steuer 1.0.4 → 1.2.0

data/lib/steuer/finanzamt_registry.rb

@@ -0,0 +1,50 @@
+# frozen_string_literal: true
+
+require 'yaml'
+
+module Steuer
+  # Lookup of Bundesfinanzamtsnummer (BUFA-Nr, the four-digit prefix of the
+  # federal-13 form) to the office's name, backed by the BZSt BUFA-Tabelle
+  # bundled in data/finanzaemter.yml.
+  #
+  # Unlike the structural data in StateMapping, these names go stale: offices
+  # merge and get renamed several times a year. Lookups therefore return nil
+  # rather than raising, so a number issued after the bundled table's revision
+  # still converts -- it just has no name yet.
+  module FinanzamtRegistry
+    DATA_PATH = File.expand_path('data/finanzaemter.yml', __dir__)
+
+    class << self
+      def name_for(code)
+        entries[code.to_s]&.fetch(:name, nil)
+      end
+
+      def state_for(code)
+        entries[code.to_s]&.fetch(:state, nil)
+      end
+
+      def known?(code)
+        entries.key?(code.to_s)
+      end
+
+      def revision
+        data[:revision]
+      end
+
+      def entries
+        @entries ||= data[:finanzaemter].to_h do |code, attrs|
+          [code.to_s, { name: attrs['name'], state: attrs['state'] }.freeze]
+        end.freeze
+      end
+
+      private
+
+      def data
+        @data ||= begin
+          loaded = YAML.safe_load_file(DATA_PATH)
+          { revision: loaded['revision'], finanzaemter: loaded['finanzaemter'] || {} }
+        end
+      end
+    end
+  end
+end

data/lib/steuer/state_mapping.rb

@@ -119,6 +119,66 @@ module Steuer
       },
     }.freeze
     class << self
+      def standard_group_sizes(state_code)
+        config = STATES[state_code]
+        return unless config
+
+        sizes = config[:standard_pattern].source.delete_prefix('^').delete_suffix('$')
+          .split('/')
+          .map { |group| standard_group_size(group) }
+
+        return if sizes.any?(&:nil?) || sizes.empty?
+
+        sizes
+      end
+
+      # Ranges such as Hessen's 0(1[3-9]) and Saarland's (01[0-2]) stand in
+      # for single digits, so groups are measured rather than read off \d{n}.
+      def standard_group_size(group)
+        size = 0
+        scanner = group.dup
+        until scanner.empty?
+          case scanner
+          when /\A\\d\{(\d+)\}/
+            size += Regexp.last_match(1).to_i
+            scanner = scanner.sub(/\A\\d\{\d+\}/, '')
+          when /\A\[[^\]]*\]/
+            size += 1
+            scanner = scanner.sub(/\A\[[^\]]*\]/, '')
+          when /\A\\d/
+            size += 1
+            scanner = scanner.delete_prefix('\\d')
+          when /\A\d/
+            size += 1
+            scanner = scanner[1..]
+          when /\A[()]/
+            scanner = scanner[1..]
+          else
+            return
+          end
+        end
+        size
+      end
+
+      # Separators are presentation only: Baden-Württemberg prints 93815/08152
+      # for the ten digits the BZSt Standardschema groups as 93/815/08152.
+      def canonicalize_standard_format(tax_number, state_code)
+        sizes = standard_group_sizes(state_code)
+        return unless sizes
+
+        digits = tax_number.gsub(/\D/, '')
+        return unless digits.length == sizes.sum
+
+        offset = 0
+        groups = sizes.map do |size|
+          part = digits[offset, size]
+          offset += size
+          part
+        end
+
+        groups.join('/')
+      end
+
       def find_state_by_standard_format(tax_number)
         # Try more specific patterns first to avoid conflicts with general patterns
         # Priority: character classes with ranges > specific digit patterns > general digit patterns

data/lib/steuer/steuernummer.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
 require_relative 'state_mapping'
+require_relative 'finanzamt_registry'
 
 module Steuer
   class Steuernummer
@@ -8,7 +9,8 @@ module Steuer
 
     def initialize(tax_number, state: nil)
       @original_input = tax_number.to_s.strip
-      @state_code = normalize_state_code(state) || auto_detect_state
+      @explicit_state_code = normalize_state_code(state)
+      @state_code = @explicit_state_code || auto_detect_state
 
       validate_tax_number!
     end
@@ -61,7 +63,7 @@ module Steuer
 
       case format_type
       when :standard
-        @original_input
+        canonical_standard_input || @original_input
       when :federal_12
         convert_federal_12_to_standard
       when :federal_13
@@ -75,6 +77,19 @@ module Steuer
       StateMapping::STATES[@state_code][:name]
     end
 
+    # The BUFA-Nr: the leading four digits of the federal-13 form, identifying
+    # the issuing Finanzamt (the Empfaenger in an ELSTER transmission).
+    def finanzamt_code
+      to_federal_13&.slice(0, 4)
+    end
+
+    def finanzamt_name
+      code = finanzamt_code
+      return unless code
+
+      FinanzamtRegistry.name_for(code)
+    end
+
     def format_type
       @format_type ||= if normalized_input.include?('/')
         :standard
@@ -82,6 +97,8 @@ module Steuer
         :federal_12
       elsif normalized_input.length == 13 && normalized_input.match?(/^\d{13}$/)
         :federal_13
+      elsif unseparated_standard?
+        :standard
       end
 
       @format_type
@@ -178,7 +195,24 @@ module Steuer
 
     def validate_standard_format
       config = StateMapping::STATES[@state_code]
-      normalized_input.match?(config[:standard_pattern])
+      return true if normalized_input.match?(config[:standard_pattern])
+
+      canonical = canonical_standard_input
+      !canonical.nil? && canonical.match?(config[:standard_pattern])
+    end
+
+    def unseparated_standard?
+      return false unless @explicit_state_code
+      return false unless normalized_input.match?(/\A\d+\z/)
+
+      sizes = StateMapping.standard_group_sizes(@explicit_state_code)
+      !sizes.nil? && normalized_input.length == sizes.sum
+    end
+
+    def canonical_standard_input
+      return unless @state_code
+
+      StateMapping.canonicalize_standard_format(normalized_input, @state_code)
     end
 
     def validate_federal_12_format
@@ -200,7 +234,8 @@ module Steuer
     def convert_standard_to_federal_12
       config = StateMapping::STATES[@state_code]
 
-      match = normalized_input.match(config[:standard_pattern])
+      match = normalized_input.match(config[:standard_pattern]) ||
+        canonical_standard_input&.match(config[:standard_pattern])
       return unless match
 
       parts = match.captures

data/lib/steuer/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Steuer
-  VERSION = '1.0.4'
+  VERSION = '1.2.0'
 end

data/lib/steuer.rb

@@ -2,6 +2,7 @@
 
 require_relative 'steuer/version'
 require_relative 'steuer/steuernummer'
+require_relative 'steuer/finanzamt_registry'
 require_relative 'steuer/errors'
 
 module Steuer

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: steuer
 version: !ruby/object:Gem::Version
-  version: 1.0.4
+  version: 1.2.0
 platform: ruby
 authors:
 - Olumuyiwa Osiname
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-12-23 00:00:00.000000000 Z
+date: 2026-09-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: pry
@@ -133,7 +133,9 @@ files:
 - LICENSE
 - README.md
 - lib/steuer.rb
+- lib/steuer/data/finanzaemter.yml
 - lib/steuer/errors.rb
+- lib/steuer/finanzamt_registry.rb
 - lib/steuer/state_mapping.rb
 - lib/steuer/steuernummer.rb
 - lib/steuer/version.rb