pubid 2.0.0.pre.alpha.7 → 2.0.0.pre.alpha.9

data/lib/pubid/nist/identifiers/base.rb

@@ -3,9 +3,7 @@
 module Pubid
   module Nist
     # NIST/NBS flavor base class. Canonical name Pubid::Nist::Identifier; every
-    # concrete NIST identifier (Identifiers::*) descends from it, and
-    # Identifiers::Base — aliased at the foot of this file — points back to it
-    # for backward compatibility. Being a real Pubid::Identifier subclass gives
+    # concrete NIST identifier (Identifiers::*) descends from it. Being a real Pubid::Identifier subclass gives
     # native `is_a?` identity and the shared polymorphic `from_hash`.
     class Identifier < ::Pubid::Identifier
       # Delegate to the flavor module so callers can use
@@ -14,6 +12,77 @@ module Pubid
         Pubid::Nist.parse(identifier)
       end
 
+      # Compact, index-friendly serialization (mirrors the key_value flavors
+      # ISO/ITU/ETSI, without a 38-field key_value block). Two transforms,
+      # applied symmetrically by to_hash/from_hash so the round-trip stays
+      # idempotent:
+      #   - flatten the single-value Code components (series, number) to bare
+      #     scalars — `value` is the only stored field. NIST's
+      #     `Components::Code#part` is a *computed* reader
+      #     (`value.split("-").last`), and prefix/subpart/parts are never
+      #     assigned on series/number, so the scalar `value` is lossless:
+      #     expand recomputes exactly what the object would compute anyway;
+      #   - flatten the plain single-value components (volume) to bare scalars
+      #     too — `Components::Volume` holds only `value`, so it expands to a
+      #     bare { "value" => … } with no part;
+      #   - drop the redundant build artifacts, decomposed pieces of the
+      #     canonical number that identity/rendering never read.
+      COMPACT_FLAT_CODES = %w[series number subseries].freeze
+      COMPACT_FLAT_VALUES = %w[volume].freeze
+      COMPACT_DROP_ATTRS = %w[first_number second_number].freeze
+
+      # Deserialize a (possibly compact) hash: expand the flattened scalars
+      # back into Code sub-hashes, then let the shared polymorphic from_hash
+      # do the real work. Nested `base` documents are expanded recursively by
+      # their own from_hash via apply_mappings, so only this level is touched.
+      def self.from_hash(data, options = {})
+        super(expand_compact_hash(data), options)
+      end
+
+      # Expand a compact hash in place-safe fashion: a scalar series/number
+      # becomes { "value" => …, "part" => <last-dash segment> }.
+      def self.expand_compact_hash(data)
+        return data unless data.is_a?(::Hash)
+
+        out = data.dup
+        COMPACT_FLAT_CODES.each do |key|
+          out[key] = expand_code(out[key]) if out[key].is_a?(::String)
+        end
+        COMPACT_FLAT_VALUES.each do |key|
+          out[key] = { "value" => out[key] } if out[key].is_a?(::String)
+        end
+        out
+      end
+
+      # A flattened Code scalar -> { "value" => …, "part" => <last-dash seg> }.
+      def self.expand_code(value)
+        code = { "value" => value }
+        code["part"] = value.split("-").last if value.include?("-")
+        code
+      end
+
+      # Serialize, then apply the compact transforms (recursing into a nested
+      # `base`, which lutaml serialized via its own transform — bypassing this
+      # override — so it is not compacted otherwise).
+      def to_hash(*args)
+        hash = super
+        compact_hash!(hash) if hash.is_a?(::Hash)
+        hash
+      end
+
+      def compact_hash!(hash)
+        COMPACT_DROP_ATTRS.each { |key| hash.delete(key) }
+        (COMPACT_FLAT_CODES + COMPACT_FLAT_VALUES).each do |key|
+          value = hash[key]
+          next unless value.is_a?(::Hash) && value.key?("value")
+
+          hash[key] = value["value"]
+        end
+        compact_hash!(hash["base"]) if hash["base"].is_a?(::Hash)
+        hash
+      end
+      private :compact_hash!
+
       # Default: no typed stages. Subclasses override as needed.
       def self.typed_stages
         []
@@ -29,11 +98,23 @@ module Pubid
       # raw string the circular/supplement builders pass straight through.
       attribute :publisher, :string
       attribute :series, Components::Code # Set by Builder from parsed data
+      # SP subseries (Appendix A.2 of the NIST PubID Syntax): for "NIST SP 800-53"
+      # the subseries is "800" and the sequence is "53". Populated by the Builder
+      # when the series is "SP" and the first number matches a known subseries
+      # code. Informational metadata derived from series+number — ignored in
+      # equality so a manually-built id (without subseries set) still equals a
+      # parsed one.
+      attribute :subseries, Components::Code
       attribute :number, Components::Code
 
       # V2 COMPONENTS (Lutaml::Model objects) - PROPER SEPARATION
+      # `edition` (Components::Edition: type + id + additional_text) is the
+      # single source of truth for edition/revision. The former
+      # `edition_component` (a byte-identical duplicate) and the `revision`
+      # string (always "#{edition.type}#{edition.id}") were unfinished V1->V2
+      # migration aliases; both removed. `#revision` survives as a derived
+      # reader below.
       attribute :edition, Components::Edition # Edition (type + id): e2, e2021, r5, -3
-      attribute :edition_component, Components::Edition # V2 edition component (alias)
       attribute :volume, Components::Volume # Volume component (v6)
       attribute :part, Components::Part  # Part component (n1 or pt1)
       attribute :stage, Components::Stage
@@ -52,7 +133,6 @@ module Pubid
 
       # LEGACY attributes (keep for backward compatibility during migration)
       attribute :parts, Components::Code, collection: true
-      attribute :revision, :string
       attribute :revision_year, :string # Year for revision (e.g., r6/1925, r1963, rJun1992)
       attribute :revision_month, :string # Month for revision (e.g., rJun1992)
       attribute :edition_year, :string # Legacy edition year for backward compatibility
@@ -108,12 +188,15 @@ module Pubid
       # Attributes that are build artifacts or rendering aliases, not part
       # of an identifier's logical identity. They diverge between equally-
       # valid spellings of the same id (e.g. long "Rev. 1" vs short "r1"):
-      #   - edition_component: redundant alias of :edition
       #   - first_number/second_number: decomposed parts of the canonical
       #     :number, retained from the parse for building
       #   - parsed_format: records the input format for round-trip rendering
+      #   - subseries: derived from series+number (which SP subseries the
+      #     document belongs to); populated by the Builder for known SP
+      #     subseries codes. Ignored in equality so a manually-built id
+      #     without subseries set still equals a parsed one.
       EQUALITY_IGNORED_ATTRS = %i[
-        edition_component first_number second_number parsed_format
+        first_number second_number parsed_format subseries
       ].freeze
 
       # Logical identity comparison: equal when every attribute except the
@@ -161,21 +244,11 @@ module Pubid
         end
       end
 
-      # Return a copy with the named attributes nil'd. Overrides
-      # Pubid::Identifier#exclude because NIST's initialize is keyword-only
-      # (initialize(**attributes)) while the inherited exclude rebuilds via
-      # the positional self.class.new(attrs) form — passing a positional
-      # hash to a keyword-only initializer raises ArgumentError. Rebuild
-      # with the keyword splat instead.
-      def exclude(*args)
-        excluded_args = args.dup
-        excluded_args << :date if excluded_args.delete(:year)
-
-        attrs = self.class.attributes.each_with_object({}) do |(name, _), h|
-          h[name] = excluded_args.include?(name) ? nil : public_send(name)
-        end
-        self.class.new(**attrs)
-      end
+      # NIST inherits Pubid::Identifier#exclude directly: it now rebuilds via
+      # `self.class.new(**attrs)`, so NIST's keyword-only initialize
+      # (initialize(**attributes)) works without a local override, and the base
+      # version additionally carries the nested-identifier recursion this class's
+      # old copy lacked.
 
       # Short-form supplement fragment ("sup", "sup1924", "supJan1924",
       # "suprev", " supJun1925-Jun1926"), rendered from the structured
@@ -190,15 +263,12 @@ module Pubid
         prefix + (rendered.empty? ? "sup" : rendered)
       end
 
-      # Compute revision from edition component for backward compatibility
+      # Derived, read-only revision string ("r5", "e1979", "-3") for backward
+      # compatibility. `edition` is the single stored representation; this is
+      # never persisted, compared, or merged — just a convenience view of it.
       # @return [String, nil] revision string (e.g., "r5") or nil
       def revision
-        return @revision if @revision
-
-        # Compute from edition component if available
-        if edition&.type && edition.id
-          "#{edition.type}#{edition.id}"
-        end
+        "#{edition.type}#{edition.id}" if edition&.type && edition.id
       end
 
       # Backward compatibility: translation method returns translation_component
@@ -243,11 +313,28 @@ module Pubid
                nist_format: format || parsed_format&.to_sym || :short)
       end
 
+      # NIST has its own defined MR format (to_mr_style) used by relaton-nist
+      # and downstream tooling. The generic Renderers::MrString does not know
+      # about NIST series, revisions, supplements, etc., so delegate to the
+      # established style for losslessness (issue #142).
+      def to_mr_string
+        to_mr_style
+      end
+
+      # NIST's MR shape is fixed by the pubid standard and stays uppercase
+      # (e.g. `NIST.SP.800-53`). The slug, however, must be lowercase to
+      # match the cross-flavor convention; project the MR through `downcase`.
+      def to_slug
+        to_mr_string.downcase
+      end
+
       # Returns weight based on amount of defined attributes
       # Used for ranking identifiers by specificity for conflict resolution
       # @return [Integer] weight score (higher = more specific)
       def weight
         self.class.attributes.keys.inject(0) do |sum, key|
+          next sum if EQUALITY_IGNORED_ATTRS.include?(key)
+
           val = public_send(key)
           val && !val.to_s.empty? ? sum + 1 : sum
         end
@@ -275,7 +362,7 @@ module Pubid
                          when :edition
                            current_val.nil? || edition_greater?(new_val,
                                                                 current_val)
-                         when :volume, :part, :version, :revision
+                         when :volume, :part, :version
                            current_val.nil? || (new_val.to_s.length > current_val.to_s.length)
                          when :supplement, :errata, :index, :insert, :section, :appendix, :translation
                            true
@@ -411,6 +498,10 @@ module Pubid
         result += " #{number}" if number
         result += parts.map { |p| "-#{p}" }.join if parts&.any?
 
+        # "NIST Research Library (YYYY)" — oddball identifier with no number;
+        # the year appears in parens rather than as a date component.
+        result += " (#{year})" if year && !number
+
         # Append every optional component the parser may have attached
         # (volume, part, edition, version, supplement, update, stage, ...).
         # Shared with the lossy subclasses so they cannot silently drop a
@@ -639,12 +730,5 @@ module Pubid
       end
     end
 
-    module Identifiers
-      # Backward-compatible alias: NIST's base class used to be
-      # Pubid::Nist::Identifiers::Base. It is now Pubid::Nist::Identifier; keep
-      # the old constant working for `class Foo < Base` declarations and any
-      # external reference.
-      Base = Pubid::Nist::Identifier
-    end
   end
 end

data/lib/pubid/nist/identifiers/circular.rb

@@ -7,7 +7,7 @@ module Pubid
       # Examples:
       # - "NBS CIRC 13e2revJune1908" = Circular 13, edition 2, revised June 1908
       # - "NBS CIRC 13" = Circular 13
-      class Circular < Base
+      class Circular < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["CIRC", "NBS CIRC"],

data/lib/pubid/nist/identifiers/commercial_standard.rb

@@ -9,7 +9,7 @@ module Pubid
       # - "NBS CS 102" - Basic commercial standard
       # - "NBS CS 102E-42" - With letter suffix and number
       # - "CS 190-58" - With year
-      class CommercialStandard < Base
+      class CommercialStandard < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["CS", "NBS CS"],

data/lib/pubid/nist/identifiers/commercial_standard_emergency.rb

@@ -6,7 +6,7 @@ module Pubid
       # NBS Commercial Standard (Emergency) Identifier
       # Format: NBS CS eNNN-YY where NNN is edition number, YY is 2-digit year
       # Example: "NBS CS e104-43" = Commercial Standard Emergency edition 104, year 1943
-      class CommercialStandardEmergency < Base
+      class CommercialStandardEmergency < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["CS-E", "NBS CS-E"],

data/lib/pubid/nist/identifiers/commercial_standards_monthly.rb

@@ -6,7 +6,7 @@ module Pubid
       # NBS Commercial Standards Monthly Identifier
       # Format: NBS CSM N where N is simple number
       # Example: "NBS CSM 1", "NBS CSM 40"
-      class CommercialStandardsMonthly < Base
+      class CommercialStandardsMonthly < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["CSM", "NBS CSM"],

data/lib/pubid/nist/identifiers/crpl_report.rb

@@ -8,7 +8,7 @@ module Pubid
       # - "NBS CRPL 1-2_3-1" = Reports 1-2 and 3-1 jointly
       # - "NBS CRPL 4-m-5" = Report 4, month m, issue 5
       # - "NBS CRPL c4-4" = Report c4, issue 4
-      class CrplReport < Base
+      class CrplReport < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["CRPL", "NBS CRPL", "CRPL-F-B", "CRPL-F-A", "NBS CRPL-F-B",

data/lib/pubid/nist/identifiers/dated_document.rb

@@ -7,7 +7,7 @@ module Pubid
       # Examples:
       # - "NIST 2022-04-15 001" (short) / "NIST.2022-04-15.001" (mr/doi)
       #   DOI 10.6028/NIST.2022-04-15.001
-      class DatedDocument < Base
+      class DatedDocument < Identifier
         # Empty so the bare "NIST" abbr never shadows a series lookup; the
         # Router guard (parsed_hash[:dated_date] && :dated_seq) is the dispatch path.
         TYPED_STAGES = [].freeze

data/lib/pubid/nist/identifiers/federal_information_processing_standards.rb

@@ -7,7 +7,7 @@ module Pubid
       # Examples:
       # - "FIPS 140-3" (no NIST prefix)
       # - "NIST FIPS 140-3" (also accepted, normalizes to FIPS 140-3)
-      class FederalInformationProcessingStandards < Base
+      class FederalInformationProcessingStandards < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["FIPS", "NIST FIPS"],

data/lib/pubid/nist/identifiers/grant_contractor_report.rb

@@ -7,7 +7,7 @@ module Pubid
       # Examples:
       # - "NIST GCR 17-917-45" - Basic 3-part number
       # - "NIST GCR 21-917-48v3B" - With volume and letter suffix
-      class GrantContractorReport < Base
+      class GrantContractorReport < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["GCR", "NIST GCR"],

data/lib/pubid/nist/identifiers/handbook.rb

@@ -7,7 +7,7 @@ module Pubid
       # Examples:
       # - "NIST HB 130" = Handbook 130
       # - "NBS HB 133" = NBS Handbook 133
-      class Handbook < Base
+      class Handbook < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["HB", "NIST HB", "NBS HB"],

data/lib/pubid/nist/identifiers/internal_report.rb

@@ -8,7 +8,7 @@ module Pubid
       # - "NIST IR 8115" = Interagency Report 8115
       # - "NIST IR 8115r1" = Interagency Report 8115 revision 1
       # - "NBS IR 73-197" = NBS Interagency Report 73-197
-      class InteragencyReport < Base
+      class InteragencyReport < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["IR", "NIST IR", "NBS IR"],

data/lib/pubid/nist/identifiers/letter_circular.rb

@@ -11,7 +11,7 @@ module Pubid
       # - "NBS LC 378sup12/1926" - With supplement date
       # - "NBS LC 378r11/1925" - With revision date
       # - "NBS LC 378(sp)" - With language code
-      class LetterCircular < Base
+      class LetterCircular < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["LCIRC"], # Single definition of truth

data/lib/pubid/nist/identifiers/miscellaneous_publication.rb

@@ -8,7 +8,7 @@ module Pubid
       # - "NBS MP 39e1" - Edition with "e" notation
       # - "NBS MP 260e1965" - Edition with year as edition ID
       # NOTE: Parenthetical edition format (e.g., "39(1)") does NOT exist for MP identifiers
-      class MiscellaneousPublication < Base
+      class MiscellaneousPublication < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["MP", "NBS MP"],

data/lib/pubid/nist/identifiers/monograph.rb

@@ -10,7 +10,7 @@ module Pubid
       # - "NBS MONO 128pt1" - Monograph with part
       # - "NIST MONO 1-1F" - Monograph with letter suffix
       # - "NIST MONO 1-2Bv1" - Monograph with letter suffix and volume
-      class Monograph < Base
+      class Monograph < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["MONO", "NBS MONO", "NIST MONO"],

data/lib/pubid/nist/identifiers/ncstar.rb

@@ -8,7 +8,7 @@ module Pubid
       # - "NIST NCSTAR 1-1Cv1" → "NIST NCSTAR 1-1Cv1" (round-trips)
       # - "NIST NCSTAR 1-1b" → "NIST NCSTAR 1-1B" (normalizes letter)
       # - "NIST NCSTAR 1-1cv1" → "NIST NCSTAR 1-1Cv1" (normalizes letter, keeps v1 format)
-      class Ncstar < Base
+      class Ncstar < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["NCSTAR", "NIST NCSTAR"],
@@ -34,7 +34,7 @@ module Pubid
 
         # NCSTAR-specific: do NOT expand Cv patterns to ", Volume" format
         # The base class would render "1-1Cv1" as is, which is correct for NCSTAR
-        # No custom rendering needed - just inherit from Base
+        # No custom rendering needed - just inherit from Identifier
       end
     end
   end

data/lib/pubid/nist/identifiers/nsrds.rb

@@ -8,7 +8,7 @@ module Pubid
       # - "NBS NSRDS 1" - Basic NSRDS
       # - "NSRDS-NBS 1" - With hyphen prefix
       # - "NBS NSRDS 1p1" - With part notation
-      class Nsrds < Base
+      class Nsrds < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["NSRDS", "NBS NSRDS"],

data/lib/pubid/nist/identifiers/owmwp.rb

@@ -6,7 +6,7 @@ module Pubid
       # NIST OWMWP (Office of Weights and Measures Workshop Proceedings)
       # Examples:
       # - "NIST OWMWP 06-13-2018" - Date-based format MM-DD-YYYY
-      class Owmwp < Base
+      class Owmwp < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["OWMWP", "NIST OWMWP"],

data/lib/pubid/nist/identifiers/report.rb

@@ -11,7 +11,7 @@ module Pubid
       # - "NBS RPT ADHOC" - Ad hoc report
       # - "NBS RPT div9" - Division report
       # - "NBS RPT 4817-A" - Report with letter suffix
-      class Report < Base
+      class Report < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["RPT", "NBS RPT"],

data/lib/pubid/nist/identifiers/special_publication.rb

@@ -8,7 +8,16 @@ module Pubid
       # - "NIST SP 800-53" = Special Publication 800-53
       # - "NIST SP 800-53r5" = Special Publication 800-53 revision 5
       # - "NIST SP 800-57pt1r4" = Special Publication 800-57 part 1 revision 4
-      class SpecialPublication < Base
+      class SpecialPublication < Identifier
+        # Per Appendix A.2 of the NIST PubID Syntax (April 2022), the SP series
+        # is subdivided into named subseries. The subseries number is the
+        # leading token of a report-num like "800-53" or "1190GB-12". The
+        # Builder exposes this as `identifier.subseries` so callers can query
+        # which subseries a document belongs to.
+        SP_SUBSERIES = %w[
+          250 260 300 400 480 500 700 800 823 960 1190GB 1200 1500 1800 1900 2000 2100
+        ].freeze
+
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["SP", "NIST SP", "NBS SP"], # Compound series

data/lib/pubid/nist/identifiers/technical_note.rb

@@ -7,7 +7,7 @@ module Pubid
       # Examples:
       # - "NIST TN 1234" = Technical Note 1234
       # - "NBS TN 567" = NBS Technical Note 567
-      class TechnicalNote < Base
+      class TechnicalNote < Identifier
         TYPED_STAGES = [
           Pubid::Components::TypedStage.new(
             abbr: ["TN", "NIST TN", "NBS TN"],

data/lib/pubid/nist/identifiers.rb

@@ -3,7 +3,6 @@
 module Pubid
   module Nist
     module Identifiers
-      autoload :Base, "#{__dir__}/identifiers/base"
       autoload :Circular, "#{__dir__}/identifiers/circular"
       autoload :CircularSupplement, "#{__dir__}/identifiers/circular_supplement"
       autoload :CommercialStandard, "#{__dir__}/identifiers/commercial_standard"

data/lib/pubid/nist/parser.rb

@@ -116,7 +116,9 @@ module Pubid
           str("AI") | str("CIRC") | str("CS") | str("CSM") |
           str("CRPL") | str("LCIRC") | str("OWMWP") | str("PC") | str("RPT") |
           str("SIBS") | str("TIBM") | str("TTB") | str("EAB") |
-          str("JPCRD") | str("JRES")
+          str("JPCRD") | str("JRES") |
+          # NEW - CHIPS Act, NWIRP, and Research Brief series
+          str("CHIPS") | str("NWIRP") | str("RB")
         ).as(:series)
       end
 
@@ -344,6 +346,16 @@ module Pubid
           match("[0-9]").repeat(1).as(:dated_seq)
       end
 
+      # "NIST Research Library (YYYY)" — one-off NIST publication that doesn't
+      # fit the standard series+number shape (no report number, year in parens).
+      # See pubid/pubid#151.
+      rule(:research_library_identifier) do
+        publisher >> space >>
+          str("Research Library").as(:series) >>
+          (space >> str("(") >>
+            match("[0-9]").repeat(4, 4).as(:year) >> str(")")).maybe
+      end
+
       # LEGACY EDITION PATTERNS (for backward compatibility during migration)
       # These will be gradually replaced as we migrate to proper Edition/Date components
       rule(:legacy_edition) do
@@ -436,7 +448,12 @@ module Pubid
 
       # Volume
       rule(:volume) do
-        (space.maybe >> (str("v") | str(" Vol. "))) >>
+        # Long-form "Vol." comes FIRST so PEG ordered choice doesn't commit to
+        # the bare "v" alternative after consuming the leading space via
+        # space.maybe (which would then starve the " Vol. " form). Both
+        # "Vol. 4" (period + space) and "Vol.4" (period only) are accepted.
+        (space.maybe >>
+          (str("Vol. ") | str("Vol.") | str("v"))) >>
           (digits >>
            # Support letter ranges (lowercase normalized in preprocessing)
            (str("a-l") | str("m-z") | str("A-L") | str("M-Z")).maybe >>
@@ -446,7 +463,8 @@ module Pubid
 
       # Part - enhanced to support patterns like p1adde1 AND pt3r1 (part with revision)
       rule(:part) do
-        ((space.maybe >> (str("pt") | str("p") | str("P"))) | str(" Part ")) >>
+        (str(" Part. ") | str(" Part.") | str(" Part ") |
+         (space.maybe >> (str("pt") | str("p") | str("P")))) >>
           (digits >>
            # NEW: Revision after part number: pt3r1, p1r1 (space.maybe for preprocessing)
            (space.maybe >> str("r") >> digits).maybe >>
@@ -472,7 +490,8 @@ module Pubid
           # Enhanced to accept letter-only revisions and space before r
           # ENHANCED: Accept BOTH lowercase and uppercase letters in suffix
           # ENHANCED: Capture original format prefix for format preservation (e.g., " Rev. 5")
-          ((str(" rev ") | str("rev") | str(" r") | str("r") | str(" Rev. ") | str(" Revision (r)")).as(:revision_prefix) >>
+          ((str(" rev ") | str("rev") | str(" r") | str("r") | str(" Rev. ") | str(" Rev.") | str(" Revision (r)")).as(:revision_prefix) >>
+           space.maybe >>
             ((digits >> match("[a-zA-Z]").maybe) | match("[a-zA-Z]").repeat(1)).as(:revision_id)).as(:revision) |
           # NEW: Standalone 'r' - MUST BE LAST to avoid consuming from other patterns
           # Matches " r" at end of input (after preprocessing: "800-56a r", "800-27 r")
@@ -654,6 +673,7 @@ module Pubid
         circ_supplement_identifier |
           dated_identifier |
           mr_identifier |
+          research_library_identifier |
           (
             # Compound series (includes publisher in series name)
             compound_series >> (space | dot) >>
@@ -688,6 +708,11 @@ module Pubid
             # Date range supplement (no base number)
             (str("supp") >> month_abbrev.as(:supp_month_start) >> digits.as(:supp_year_start) >>
              dash >> month_abbrev.as(:supp_month_end) >> digits.as(:supp_year_end)).as(:supplement_date_range) |
+            # Year-only date range supplement to the whole series, no months
+            # (e.g. "NBS CIRC sup1925-1926" — pubid/pubid#152). Requires two
+            # 4-digit years so it can't swallow "sup3/1926" or a base number.
+            ((str("supp") | str("sup")) >> match("[0-9]").repeat(4, 4).as(:supp_year_start) >>
+             dash >> match("[0-9]").repeat(4, 4).as(:supp_year_end)).as(:supplement_date_range) |
             # With base identifier + supplement
             (
               # Capture base portion (everything before "supp" or "sup" or slash+year)

data/lib/pubid/nist/preprocessor.rb

@@ -343,6 +343,21 @@ module Pubid
         revert_handbook_edition!
         revert_owmwp_date!
         revert_report_year_range!
+        revert_circ_supplement_year_range!
+      end
+
+      # CIRC/LCIRC year-only series supplements: "NBS CIRC sup1925-1926" is a
+      # supplement spanning a year range, so the trailing "-1926" is a range
+      # end, not an edition. convert_dashyear_to_edition! would have rewritten
+      # it to "sup1925e1926"; restore the dash so the circ_supplement_identifier
+      # grammar can read the range (pubid/pubid#152).
+      def revert_circ_supplement_year_range!
+        return unless @cleaned.include?("CIRC")
+
+        # Anchor on the series prefix ("CIRC " / "CIRC." / "LCIRC ") so only the
+        # whole-series form (sup right after the series, no base number) is
+        # reverted — a based "24sup1940e1950" keeps its genuine edition.
+        @cleaned = @cleaned.gsub(/(CIRC[ .])(supp?)(\d{4})e(\d{4})(?=\s|$)/, '\1\2\3-\4')
       end
 
       # HB handbooks: "HB 130e1979" → "HB 130-1979" (year is part of

data/lib/pubid/nist/router.rb

@@ -18,6 +18,7 @@ module Pubid
         "GCR", "AMS", "BSS", "BMS", "BH", "MONO", "MP",
         "NCSTAR", "NSRDS", "CSWP", "VTS", "AI", "OWMWP",
         "PC", "RPT", "SIBS", "TIBM", "TTB", "EAB",
+        "CHIPS", "NWIRP", "RB",
         "JPCRD", "JRES", "CIS", "HR", "IRPL", "IP",
         "LC", "PS", "LCIRC",
         # Compound series
@@ -126,10 +127,10 @@ module Pubid
         # instead of an injected dependency.
         typed_stage = Pubid::Nist.locate_stage(series)
         if typed_stage
-          Pubid::Nist.locate_type(typed_stage.type_code) || Identifiers::Base
+          Pubid::Nist.locate_type(typed_stage.type_code) || Identifier
         else
           # Fallback to Base for unmapped series (e.g., "AMS", "VTS")
-          Identifiers::Base
+          Identifier
         end
       end
 

data/lib/pubid/nist/series/ir.rb

@@ -36,8 +36,6 @@ module Pubid
           )
           edition_obj = Components::Edition.new(type: "r", id: "1")
           identifier.edition = edition_obj
-          identifier.edition_component = edition_obj
-          identifier.revision = "r1"
           true
         end
 
@@ -52,7 +50,6 @@ module Pubid
             value: "#{match[1]}-#{match[2]}",
           )
           identifier.edition = nil
-          identifier.edition_component = nil
         end
       end
     end