unitsdb 1.0.0 → 2.1.0

data/lib/unitsdb/dimension.rb

@@ -0,0 +1,49 @@
+# frozen_string_literal: true
+
+require_relative "identifier"
+require_relative "dimension_details"
+require_relative "quantity"
+require_relative "localized_string"
+# NISTd1:
+#   length:
+#     power: 1
+#     symbol: L
+#     dim_symbols:
+#     - id: "dim_L"
+#       ascii: "L"
+#       html: "𝖫"
+#       mathml: "<mi mathvariant='sans-serif'>L</mi>"
+#       latex: \ensuremath{\mathsf{L}}
+#       unicode: "𝖫"
+
+# NISTd9:
+# -dimensionless: true
+# -plane_angle:
+# -  dim_symbols:
+# -  - ascii: phi
+# -    html: "&#x1D785;"
+# -    id: dim_phi
+# -    latex: "\\ensuremath{\\mathsf{\\phi}}"
+# -    mathml: "<mi mathvariant='sans-serif'>&#x3c6;</mi>"
+# -    unicode: "\U0001D785"
+# -  power: 1
+# -  symbol: phi
+
+module Unitsdb
+  class Dimension < Lutaml::Model::Serializable
+    # model Config.model_for(:dimension)
+
+    attribute :identifiers, Identifier, collection: true
+    attribute :dimensionless, :boolean
+    attribute :length, DimensionDetails
+    attribute :mass, DimensionDetails
+    attribute :time, DimensionDetails
+    attribute :electric_current, DimensionDetails
+    attribute :thermodynamic_temperature, DimensionDetails
+    attribute :amount_of_substance, DimensionDetails
+    attribute :luminous_intensity, DimensionDetails
+    attribute :plane_angle, DimensionDetails
+    attribute :short, :string
+    attribute :names, LocalizedString, collection: true
+  end
+end

data/lib/unitsdb/dimension_details.rb

@@ -0,0 +1,20 @@
+# frozen_string_literal: true
+
+# power: 1
+# symbol: M
+# symbols:
+# - ascii: M
+#   html: "𝖬"
+#   id: dim_M
+#   latex: "\\ensuremath{\\mathsf{M}}"
+#   mathml: "<mi mathvariant='sans-serif'>M</mi>"
+#   unicode: "\U0001D5AC"
+
+require_relative "symbol_presentations"
+module Unitsdb
+  class DimensionDetails < Lutaml::Model::Serializable
+    attribute :power, :integer
+    attribute :symbol, :string
+    attribute :symbols, SymbolPresentations, collection: true
+  end
+end

data/lib/unitsdb/dimension_reference.rb

@@ -0,0 +1,8 @@
+# frozen_string_literal: true
+
+module Unitsdb
+  class DimensionReference < Identifier
+    attribute :id, :string
+    attribute :type, :string
+  end
+end

data/lib/unitsdb/dimensions.rb

@@ -1,18 +1,12 @@
 # frozen_string_literal: true
 
-require_relative "dimensions/dimension"
+require_relative "dimension"
 
 module Unitsdb
-  class Dimensions
-    include Lutaml::Model::Serialize
-    model Config.model_for(:dimensions)
+  class Dimensions < Lutaml::Model::Serializable
+    # model Config.model_for(:dimensions)
 
+    attribute :schema_version, :string
     attribute :dimensions, Dimension, collection: true
-
-    key_value do
-      map to: :dimensions, root_mappings: {
-        id: :key
-      }
-    end
   end
 end

data/lib/unitsdb/errors.rb

@@ -0,0 +1,13 @@
+# frozen_string_literal: true
+
+module Unitsdb
+  module Errors
+    class DatabaseError < StandardError; end
+    class DatabaseNotFoundError < DatabaseError; end
+    class DatabaseFileNotFoundError < DatabaseError; end
+    class DatabaseFileInvalidError < DatabaseError; end
+    class DatabaseLoadError < DatabaseError; end
+    class VersionMismatchError < DatabaseError; end
+    class UnsupportedVersionError < DatabaseError; end
+  end
+end

data/lib/unitsdb/external_reference.rb

@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+
+# references:
+# - uri: http://si-digital-framework.org/quantities/LENG
+#   type: normative
+#   authority: si-digital-framework
+
+module Unitsdb
+  class ExternalReference < Identifier
+    attribute :uri, :string
+    attribute :type, :string, values: %w[normative informative]
+    attribute :authority, :string
+  end
+end

data/lib/unitsdb/identifier.rb

@@ -0,0 +1,8 @@
+# frozen_string_literal: true
+
+module Unitsdb
+  class Identifier < Lutaml::Model::Serializable
+    attribute :id, :string
+    attribute :type, :string
+  end
+end

data/lib/unitsdb/localized_string.rb

@@ -0,0 +1,17 @@
+# frozen_string_literal: true
+
+module Unitsdb
+  # Represents a localized string with a language tag
+  class LocalizedString < Lutaml::Model::Serializable
+    attribute :value, :string
+    attribute :lang, :string
+
+    def to_s
+      "#{value} (#{lang})"
+    end
+
+    def downcase
+      value&.downcase
+    end
+  end
+end

data/lib/unitsdb/prefix.rb

@@ -0,0 +1,30 @@
+# frozen_string_literal: true
+
+require_relative "identifier"
+require_relative "symbol_presentations"
+require_relative "external_reference"
+require_relative "localized_string"
+# ---
+# NISTp10_30:
+#   name: quetta
+#   symbol:
+#     ascii: Q
+#     html: Q
+#     latex: Q
+#     unicode: Q
+#   base: 10
+#   power: 30
+
+module Unitsdb
+  class Prefix < Lutaml::Model::Serializable
+    # model Config.model_for(:prefix)
+
+    attribute :identifiers, Identifier, collection: true
+    attribute :names, LocalizedString, collection: true
+    attribute :short, :string
+    attribute :symbols, SymbolPresentations, collection: true
+    attribute :base, :integer
+    attribute :power, :integer
+    attribute :references, ExternalReference, collection: true
+  end
+end

data/lib/unitsdb/prefix_reference.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+require_relative "identifier"
+
+module Unitsdb
+  class PrefixReference < Identifier
+    attribute :id, :string
+    attribute :type, :string
+  end
+end

data/lib/unitsdb/prefixes.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require_relative "prefixes/prefix"
+require_relative "prefix"
 # ---
 # NISTp10_30:
 #   name: quetta
@@ -13,17 +13,10 @@ require_relative "prefixes/prefix"
 #   power: 30
 
 module Unitsdb
-  class Prefixes
-    include Lutaml::Model::Serialize
-
-    model Config.model_for(:prefixes)
+  class Prefixes < Lutaml::Model::Serializable
+    # model Config.model_for(:prefixes)
 
+    attribute :schema_version, :string
     attribute :prefixes, Prefix, collection: true
-
-    key_value do
-      map to: :prefixes, root_mappings: {
-        id: :key
-      }
-    end
   end
 end

data/lib/unitsdb/quantities.rb

@@ -1,38 +1,11 @@
 # frozen_string_literal: true
 
-# # Quantities
-# NISTq156:
-#   dimension_url: "#NISTd68"
-#   quantity_type: derived
-#   quantity_name:
-#       - linear expansion coefficient
-#   unit_reference:
-#       - name: kelvin to the power minus one
-#         url: "#NISTu5e-1/1"
-
-# NISTq155:
-#   dimension_url: "#NISTd57"
-#   quantity_type: derived
-#   quantity_name:
-#     - area moment of inertia
-#     - second moment of area
-#   unit_reference:
-#     - name: inch to the fourth power
-#       url: "#NISTu208"
-
-require_relative "quantities/quantity"
+require_relative "quantity"
 
 module Unitsdb
-  class Quantities
-    include Lutaml::Model::Serialize
-    model Config.model_for(:quantities)
-
+  class Quantities < Lutaml::Model::Serializable
+    # model Config.model_for(:quantities)
+    attribute :schema_version, :string
     attribute :quantities, Quantity, collection: true
-
-    key_value do
-      map to: :quantities, root_mappings: {
-        id: :key
-      }
-    end
   end
 end

data/lib/unitsdb/quantity.rb

@@ -0,0 +1,21 @@
+# frozen_string_literal: true
+
+require_relative "identifier"
+require_relative "unit_reference"
+require_relative "dimension_reference"
+require_relative "external_reference"
+require_relative "localized_string"
+
+module Unitsdb
+  class Quantity < Lutaml::Model::Serializable
+    # model Config.model_for(:quantity)
+
+    attribute :identifiers, Identifier, collection: true
+    attribute :quantity_type, :string
+    attribute :names, LocalizedString, collection: true
+    attribute :short, :string
+    attribute :unit_references, UnitReference, collection: true
+    attribute :dimension_reference, DimensionReference
+    attribute :references, ExternalReference, collection: true
+  end
+end

data/lib/unitsdb/quantity_reference.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+module Unitsdb
+  class QuantityReference < Identifier
+    # model Config.model_for(:quantity_reference)
+
+    attribute :id, :string
+    attribute :type, :string
+  end
+end

data/lib/unitsdb/root_unit_reference.rb

@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+
+require_relative "unit_reference"
+require_relative "prefix_reference"
+
+module Unitsdb
+  class RootUnitReference < Lutaml::Model::Serializable
+    # model Config.model_for(:root_unit)
+
+    attribute :power, :integer
+    attribute :unit_reference, UnitReference
+    attribute :prefix_reference, PrefixReference
+  end
+end

data/lib/unitsdb/scale.rb

@@ -0,0 +1,17 @@
+# frozen_string_literal: true
+
+require_relative "identifier"
+require_relative "localized_string"
+require_relative "scale_properties"
+
+module Unitsdb
+  class Scale < Lutaml::Model::Serializable
+    # model Config.model_for(:quantity)
+
+    attribute :identifiers, Identifier, collection: true
+    attribute :names, LocalizedString, collection: true
+    attribute :description, LocalizedString, collection: true
+    attribute :short, :string
+    attribute :properties, ScaleProperties
+  end
+end

data/lib/unitsdb/scale_properties.rb

@@ -0,0 +1,12 @@
+# frozen_string_literal: true
+
+module Unitsdb
+  class ScaleProperties < Lutaml::Model::Serializable
+    # model Config.model_for(:quantity)
+    attribute :continuous, :boolean
+    attribute :ordered, :boolean
+    attribute :logarithmic, :boolean
+    attribute :interval, :boolean
+    attribute :ratio, :boolean
+  end
+end

data/lib/unitsdb/scale_reference.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+require_relative "identifier"
+
+module Unitsdb
+  class ScaleReference < Identifier
+    attribute :id, :string
+    attribute :type, :string
+  end
+end

data/lib/unitsdb/scales.rb

@@ -0,0 +1,11 @@
+# frozen_string_literal: true
+
+require_relative "scale"
+
+module Unitsdb
+  class Scales < Lutaml::Model::Serializable
+    # model Config.model_for(:Scale)
+    attribute :schema_version, :string
+    attribute :scales, Scale, collection: true
+  end
+end

data/lib/unitsdb/si_derived_base.rb

@@ -0,0 +1,19 @@
+# frozen_string_literal: true
+
+require_relative "root_unit_reference"
+
+# si_derived_bases:
+# - power: 2
+#   unit_reference:
+#     id: NISTu1
+#     type: nist
+# - power: -2
+#   unit_reference:
+#     id: NISTu1
+#     type: nist
+
+module Unitsdb
+  class SiDerivedBase < RootUnitReference
+    # model Config.model_for(:si_derived_base)
+  end
+end

data/lib/unitsdb/symbol_presentations.rb

@@ -2,18 +2,13 @@
 
 module Unitsdb
   class SymbolPresentations < Lutaml::Model::Serializable
-    model Config.model_for(:symbol_presentations)
+    # model Config.model_for(:symbol_presentations)
 
+    attribute :id, :string
     attribute :ascii, :string
     attribute :html, :string
     attribute :latex, :string
+    attribute :mathml, :string
     attribute :unicode, :string
-
-    key_value do
-      map :ascii, to: :ascii
-      map :html, to: :html
-      map :latex, to: :latex
-      map :unicode, to: :unicode
-    end
   end
 end

data/lib/unitsdb/unit.rb

@@ -0,0 +1,63 @@
+# frozen_string_literal: true
+
+require_relative "symbol_presentations"
+require_relative "unit_system_reference"
+require_relative "root_unit_reference"
+require_relative "si_derived_base"
+require_relative "quantity_reference"
+require_relative "dimension_reference"
+require_relative "external_reference"
+require_relative "localized_string"
+require_relative "scale_reference"
+
+# "NISTu10":
+#   dimension_url: "#NISTd9"
+#   short: steradian
+#   root: true
+#   unit_system:
+#     type: "SI_derived_special"
+#     name: "SI"
+#   names:
+#     - "steradian"
+#   unit_symbols:
+#     - id: "sr"
+#       ascii: "sr"
+#       html: "sr"
+#       mathml: "<mi mathvariant='normal'>sr</mi>"
+#       latex: \ensuremath{\mathrm{sr}}
+#       unicode: "sr"
+#   root_units:
+#     enumerated_root_units:
+#       - unit: "steradian"
+#         power_denominator: 1
+#         power: 1
+#   quantity_reference:
+#     - name: "solid angle"
+#       url: "#NISTq11"
+#   si_derived_bases:
+#     - id: NISTu1
+#       prefix:
+#       power: 1
+#     - id: NISTu1
+#       prefix:
+#       power: -1
+
+module Unitsdb
+  class Unit < Lutaml::Model::Serializable
+    # model Config.model_for(:unit)
+
+    attribute :identifiers, Identifier, collection: true
+    attribute :short, :string
+    attribute :root, :boolean
+    attribute :prefixed, :boolean
+    attribute :dimension_reference, DimensionReference
+    attribute :unit_system_reference, UnitSystemReference, collection: true
+    attribute :names, LocalizedString, collection: true
+    attribute :symbols, SymbolPresentations, collection: true
+    attribute :quantity_references, QuantityReference, collection: true
+    attribute :si_derived_bases, SiDerivedBase, collection: true
+    attribute :root_units, RootUnitReference, collection: true
+    attribute :references, ExternalReference, collection: true
+    attribute :scale_reference, ScaleReference
+  end
+end

data/lib/unitsdb/unit_reference.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+require_relative "identifier"
+
+module Unitsdb
+  class UnitReference < Identifier
+    attribute :id, :string
+    attribute :type, :string
+  end
+end

data/lib/unitsdb/unit_system.rb

@@ -0,0 +1,15 @@
+# frozen_string_literal: true
+
+require_relative "identifier"
+require_relative "localized_string"
+
+module Unitsdb
+  class UnitSystem < Lutaml::Model::Serializable
+    # model Config.model_for(:unit_system)
+
+    attribute :identifiers, Identifier, collection: true
+    attribute :names, LocalizedString, collection: true
+    attribute :short, :string
+    attribute :acceptable, :boolean
+  end
+end

data/lib/unitsdb/unit_system_reference.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+require_relative "identifier"
+
+module Unitsdb
+  class UnitSystemReference < Identifier
+    attribute :id, :string
+    attribute :type, :string
+  end
+end

data/lib/unitsdb/unit_systems.rb

@@ -1,18 +1,12 @@
 # frozen_string_literal: true
 
-require_relative "unit_systems/unit_system"
+require_relative "unit_system"
 
 module Unitsdb
-  class UnitSystems
-    include Lutaml::Model::Serialize
-    model Config.model_for(:unit_systems)
+  class UnitSystems < Lutaml::Model::Serializable
+    # model Config.model_for(:unit_systems)
 
+    attribute :schema_version, :string
     attribute :unit_systems, UnitSystem, collection: true
-
-    key_value do
-      map to: :unit_systems, root_mappings: {
-        id: :key
-      }
-    end
   end
 end

data/lib/unitsdb/units.rb

@@ -1,18 +1,12 @@
 # frozen_string_literal: true
 
-require "unitsdb/units/unit"
+require_relative "unit"
 
 module Unitsdb
-  class Units
-    include Lutaml::Model::Serialize
-    model Config.model_for(:units)
+  class Units < Lutaml::Model::Serializable
+    # model Config.model_for(:units)
 
+    attribute :schema_version, :string
     attribute :units, Unit, collection: true
-
-    key_value do
-      map to: :units, root_mappings: {
-            id: :key,
-          }
-    end
   end
 end

data/lib/unitsdb/utils.rb

@@ -0,0 +1,84 @@
+# frozen_string_literal: true
+
+module Unitsdb
+  module Utils
+    DEFAULT_YAML_FILES = %w[dimensions prefixes quantities unit_systems units].map { |f| "#{f}.yaml" }
+
+    def self.levenshtein_distance(s, t)
+      m = s.length
+      n = t.length
+
+      return m if n.zero?
+      return n if m.zero?
+
+      d = Array.new(m + 1) { Array.new(n + 1) }
+
+      (0..m).each { |i| d[i][0] = i }
+      (0..n).each { |j| d[0][j] = j }
+
+      (1..n).each do |j|
+        (1..m).each do |i|
+          d[i][j] = if s[i - 1] == t[j - 1]
+                      d[i - 1][j - 1]
+                    else
+                      [d[i - 1][j] + 1, d[i][j - 1] + 1, d[i - 1][j - 1] + 1].min
+                    end
+        end
+      end
+
+      d[m][n]
+    end
+
+    def self.find_similar_ids(id, valid_ids, max_results = 5)
+      # Simple similarity measure based on Levenshtein distance
+      valid_ids
+        .map { |valid_id| [valid_id, levenshtein_distance(id.to_s, valid_id.to_s)] }
+        .sort_by { |_, distance| distance }
+        .take(max_results)
+        .select { |_, distance| distance < [id.to_s.length, 10].min }
+        .map { |valid_id, _| valid_id }
+    end
+
+    def self.sort_yaml_keys(obj)
+      case obj
+      when Hash
+        # Transform values first, then sort keys
+        obj.transform_values { |v| sort_yaml_keys(v) }
+           .sort_by do |key, _|
+             key_str = key.to_s
+
+             # Handle NIST IDs with proper category-based sorting
+             case key_str
+             when /\ANIST([a-z])(\d+)_(-?\d+)\Z/ # Prefixes with power (NISTp2_10)
+               [0, ::Regexp.last_match(1), ::Regexp.last_match(2).to_i, ::Regexp.last_match(3).to_i]
+             when /\ANIST([a-z])(\d+)\Z/ # Prefixes without power
+               [0, ::Regexp.last_match(1), ::Regexp.last_match(2).to_i, 0]
+             when %r{[/e]} # Composite IDs with '/' or 'e' to the end
+               [2, key_str]
+             when /\ANISTu(\d+)\.(.+)\Z/ # Simple unit IDs (NISTu10.1)
+               [1, "u", ::Regexp.last_match(1).to_i, ::Regexp.last_match(2)]
+             else
+               [3, key_str] # Everything else
+             end
+           end.to_h
+      when Array
+        # For arrays in the new structure, sort them if they're arrays of hashes with 'id' or 'short' keys
+        if !obj.empty? && obj.all? { |item| item.is_a?(Hash) }
+          # Sort by 'id' within identifiers array
+          if obj.first.key?("id") && obj.first.key?("type")
+            obj.sort_by { |item| item["id"].to_s }.map { |item| sort_yaml_keys(item) }
+          # Sort by 'short' for other arrays (like main unit arrays)
+          elsif obj.first.key?("short")
+            obj.sort_by { |item| item["short"].to_s }.map { |item| sort_yaml_keys(item) }
+          else
+            obj.map { |item| sort_yaml_keys(item) }
+          end
+        else
+          obj.map { |item| sort_yaml_keys(item) }
+        end
+      else
+        obj
+      end
+    end
+  end
+end

data/lib/unitsdb/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Unitsdb
-  VERSION = "1.0.0"
+  VERSION = "2.1.0"
 end