mods 3.0.0.alpha2 → 3.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6490e1cc58f876632e1d50065bab857eb104c4e0af18e8d1c3c18624fcdcf228
-  data.tar.gz: 4c16e2ca22fa7803d20f521b21514317f31c13f689eca1d5d7eaee3e5d2ecd80
+  metadata.gz: 30381fee9ce7d47827010ce02902834f10118daeb381bdfa65bcc5adf778b65f
+  data.tar.gz: 65b9fac4ebee7589010039dbba9cbd8b204a94987443ec2c68e36a7509f39685
 SHA512:
-  metadata.gz: a40bedc72f0282d80f771d6808a924ece95447092a4bb9f48108977387e73b80cdc5ebd5f6c3af1aef938a2b952fde5a1ee8d320635e808b8f06a47f90cba41d
-  data.tar.gz: a689cc2b3415bb73af3940809fca8795b0de6cbb9f9768a62a347ea7016d7e18c005b5a33a0233ee61dad9f37dc324a20fb0a8f88068ca196fe291bc02d704a9
+  metadata.gz: e9ecec0384d8c4d931cace14b32ddc9e2b5161ed6ef9f61af27fb89d9993643db790f98966f8ddea419008bd03588cbc6aaf4dfee25b404c5a392deae1b3e78e
+  data.tar.gz: 79b733073abbf0ee6dc8da9835c86e44be67f1f298a4e09f3d6380772ff51713cd63c121daf2bb0ae0751ab9ffb3954003eb24ecffde1a7e89ed3c542cead994

data/lib/mods/date.rb

@@ -11,7 +11,7 @@ module Mods
     # @param [Nokogiri::XML::Element] xml A date-flavored MODS field from the XML
     # @return [Mods::Date]
     def self.from_element(xml)
-      case xml.attr(:encoding)
+      case xml.attr(:encoding)&.downcase
       when 'w3cdtf'
         Mods::Date::W3cdtfFormat.new(xml)
       when 'iso8601'
@@ -23,12 +23,13 @@ module Mods
       # when 'temper'
       #   Mods::Date::TemperFormat.new(xml)
       else
-        date_class = Mods::Date if xml.text =~ /\p{Hebrew}/
+        date_class = UnparseableDate if xml.text =~ /\p{Hebrew}/ || xml.text =~ /^-/
         date_class ||= [
           MMDDYYYYFormat,
           MMDDYYFormat,
           YearRangeFormat,
           DecadeAsYearDashFormat,
+          DecadeStringFormat,
           EmbeddedBCYearFormat,
           EmbeddedYearFormat,
           EmbeddedThreeDigitYearFormat,
@@ -49,7 +50,7 @@ module Mods
     # Strict ISO8601-encoded date parser
     class Iso8601Format < Date
       def self.parse_date(text)
-        @date = ::Date.parse(normalize_to_edtf(text))
+        ::Date.parse(normalize_to_edtf(text))
       end
     end
 
@@ -62,7 +63,16 @@ module Mods
       attr_reader :date
 
       def self.normalize_to_edtf(text)
-        text
+        return '0000' if text.strip == '0'
+
+        case text
+        when /^\d{1,3}$/
+          text.rjust(4, "0") if text =~ /^\d{1,3}$/
+        when /^-\d{1,3}$/
+          "-#{text.sub(/^-/, '').rjust(4, "0")}"
+        else
+          text
+        end
       end
     end
 
@@ -100,6 +110,12 @@ module Mods
       end
     end
 
+    class UnparseableDate < ExtractorDateFormat
+      def self.parse_date(text)
+        nil
+      end
+    end
+
     # Full text extractor for MM/DD/YYYY and MM/DD/YYY-formatted dates
     class MMDDYYYYFormat < ExtractorDateFormat
       REGEX = /(?<month>\d{1,2})\/(?<day>\d{1,2})\/(?<year>\d{3,4})/
@@ -201,7 +217,17 @@ module Mods
 
     # Full-text extractor for data formatted as YYY-
     class DecadeAsYearDashFormat < ExtractorDateFormat
-      REGEX = /(?<!\d)(?<year>\d{3})[-_x?](?!\d)/
+      REGEX = /(?<!\d)(?<year>\d{3})[-_xu?](?!\d)/
+
+      def self.normalize_to_edtf(text)
+        matches = text.match(REGEX)
+        "#{matches[:year]}X"
+      end
+    end
+
+    # Full-text extractor for data formatted as YYY0s
+    class DecadeStringFormat < ExtractorDateFormat
+      REGEX = /(?<!\d)(?<year>\d{3})0s(?!\d)/
 
       def self.normalize_to_edtf(text)
         matches = text.match(REGEX)
@@ -221,42 +247,42 @@ module Mods
 
     # Full-text extractor that tries hard to pick any year present in the data
     class EmbeddedYearFormat < ExtractorDateFormat
-      REGEX = /(?<prefix>-)?(?<!\d)(?<year>\d{4})(?!\d)/
+      REGEX = /(?<!\d)(?<year>\d{4})(?!\d)/
 
       def self.normalize_to_edtf(text)
         matches = text.match(REGEX)
-        "#{matches[:prefix]}#{matches[:year].rjust(4, "0")}"
+        "#{matches[:year].rjust(4, "0")}"
       end
     end
 
     # Full-text extractor that tries hard to pick any year present in the data
     class EmbeddedThreeDigitYearFormat < ExtractorDateFormat
-      REGEX = /(?<prefix>-)?(?<!\d)(?<year>\d{3})(?!\d)(?!\d)/
+      REGEX = /(?<!\d)(?<year>\d{3})(?!\d)(?!\d)/
 
       def self.normalize_to_edtf(text)
         matches = text.match(REGEX)
-        "#{matches[:prefix]}#{matches[:year].rjust(4, "0")}"
+        "#{matches[:year].rjust(4, "0")}"
       end
     end
 
     # Full-text extractor that tries hard to pick any year present in the data
     class OneOrTwoDigitYearFormat < ExtractorDateFormat
-      REGEX = /^(?<prefix>-)?(?<year>\d{1,2})$/
+      REGEX = /^(?<year>\d{1,2})$/
 
       def self.normalize_to_edtf(text)
         matches = text.match(REGEX)
-        "#{matches[:prefix]}#{matches[:year].rjust(4, "0")}"
+        "#{matches[:year].rjust(4, "0")}"
       end
     end
 
     # Full-text extractor that tries hard to pick any year present in the data
     class EmbeddedYearWithBracketsFormat < ExtractorDateFormat
       # [YYY]Y Y[YYY] [YY]YY Y[YY]Y YY[YY] YYY[Y] YY[Y]Y Y[Y]YY [Y]YYY
-      REGEX = /(?<prefix>-)?(?<year>[\d\[\]]{6})(?!\d)/
+      REGEX = /(?<year>[\d\[\]]{6})(?!\d)/
 
       def self.normalize_to_edtf(text)
         matches = text.match(REGEX)
-        "#{matches[:prefix]}#{matches[:year].gsub('[', '').gsub(']', '')}"
+        "#{matches[:year].gsub('[', '').gsub(']', '')}"
       end
     end
 
@@ -327,7 +353,9 @@ module Mods
     #
     # @return [String]
     def type
-      xml.attr(:type)
+      return if xml.attr(:type)&.empty?
+
+      xml.attr(:type)&.downcase
     end
 
     ##
@@ -335,7 +363,9 @@ module Mods
     #
     # @return [String]
     def encoding
-      xml.attr(:encoding)
+      return if xml.attr(:encoding)&.empty?
+
+      xml.attr(:encoding)&.downcase
     end
 
     ##
@@ -359,7 +389,9 @@ module Mods
     #
     # @return [String]
     def point
-      xml.attr(:point)
+      return if xml.attr(:point)&.empty?
+
+      xml.attr(:point)&.downcase
     end
 
     ##
@@ -391,7 +423,7 @@ module Mods
     #
     # @return [String]
     def qualifier
-      xml.attr(:qualifier)
+      xml.attr(:qualifier)&.downcase
     end
 
     ##
@@ -419,6 +451,8 @@ module Mods
     end
 
     def precision
+      return :unknown unless date_range || date
+
       if date_range.is_a? EDTF::Century
         :century
       elsif date_range.is_a? EDTF::Decade

data/lib/mods/nom_terminology.rb

@@ -258,7 +258,7 @@ module Mods
           n.publisher :path => 'm:publisher'
           Mods::OriginInfo::DATE_ELEMENTS.each { |date_el|
             n.send date_el, :path => "m:#{date_el}" do |d|
-              d.as_object :path => '.', :accessor => lambda { |a| Mods::Date.from_element(a) }
+              d.as_object :path => '.', :single => true, :accessor => lambda { |a| Mods::Date.from_element(a) }
 
               with_attributes(d, Mods::DATE_ATTRIBS)
 

data/lib/mods/version.rb

@@ -1,4 +1,4 @@
 module Mods
   # this is the Ruby Gem version
-  VERSION = '3.0.0.alpha2'.freeze
+  VERSION = '3.0.0'.freeze
 end

data/spec/lib/date_spec.rb

@@ -209,6 +209,9 @@ RSpec.describe Mods::Date do
       '1900-uu-uu' => Date.parse('1900-01-01')..Date.parse('1900-12-31'),
       '1900-uu-15' => Date.parse('1900-01-15')..Date.parse('1900-12-15'),
       '1900-06-uu' => Date.parse('1900-06-01')..Date.parse('1900-06-30'),
+      '-250' => Date.parse('-250-01-01')..Date.parse('-250-12-31'), # EDTF requires a 4 digit year, but what can you do.
+      '63' => Date.parse('0063-01-01')..Date.parse('0063-12-31'),
+      '125' => Date.parse('125-01-01')..Date.parse('125-12-31'),
     }.each do |data, expected|
       describe "with #{data}" do
         let(:date_element) { "<dateCreated encoding=\"edtf\">#{data}</dateCreated>" }
@@ -303,7 +306,8 @@ RSpec.describe Mods::Date do
     {
       'Minguo 19 [1930]' => Date.parse('1930-01-01')..Date.parse('1930-12-31'),
       '1745 mag. 14' => Date.parse('1745-01-01')..Date.parse('1745-12-31'),
-      '-745' => Date.parse('-745-01-01')..Date.parse('-745-12-31'),
+      '-745' => '', # too ambiguious to even attempt.
+      '-1999' => '', # too ambiguious to even attempt.
       '[1923]' => Date.parse('1923-01-01')..Date.parse('1923-12-31'),
       '1532.' => Date.parse('1532-01-01')..Date.parse('1532-12-31'),
       '[ca 1834]' => Date.parse('1834-01-01')..Date.parse('1834-12-31'),
@@ -319,8 +323,11 @@ RSpec.describe Mods::Date do
       '193-' => Date.parse('1930-01-01')..Date.parse('1939-12-31'),
       '196_' => Date.parse('1960-01-01')..Date.parse('1969-12-31'),
       '196x' => Date.parse('1960-01-01')..Date.parse('1969-12-31'),
+      '196u' => Date.parse('1960-01-01')..Date.parse('1969-12-31'),
+      '1960s' => Date.parse('1960-01-01')..Date.parse('1969-12-31'),
       '186?' => Date.parse('1860-01-01')..Date.parse('1869-12-31'),
       '1700?' => Date.parse('1700-01-01')..Date.parse('1700-12-31'),
+      'early 1730s' => Date.parse('1730-01-01')..Date.parse('1739-12-31'),
       '[1670-1684]' => Date.parse('1670-01-01')..Date.parse('1684-12-31'),
       '[18]74' => Date.parse('1874-01-01')..Date.parse('1874-12-31'),
       '250 B.C.' => Date.parse('-0249-01-01')..Date.parse('-249-12-31'),

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: mods
 version: !ruby/object:Gem::Version
-  version: 3.0.0.alpha2
+  version: 3.0.0
 platform: ruby
 authors:
 - Naomi Dushay
@@ -9,7 +9,7 @@ authors:
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-01-27 00:00:00.000000000 Z
+date: 2022-02-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -218,9 +218,9 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ">"
+  - - ">="
     - !ruby/object:Gem::Version
-      version: 1.3.1
+      version: '0'
 requirements: []
 rubygems_version: 3.2.32
 signing_key: