-/, "."),
@@ -149,10 +149,11 @@ module RelatonNist
     #
     def full_ref # rubocop:disable Metrics/AbcSize
       @full_ref ||= begin
-        ref = "#{refparts[:perfix]}#{refparts[:series]} #{refparts[:code]}"
+        ref = [refparts[:perfix], refparts[:series], refparts[:code]].compact.join " "
         ref += "pt#{refparts[:prt]}" if refparts[:prt] # long_to_short(refparts, "prt").to_s
         ref += "ver#{refparts[:ver]}" if refparts[:ver] # long_to_short(refparts, "vol").to_s
         ref += "v#{refparts[:vol]}" if refparts[:vol]
+        ref += "r#{refparts[:rev]}" if refparts[:rev]
         ref
       end
     end
@@ -164,11 +165,10 @@ module RelatonNist
     #
     def sort_hits!
       @array.sort! do |a, b|
-        if a.sort_value == b.sort_value
-          (b.hit[:release_date] - a.hit[:release_date]).to_i
-        else
-          b.sort_value - a.sort_value
-        end
+        code = a.hit[:code] <=> b.hit[:code]
+        next code unless code.zero?
+
+        b.hit[:release_date] <=> a.hit[:release_date]
       end
       self
     end
@@ -184,16 +184,20 @@ module RelatonNist
       ref = full_ref
       # fn = ref.gsub(%r{[/\s:.]}, "_").upcase
       index = Relaton::Index.find_or_create :nist, url: "#{GHNISTDATA}index-v1.zip", file: INDEX_FILE
-      row = index.search { |r| r[:id] == ref }.min_by { |r| r[:id] }
-      return [] unless row
+      rows = index.search(ref).sort_by { |r| r[:id] }
+      # return [] unless row
 
-      yaml = OpenURI.open_uri "#{GHNISTDATA}#{row[:file]}"
-      hash = YAML.safe_load yaml
-      hash["fetched"] = Date.today.to_s
-      bib = RelatonNist::NistBibliographicItem.from_hash hash
-      hit = Hit.new({ code: text }, self)
-      hit.fetch = bib
-      [hit]
+      # yaml = OpenURI.open_uri "#{GHNISTDATA}#{row[:file]}"
+      # hash = YAML.safe_load yaml
+      # hash["fetched"] = Date.today.to_s
+      # bib = RelatonNist::NistBibliographicItem.from_hash hash
+      # id = bib.docidentifier.find(&:primary).id
+
+      rows.map do |row|
+        Hit.new({ code: row[:id], path: row[:file] }, self)
+      end
+      # hit.fetch = bib
+      # [hit]
     rescue OpenURI::HTTPError => e
       return [] if e.io.status[0] == "404"
 
@@ -263,8 +267,9 @@ module RelatonNist
     def match_year?(doc, date)
       return true unless year
 
-      idate = RelatonBib.parse_date doc["issued-date"], false
-      idate.between? date, date.next_year.prev_day
+      d = doc["issued-date"] || doc["published-date"]
+      pidate = RelatonBib.parse_date d, false
+      pidate.between? date, date.next_year.prev_day
     end
   end
 end

data/lib/relaton_nist/nist_bibliography.rb

@@ -12,7 +12,7 @@ module RelatonNist
   class NistBibliography
     class << self
       #
-      # Search NIST docuemnts by reference
+      # Search NIST documents by reference
       #
       # @param text [String] reference
       #
@@ -47,9 +47,9 @@ module RelatonNist
           if date2
             case date2
             when /\w+\s\d{4}/
-              opts[:issued_date] = Date.strptime date2, "%B %Y"
+              opts[:date] = Date.strptime date2, "%B %Y"
             when /\w+\s\d{2},\s\d{4}/
-              opts[:updated_date] = Date.strptime date2, "%B %d, %Y"
+              opts[:date] = Date.strptime date2, "%B %d, %Y"
             end
           end
           opts[:stage] = stage if stage
@@ -117,16 +117,9 @@ module RelatonNist
                     end
         result.each_slice(3) do |s| # ISO website only allows 3 connections
           fetch_pages(s, 3).each_with_index do |r, _i|
-            if opts[:issued_date]
-              ids = r.date.select do |d|
-                d.type == "issued" && d.on(:date) == opts[:issued_date]
-              end
-              next if ids.empty?
-            elsif opts[:updated_date]
-              pds = r.date.select do |d|
-                d.type == "published" && d.on(:date) == opts[:updated_date]
-              end
-              next if pds.empty?
+            if opts[:date]
+              dates = r.date.select { |d| d.on(:date) == opts[:date] }
+              next if dates.empty?
             end
             next if iter && r.status.iteration != iteration
             return { ret: r } if !year

data/lib/relaton_nist/scrapper.rb

@@ -11,6 +11,17 @@ module RelatonNist
       # @param hit_data [Hash]
       # @return [Hash]
       def parse_page(hit_data)
+        hit_data[:url] ? parse_json(hit_data) : fetch_gh(hit_data)
+      end
+
+      def fetch_gh(hit_data)
+        yaml = OpenURI.open_uri "#{HitCollection::GHNISTDATA}#{hit_data[:path]}"
+        hash = YAML.safe_load yaml
+        hash["fetched"] = Date.today.to_s
+        NistBibliographicItem.from_hash hash
+      end
+
+      def parse_json(hit_data)
         item_data = from_json hit_data
         titles = fetch_titles(hit_data)
         unless /^(SP|NISTIR|FIPS) /.match? item_data[:docid][0].id

data/lib/relaton_nist/series.yaml

@@ -22,7 +22,7 @@ MONO: Monographs
 MP: Miscellaneous Publications
 NCSTAR: National Construction Safety Team Act Reports
 NSRDS: National Standard Reference Data Series
-IR: NISTIRs (Interagency/Internal Reports)
+IR: IR (Interagency/Internal Reports)
 OWMWP: Office of Weights and Measures White Papers
 PC: Photographic Circulars
 RPT: NBS Reports

data/lib/relaton_nist/version.rb

@@ -1,3 +1,3 @@
 module RelatonNist
-  VERSION = "1.14.6".freeze
+  VERSION = "1.14.8".freeze
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-nist
 version: !ruby/object:Gem::Version
-  version: 1.14.6
+  version: 1.14.8
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2023-06-06 00:00:00.000000000 Z
+date: 2023-08-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: relaton-bib