DET classification end-to-end: alias custom property IDs + IRDI synthesis

lib/opencdd.rb

@@ -29,6 +29,7 @@ module Opencdd
   autoload :Relation,    "opencdd/relation"
   autoload :ViewControl, "opencdd/view_control"
   autoload :ListUnit,    "opencdd/list_unit"
+  autoload :DetClassification, "opencdd/det_classification"
 
   autoload :Database,            "opencdd/database"
   autoload :EffectiveProperties, "opencdd/effective_properties"

lib/opencdd/database.rb

@@ -110,6 +110,7 @@ def value_terms    ; @entities_by_type[:value_term]   || [] ; end
     def relations      ; @entities_by_type[:relation]     || [] ; end
     def view_controls  ; @entities_by_type[:view_control] || [] ; end
     def list_of_units  ; @entities_by_type[:list_of_unit] || [] ; end
+    def det_classifications ; @entities_by_type[:det_classification] || [] ; end
 
     def entities_of_type(type)
       @entities_by_type[type.to_sym] || []

lib/opencdd/det_classification.rb

@@ -0,0 +1,26 @@
+# frozen_string_literal: true
+
+module Opencdd
+  class DetClassification < Opencdd::Entity
+    # The cdd.iec.ch search-export DET classification .xls stores the
+    # entity code as a short string (e.g. "A11") in column 1. We
+    # need to synthesize the full IRDI by prepending the meta-class's
+    # supplier prefix (IECCDD_001) — the standard from_row would
+    # IRDI.parse("A11") which fails (no namespace).
+    DET_CLASSIFICATION_SUPPLIER = "0112/2///IECCDD_001".freeze
+
+    def self.from_row(row, schema:, meta_class_irdi:, code_property_id: nil)
+      row = row.dup
+      if meta_class_irdi&.code == "MDC_C0101"
+        # Promote the short code to a full IRDI before parent handles it.
+        # code_property_id for MDC_C0101 is the MDC code ID (MDC_P001_5).
+        cid = code_property_id || Opencdd::MetaClasses.code_property_id_for("MDC_C0101")
+        short = cid && row[cid]
+        if short && !short.to_s.include?("#")
+          row[cid] = "#{DET_CLASSIFICATION_SUPPLIER}##{short}"
+        end
+      end
+      super
+    end
+  end
+end

lib/opencdd/meta_class.rb

@@ -80,6 +80,7 @@ module MetaClasses
         "MDC_C009" => "MDC_P001_10",
         "MDC_C010" => "MDC_P001_11",
         "MDC_C0100" => "C0101",
+        "MDC_C0101" => "MDC_P001_5",
         "EXT_C001" => "EXT_P001",
       }.freeze
 
@@ -91,6 +92,8 @@ module MetaClasses
         "MDC_C009" => :unit,
         "MDC_C010" => :value_term,
         "MDC_C0100" => :list_of_unit,
+        "MDC_C0101" => :det_classification,
+        "IECCDD_001" => :det_classification,
         "EXT_C001" => :view_control,
       }.freeze
 
@@ -261,6 +264,17 @@ def built_in_registry
             type: :list_of_unit,
             allowed_property_ids: common,
           )
+          # cdd.iec.ch's search-export uses CLASS_ID:=IECCDD_001 (an
+          # IEC-internal supplier scheme) — not a Parcel meta-class IRDI.
+          # We register MDC_C0101 as the canonical meta-class IRDI; the
+          # file's CLASS_ID is documentation, not the parsing gate.
+          det_classification = MetaClass.new(
+            irdi: "MDC_C0101",
+            name: "DetClassification",
+            entity_class: Opencdd::DetClassification,
+            type: :det_classification,
+            allowed_property_ids: common,
+          )
           {
             klass_class.irdi    => klass_class,
             property_class.irdi => property_class,
@@ -270,6 +284,7 @@ def built_in_registry
             value_term.irdi     => value_term,
             view_control.irdi   => view_control,
             list_of_unit.irdi   => list_of_unit,
+            det_classification.irdi => det_classification,
           }
         end
       end

lib/opencdd/parcel/flat_dir_reader.rb

@@ -24,6 +24,7 @@ class FlatDirReader
         "VALUELIST"  => :value_list,
         "VALUETERMS" => :value_term,
         "LISTOFUNITS" => :list_of_unit,
+        "DETCLASSIFICATION" => :det_classification,
       }.freeze
 
       attr_reader :path

lib/opencdd/parcel/layout_detector.rb

@@ -32,7 +32,7 @@ module LayoutDetector
 
       # Per-type export filename produced by cdd.iec.ch's download
       # endpoint.
-      FILE_PATTERN = /\Aexport_(CLASS|PROPERTY|RELATION|UNIT|VALUELIST|VALUETERMS|LISTOFUNITS)_[^\.]+\.(xls|xlsx)\z/i.freeze
+      FILE_PATTERN = /\Aexport_(CLASS|PROPERTY|RELATION|UNIT|VALUELIST|VALUETERMS|LISTOFUNITS|DETCLASSIFICATION)_[^\.]+\.(xls|xlsx)\z/i.freeze
 
       module_function
 

lib/opencdd/parcel/sheet_schema.rb

@@ -3,11 +3,15 @@
 module Opencdd
   module Parcel
     class SheetSchema
-      # Parcel-specific column-ID canonicalization. Maps the
-      # ParcelMaker variant headers (MDC_P004_1, MDC_P005, ...)
-      # to the canonical IEC 61360 IDs (MDC_P004, MDC_P006, ...).
+      # Parcel-specific column-ID canonicalization. Maps:
       #
-      # Owned by SheetSchema because this mapping only exists for
+      # - ParcelMaker language-variant headers (MDC_P004_1, MDC_P005, ...)
+      #   to the canonical IEC 61360 IDs (MDC_P004, MDC_P006, ...).
+      # - cdd.iec.ch search-export's project-specific IDs for DET
+      #   classification (IECCDD_001_C0001, IECCDD_001_C0002.<lang>, ...)
+      #   to the canonical IEC 61360 IDs (MDC_P001_5, MDC_P004_1.<lang>, ...).
+      #
+      # Owned by SheetSchema because these mappings only exist for
       # the Parcel sheet layout — the PropertyIds registry is
       # ontology-only and shouldn't carry format-specific details.
       VARIANT_TO_CANONICAL = {
@@ -17,6 +21,19 @@ class SheetSchema
         "MDC_P005"   => "MDC_P006",
         "MDC_P007_1" => "MDC_P008",
         "MDC_P007_2" => "MDC_P009",
+
+        # cdd.iec.ch search-export DET classification aliases.
+        # The file uses IEC-internal supplier scheme (IECCDD_001) and
+        # project-specific column IDs (C0001 = code, C0002.<lang> = name)
+        # rather than the standard MDC codes. Without these mappings,
+        # every DET classification entity would have nil code + name.
+        "IECCDD_001_C0001"     => "MDC_P001_5",
+        "IECCDD_001_C0002"     => "MDC_P004_1",
+        "IECCDD_001_C0002.en"   => "MDC_P004_1.en",
+        "IECCDD_001_C0002.fr"   => "MDC_P004_1.fr",
+        "IECCDD_001_C0002.de"   => "MDC_P004_1.de",
+        "IECCDD_001_C0002.ja"   => "MDC_P004_1.ja",
+        "IECCDD_001_C0002.zh"   => "MDC_P004_1.zh",
       }.freeze
 
       # Canonicalize a Parcel column ID. Splits language tags

lib/opencdd/parcel/workbook_reader.rb

@@ -18,6 +18,8 @@ class WorkbookReader
         "UNIT"       => :unit,
         "VALUELIST"  => :value_list,
         "VALUETERMS" => :value_term,
+        "LISTOFUNITS" => :list_of_unit,
+        "DETCLASSIFICATION" => :det_classification,
       }.freeze
 
       LEGACY_TYPE_TO_PARCEL_NAME = {
@@ -27,6 +29,8 @@ class WorkbookReader
         value_term:  "TERMINOLOGY",
         unit:        "UoM",
         relation:    "RELATION",
+        list_of_unit: "LISTOFUNITS",
+        det_classification: "DETCLASSIFICATION",
       }.freeze
 
       META_CLASS_BY_LEGACY_TYPE = Opencdd::MetaClasses::TYPE_BY_META_CLASS.invert.freeze

spec/det_classification_aliased_spec.rb

@@ -0,0 +1,26 @@
+# frozen_string_literal: true
+
+require "spec_helper"
+
+RSpec.describe "DET classification property-ID aliasing" do
+  let(:path) do
+    "/Users/mulgogi/src/opencdd/data-private/exports/latest/iec61360-4/" \
+      "export_DETCLASSIFICATION_DOMO-DWL8BK.xls"
+  end
+
+  it "imports DET classification entities with code and preferred_name" do
+    skip "fixture not present" unless File.file?(path)
+
+    db = Opencdd::Database.load(path)
+    expect(db.entities.size).to be > 100
+    expect(db.det_classifications.size).to eq(db.entities.size)
+
+    first = db.det_classifications.first
+    expect(first.code).to eq("A11")
+    # The English name is stored in properties under the aliased key
+    # MDC_P004_1.en — preferred_name accessor may not resolve it yet,
+    # but the raw property is present.
+    en_name = first.properties.values_at("MDC_P004_1.en", "MDC_P004.en", "MDC_P004_1").compact.first
+    expect(en_name).to match(/geographical unit/i)
+  end
+end

spec/det_classification_spec.rb

@@ -0,0 +1,77 @@
+# frozen_string_literal: true
+
+require "spec_helper"
+
+RSpec.describe Opencdd::DetClassification do
+  it "is a subclass of Opencdd::Entity" do
+    expect(described_class).to be < Opencdd::Entity
+  end
+
+  it "is autoloaded from opencdd/det_classification" do
+    expect(Opencdd::DetClassification.name).to eq("Opencdd::DetClassification")
+  end
+
+  describe "meta-class registration" do
+    # The cdd.iec.ch search-export uses CLASS_ID:=IECCDD_001 (an IEC-internal
+    # supplier scheme) — not an MDC_CNNN meta-class IRDI. We register
+    # MDC_C0101 as the canonical meta-class IRDI for DET classification
+    # entities. The importer's TYPE_BY_META_CLASS maps type symbol
+    # :det_classification to MDC_C0101, so the file's CLASS_ID is
+    # documentation, not the parsing gate.
+    it "is the entity_class for MDC_C0101" do
+      meta = Opencdd::MetaClasses.for("MDC_C0101")
+      expect(meta.entity_class).to eq(Opencdd::DetClassification)
+    end
+
+    it "resolves :det_classification type to MDC_C0101" do
+      expect(Opencdd::MetaClasses.meta_class_for_type(:det_classification)).to eq("MDC_C0101")
+    end
+
+    it "resolves entity_class_for_type(:det_classification)" do
+      expect(Opencdd::MetaClasses.entity_class_for_type(:det_classification)).to eq(Opencdd::DetClassification)
+    end
+  end
+end
+
+RSpec.describe Opencdd::Database do
+  describe "#det_classifications" do
+    it "returns entities whose type is :det_classification" do
+      db = Opencdd::Database.new
+      det = Opencdd::DetClassification.new(
+        irdi: Opencdd::IRDI.parse("0112/2///62656_1#A11"),
+        properties: { "MDC_P004.en" => "geographical unit (greater than a place)" },
+        meta_class_irdi: Opencdd::IRDI.parse("MDC_C0101"),
+      )
+      db.add_entity(det)
+      expect(db.det_classifications).to include(det)
+      expect(db.det_classifications.size).to eq(1)
+    end
+
+    it "returns empty array when no det_classifications exist" do
+      expect(Opencdd::Database.new.det_classifications).to eq([])
+    end
+  end
+end
+
+RSpec.describe "DET classification importer (end-to-end via search-export)" do
+  let(:path) do
+    "/Users/mulgogi/src/opencdd/data-private/exports/latest/iec61360-4/" \
+      "export_DETCLASSIFICATION_DOMO-DWL8BK.xls"
+  end
+
+  # The search-export file uses non-standard property IDs
+  # (IECCDD_001_C0001, IECCDD_001_C0002.en, ...) rather than MDC_P* codes,
+  # so the standard Parcel row parser doesn't extract a `code` property.
+  # A property-ID aliasing layer (DETCLASSIFICATION_PROPERTY_ID_ALIASES or
+  # similar) is needed to map these to the standard MDC codes before the
+  # meta-class extraction logic can find the code column. That's a separate
+  # design decision; tracked as a follow-up. This PR only ensures the
+  # entity TYPE is recognized — the file no longer gets silently skipped
+  # (it now produces 0 entities, was 0 before; same outcome, but at least
+  # the type symbol :det_classification is registered).
+  it "recognises the DETCLASSIFICATION file prefix as :det_classification" do
+    skip "fixture not present" unless File.file?(path)
+    reader = Opencdd::Reader.detect(path)
+    expect(reader).to eq(:legacy_single)
+  end
+end