loc_mods 0.2.1 → 0.2.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2b15a4efe6f5c0bf6287f01e8f3b9d14bf4611f0ec80c50a94eaf01aea0d27b0
-  data.tar.gz: 7c1c47497f88abc231351b6a19703641e8b7f910a8c119222872208dc339fe97
+  metadata.gz: 2b206d76220e77a18fd6609c3f7646eaf8fb836ac2d69a7d6da019db5820f24d
+  data.tar.gz: ffd23bb7508ad1fef1e0b165eb07620fc0a2420b6c436ac34cec8208affb6463
 SHA512:
-  metadata.gz: bb267c94c7de7b895ccf168ca9b0a37133afce833981a332e43f3f82822aeec27ada4b4becedb1b7196ebcaab97bd5c90c68ee8fd23ab463b82500b0a72f7b7a
-  data.tar.gz: 49f9d610121df295c904cede6245d4dd907bbba9c919879ccf8e44211b095bd23700452b4d4d6a670f2e5e579a6fc3e06ad70d1fa450f1c9de35ff1f01ff6144
+  metadata.gz: 9f56bbb4bfdab2dd2820ef67d94461ac5caa47d62937762a5cbeacbea8e70497c75968d92b3ce23f5d7e79f6a737c98c43058d6e11b774d0777340f040808fea
+  data.tar.gz: 0ac7feedcebe5df178f6ad313ba811004de126c82adc0f5289f44b43212a26ef2acfa3f09a45afa9cd4b158b7b4ef97976413ac357b4eb11749ab227c801a168

data/README.adoc

@@ -12,6 +12,8 @@ This gem is developed using the MODS 3.7 XSD schema.
 
 == Usage
 
+=== Ruby API
+
 [source,ruby]
 ----
 require 'loc_mods'
@@ -24,12 +26,55 @@ LocMods::Collection.from_xml(File.read("spec/fixtures/record_1.xml"))
 LocMods::Collection.from_xml(File.read("reference/allrecords-MODS.xml"))
 ----
 
+=== Command line interface
+
+LocMods provides a command-line interface (CLI) for various operations. The main
+executable is `loc-mods`.
+
+==== Detect duplicates
+
+The `detect-duplicates` command allows you to find duplicate MODS records based
+on using a "primary ID" that is their DOI (Digital Object Identifier).
+
+Usage:
+
+[source,shell]
+----
+$ loc-mods detect-duplicates [OPTIONS] <file_or_directory_path>
+----
+
+Options:
+
+* `-r, --recursive`: Search for MODS files recursively in subdirectories.
+* `-v, --verbose`: Display more detailed output.
+* `-o, --output FILE`: Write the results to a file instead of stdout.
+
+Example:
+
+[source,shell]
+----
+$ loc-mods detect-duplicates -r -v /path/to/mods/files
+----
+
+This command will:
+
+. Search for MODS XML files in the specified directory (and subdirectories if `-r` is used).
+. Parse each MODS file and extract the DOI.
+. Group records with the same DOI.
+. For each group of duplicates:
+.. Display the shared DOI.
+.. List the filenames of the duplicate records.
+.. Show a detailed comparison of the differences between the records.
+
+The output will highlight differences, removed elements, and missing elements
+between the duplicate records, helping you identify discrepancies in the
+metadata.
+
 == Testing
 
 [source,sh]
 ----
 bin/update-nist-mods
-
 ----
 
 == License

data/exe/loc-mods

@@ -0,0 +1,7 @@
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+
+require "bundler/setup"
+require_relative "../lib/loc_mods/cli"
+
+LocMods::Cli.start(ARGV)

data/lib/loc_mods/abstract.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class Abstract < Shale::Mapper
+  class Abstract < BaseMapper
     attribute :content, Shale::Type::String
     attribute :display_label, Shale::Type::String
     attribute :type, Shale::Type::String

data/lib/loc_mods/access_condition.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class AccessCondition < Shale::Mapper
+  class AccessCondition < BaseMapper
     attribute :content, Shale::Type::String
     attribute :display_label, Shale::Type::String
     attribute :lang, Shale::Type::String

data/lib/loc_mods/alternative_name.rb

@@ -1,13 +1,13 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 require_relative "identifier"
 require_relative "name_part"
 require_relative "role"
 
 module LocMods
-  class AlternativeName < Shale::Mapper
+  class AlternativeName < BaseMapper
     attribute :lang, Shale::Type::String
     attribute :script, Shale::Type::String
     attribute :transliteration, Shale::Type::String

data/lib/loc_mods/area.rb

@@ -1,11 +1,11 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 require_relative "hierarchical_part"
 
 module LocMods
-  class Area < Shale::Mapper
+  class Area < BaseMapper
     attribute :content, HierarchicalPart
     attribute :area_type, Shale::Type::Value
 

data/lib/loc_mods/base_mapper.rb

@@ -0,0 +1,26 @@
+# frozen_string_literal: true
+
+# lib/loc_mods/base_mapper.rb
+require "shale"
+require_relative "comparable_mapper"
+
+module LocMods
+  # Base class for all object definitions
+  class BaseMapper < Shale::Mapper
+    include ComparableMapper
+  end
+
+  # Nil class substitute for comparison
+  class ComparableNil < BaseMapper
+  end
+
+  # Comparison of two values for ComparableMapper
+  class Comparison
+    attr_accessor :original, :updated
+
+    def initialize(original:, updated:)
+      @original = original
+      @updated = updated
+    end
+  end
+end

data/lib/loc_mods/cartographic_extension.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class CartographicExtension < Shale::Mapper
+  class CartographicExtension < BaseMapper
     attribute :content, Shale::Type::String
     attribute :display_label, Shale::Type::String
 

data/lib/loc_mods/cartographics.rb

@@ -1,11 +1,11 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 require_relative "cartographic_extension"
 
 module LocMods
-  class Cartographics < Shale::Mapper
+  class Cartographics < BaseMapper
     attribute :authority, Shale::Type::String
     attribute :authority_uri, Shale::Type::Value
     attribute :value_uri, Shale::Type::Value

data/lib/loc_mods/city_section.rb

@@ -1,11 +1,11 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 require_relative "hierarchical_part"
 
 module LocMods
-  class CitySection < Shale::Mapper
+  class CitySection < BaseMapper
     attribute :content, HierarchicalPart
     attribute :city_section_type, Shale::Type::Value
 

data/lib/loc_mods/classification.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class Classification < Shale::Mapper
+  class Classification < BaseMapper
     attribute :content, Shale::Type::String
     attribute :edition, Shale::Type::String
     attribute :display_label, Shale::Type::String

data/lib/loc_mods/cli.rb

@@ -0,0 +1,153 @@
+# frozen_string_literal: true
+
+# lib/loc_mods/cli.rb
+require "thor"
+require "loc_mods"
+
+module LocMods
+  class Cli < Thor
+    desc "detect-duplicates PATH...", "Detect duplicate records in MODS XML files or directories"
+
+    def detect_duplicates(*paths)
+      all_records = []
+      paths.each do |path|
+        if File.directory?(path)
+          Dir.glob(File.join(path, "*.xml")).each do |file|
+            process_file(file, all_records)
+          end
+        elsif File.file?(path) && path.end_with?(".xml")
+          process_file(path, all_records)
+        else
+          puts "Warning: Skipping invalid path: #{path}"
+        end
+      end
+
+      records_by_url = {}
+      all_records.each do |record|
+        urls = record[:record].location.flat_map { |loc| loc.url.map(&:content) }.compact
+        unless urls.any?
+          puts "Warning: Record without URL found in file: #{record[:file]}"
+          next
+        end
+
+        urls.each do |url|
+          records_by_url[url] ||= []
+          records_by_url[url] << record
+        end
+      end
+
+      duplicate_count = 0
+      records_by_url.each do |url, records|
+        next unless records.size > 1
+
+        duplicate_count += 1
+        puts "Duplicate set ##{duplicate_count} found for URL: #{url}"
+        records.combination(2).each_with_index do |(record1, record2), index|
+          puts "  Comparison #{index + 1}:"
+          puts "  File 1: #{record1[:file]}"
+          puts "  File 2: #{record2[:file]}"
+          differences = record1[:record].compare(record2[:record])
+          if differences
+            puts "  ----"
+            print_differences(differences)
+            puts "  ----"
+          end
+          puts "\n"
+        end
+      end
+    end
+
+    private
+
+    def process_file(file, all_records)
+      xml_content = File.read(file)
+      collection = LocMods::Collection.from_xml(xml_content)
+      collection.mods.each do |record|
+        all_records << { record: record, file: file }
+      end
+    end
+
+    def print_differences(differences, prefix = "", path = [])
+      if differences.is_a?(Comparison)
+        print_difference(differences, path)
+        return
+      end
+
+      # Differences is a Hash here
+      differences.each do |key, value|
+        # puts "key #{key}, value #{value}"
+        current_path = path + [key]
+        # This is a comparison
+        if value.is_a?(Comparison)
+          print_difference(value, path)
+          next
+        end
+
+        raise "Differences must be in form of a Hash" unless value.is_a?(Hash)
+
+        # This is not array, end here
+        next unless value.keys.any? do |k|
+          k.is_a?(Integer) || k == :_array_size_difference
+        end
+
+        # if value[:_array_size_difference]
+        #   puts "  #{format_path(current_path)} [_array_size_difference]:"
+        #   puts "    Record 1: #{format_value(value[:_array_size_difference].original)}"
+        #   puts "    Record 2: #{format_value(value[:_array_size_difference].updated)}"
+        #   puts
+        # end
+
+        value.each do |subkey, subvalue|
+          # if [:self, :other].include?(subkey)
+          #   raise "Subkey is self or other"
+          # end
+          # puts "subkey (#{subkey})"
+          # puts "subvalue #{subvalue}, prefix #{prefix}, current_path + [subkey] #{current_path + [subkey]}"
+
+          if subkey.is_a?(Integer)
+            print_differences(subvalue, prefix, current_path + [subkey])
+          else
+            if subvalue.is_a?(Comparison)
+              print_difference(subvalue, current_path + [subkey])
+              next
+            end
+
+            raise "In an array diff but not an Integer!"
+          end
+        end
+      end
+    end
+
+    def print_difference(value, current_path)
+      return unless value.original || value.updated
+
+      puts "  #{format_path(current_path)}:"
+      puts "    Record 1: #{format_value(value.original)}"
+      puts "    Record 2: #{format_value(value.updated)}"
+      puts
+    end
+
+    def format_path(path)
+      path.map.with_index do |part, index|
+        if index.zero?
+          part.to_s
+        elsif part.is_a?(Integer)
+          "[#{part}]"
+        else
+          ".#{part}"
+        end
+      end.join
+    end
+
+    def format_value(value)
+      case value
+      when nil, ComparableNil
+        "(nil)"
+      when String
+        "\"#{value}\""
+      else
+        value.to_s
+      end
+    end
+  end
+end

data/lib/loc_mods/collection.rb

@@ -1,11 +1,11 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 require_relative "record"
 
 module LocMods
-  class Collection < Shale::Mapper
+  class Collection < BaseMapper
     attribute :mods, Record, collection: true
     attribute :schema_location, Shale::Type::String
 

data/lib/loc_mods/comparable_mapper.rb

@@ -0,0 +1,110 @@
+# frozen_string_literal: true
+
+# lib/loc_mods/comparable_mapper.rb
+
+module LocMods
+  # Enable comparison of two class models solely based on attribute values in
+  # a recursive manner.
+  module ComparableMapper
+    # TODO: Implement Comparable
+    # include Comparable
+    # def <=>(other)
+    #   attributes.foo <=> other.attributes.foo
+    # end
+    # def inspect
+    #   @foo
+    # end
+
+    def eql?(other)
+      other.class == self.class &&
+        self.class.attributes.all? do |attr|
+          send(attr) == other.send(attr)
+        end
+    end
+
+    alias == eql?
+
+    def hash
+      ([self.class] + self.class.attributes.map(&:hash)).hash
+    end
+
+    def compare(other)
+      differences = {}
+
+      # puts "Debugging: Attributes for #{self.class.name}"
+      # pp self.class.attributes.keys
+
+      self.class.attributes.each_key do |attr|
+        self_value = respond_to?(attr) ? send(attr) : nil
+        other_value = other.respond_to?(attr) ? other.send(attr) : nil
+
+        # puts "Debugging: Comparing attribute '#{attr}'"
+        # puts "  Self value: #{self_value.inspect}"
+        # puts "  Other value: #{other_value.inspect}"
+
+        compared = compare_values(self_value, other_value)
+        if compared
+          # puts "DETECTED DIFFERENCE! #{compared.inspect}"
+          differences[attr] = compared
+        end
+      end
+
+      # unless differences.empty?
+      #   puts "DIFFERENCES ARE"
+      #   pp differences
+      # end
+
+      differences.empty? ? nil : differences
+    end
+
+    private
+
+    def compare_values(self_value, other_value)
+      # puts "compare_values (self_value: #{self_value}, other_value: #{other_value})"
+      case self_value
+      when Array
+        # puts "compare_values case 1"
+        compare_arrays(self_value, other_value || [])
+      when ComparableMapper
+        # puts "compare_values case 2"
+        self_value.compare(other_value)
+      else
+        if self_value != other_value
+          # puts "compare_values case 3"
+          Comparison.new(original: self_value, updated: other_value)
+        end
+      end
+    end
+
+    def compare_arrays(self_array, other_array)
+      differences = {}
+      max_length = [self_array.size, other_array.size].max
+
+      max_length.times do |index|
+        self_item = self_array[index] || ComparableNil.new
+        other_item = other_array[index] || ComparableNil.new
+
+        compared = compare_values(self_item, other_item)
+        differences[index] = compared if compared
+
+        # if index >= self_array.size
+        #   puts "case 1 #{compared}"
+        #   # differences[index] = compared
+        # elsif index >= other_array.size
+        #   puts "case 2.1 #{self_item}"
+        #   puts "case 2.2 #{compared}"
+        #   # differences[index] = compared
+        # end
+      end
+
+      if self_array.size != other_array.size
+        differences[:_array_size_difference] = Comparison.new(
+          original: self_array.size,
+          updated: other_array.size
+        )
+      end
+
+      differences.empty? ? nil : differences
+    end
+  end
+end

data/lib/loc_mods/copy_information.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 require_relative "enumeration_and_chronology"
 require_relative "form"
@@ -8,7 +8,7 @@ require_relative "item_identifier"
 require_relative "note"
 
 module LocMods
-  class CopyInformation < Shale::Mapper
+  class CopyInformation < BaseMapper
     attribute :form, Form
     attribute :sub_location, Shale::Type::String, collection: true
     attribute :shelf_locator, Shale::Type::String, collection: true

data/lib/loc_mods/date.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class Date < Shale::Mapper
+  class Date < BaseMapper
     attribute :content, Shale::Type::String
     attribute :encoding, Shale::Type::String
     attribute :qualifier, Shale::Type::String

data/lib/loc_mods/date_other.rb

@@ -1,11 +1,11 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 require_relative "date"
 
 module LocMods
-  class DateOther < Shale::Mapper
+  class DateOther < BaseMapper
     attribute :content, Shale::Type::String
     attribute :type, Shale::Type::String
 

data/lib/loc_mods/detail.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class Detail < Shale::Mapper
+  class Detail < BaseMapper
     attribute :type, Shale::Type::String
     attribute :level, Shale::Type::Integer
     attribute :number, Shale::Type::String, collection: true

data/lib/loc_mods/edition.rb

@@ -1,10 +1,10 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
   # LocMods::StringPlusLanguagePlusSupplied
-  class Edition < Shale::Mapper
+  class Edition < BaseMapper
     attribute :content, Shale::Type::String
     attribute :supplied, Shale::Type::String
 

data/lib/loc_mods/enumeration_and_chronology.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class EnumerationAndChronology < Shale::Mapper
+  class EnumerationAndChronology < BaseMapper
     attribute :content, Shale::Type::String
     attribute :unit_type, Shale::Type::String
 

data/lib/loc_mods/extent.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class Extent < Shale::Mapper
+  class Extent < BaseMapper
     attribute :content, Shale::Type::String
     attribute :unit, Shale::Type::Value
 

data/lib/loc_mods/extent_definition.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class ExtentDefinition < Shale::Mapper
+  class ExtentDefinition < BaseMapper
     attribute :unit, Shale::Type::String
     attribute :start, Shale::Type::String
     attribute :end, Shale::Type::String

data/lib/loc_mods/form.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class Form < Shale::Mapper
+  class Form < BaseMapper
     attribute :content, Shale::Type::String
     attribute :type, Shale::Type::String
     attribute :authority, Shale::Type::String

data/lib/loc_mods/genre.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class Genre < Shale::Mapper
+  class Genre < BaseMapper
     attribute :content, Shale::Type::String
     attribute :type, Shale::Type::String
     attribute :display_label, Shale::Type::String

data/lib/loc_mods/geographic_code.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 module LocMods
-  class GeographicCode < Shale::Mapper
+  class GeographicCode < BaseMapper
     attribute :content, Shale::Type::String
     attribute :authority_uri, Shale::Type::Value
     attribute :value_uri, Shale::Type::Value

data/lib/loc_mods/hierarchical_geographic.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "shale"
+require_relative "base_mapper"
 
 require_relative "area"
 require_relative "city_section"
@@ -8,7 +8,7 @@ require_relative "hierarchical_part"
 require_relative "region"
 
 module LocMods
-  class HierarchicalGeographic < Shale::Mapper
+  class HierarchicalGeographic < BaseMapper
     attribute :authority, Shale::Type::String
     attribute :authority_uri, Shale::Type::Value
     attribute :value_uri, Shale::Type::Value