mods 2.4.1 → 3.0.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4ee9e9f682562bec8eeee3698cb1dc20fedee759172e0f6ef84a39ab01679550
-  data.tar.gz: 67a3d7a70e895de43831a40b30195bd8fe2ffb27315c74ec28e805416169512d
+  metadata.gz: 566f8acde01c8696c1a183588b663934dc50d500999fcc83ba67516620e57ae7
+  data.tar.gz: 87aeca4a314870bcace3fc1b3fe2260b67134becc4ca2a7a22ac3bb5edf73be2
 SHA512:
-  metadata.gz: 90ba88b4a9354acfa9714711cd20a932473fafc62445ee40ea33725b4a03927b5f499cff78fab004a7ed307741424ef9aa19dd65e4b821f416e31a95bb116f83
-  data.tar.gz: 6e343ecf819c236774468511abb31be4a80aa2cfa8725231736f352c34c51cae0e8fe8fd9cd93d527c97fda52160fe1fe65ae0d204fbcff8fab54e5bad4379ec
+  metadata.gz: 0b5d979ab8b4fbcb7e5dce0b5d8117e76f1c95a284cf531fd659a1e78d74af5571af309c9ab3e77b6d105f9e9d3427ba0fa88d82f4feeb5d54c18bc306079332
+  data.tar.gz: 0bea0554c539de56fcfee76380f0478133d9288fba63b1ff6c6a922d3877ab04e12070590c2a10339599cf2615bc75433d23fb546e75ad4f8627b75dd563efb7

data/.github/workflows/ruby.yml

@@ -0,0 +1,24 @@
+name: CI
+
+on:
+  push:
+    branches: [ master ]
+  pull_request:
+    branches: [ master ]
+
+jobs:
+  tests:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        ruby: [jruby-9.3.2.0, 2.7, '3.0', 3.1]
+    steps:
+    - uses: actions/checkout@v2
+    - name: Set up Ruby
+      uses: ruby/setup-ruby@v1
+      with:
+        ruby-version: ${{ matrix.ruby }}
+    - name: Install dependencies
+      run: bundle install
+    - name: Run tests
+      run: bundle exec rake

data/.gitignore

@@ -22,3 +22,4 @@ tmtags
 .idea/*
 to_delete
 .byebug_history
+/spec/examples.txt

data/README.md

@@ -1,7 +1,6 @@
 # Mods
 
 [<img src="https://secure.travis-ci.org/sul-dlss/mods.png?branch=master" alt="Build Status"/>](http://travis-ci.org/sul-dlss/mods) [![Code Climate Test Coverage](https://codeclimate.com/github/sul-dlss/mods/badges/coverage.svg)](https://codeclimate.com/github/sul-dlss/mods/coverage) [<img
-src="https://gemnasium.com/sul-dlss/mods.png" alt="Dependency Status"/>](https://gemnasium.com/sul-dlss/mods) [<img
 src="https://badge.fury.io/rb/mods.svg" alt="Gem Version"/>](http://badge.fury.io/rb/mods)
 
 A Gem to parse MODS (Metadata Object Description Schema) records.  More information about MODS can be found at

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,
@@ -49,12 +50,15 @@ 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
 
     # Less strict W3CDTF-encoded date parser
     class W3cdtfFormat < Date
+      def self.normalize_to_edtf(text)
+        super.gsub('-00', '')
+      end
     end
 
     # Strict EDTF parser
@@ -62,7 +66,16 @@ 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 +113,12 @@ 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 +220,17 @@ 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 +250,42 @@ 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 +356,9 @@ module Mods
     #
     # @return [String]
     def type
-      xml.attr(:type)
+      return if xml.attr(:type)&.empty?
+
+      xml.attr(:type)&.downcase
     end
 
     ##
@@ -335,7 +366,9 @@ module Mods
     #
     # @return [String]
     def encoding
-      xml.attr(:encoding)
+      return if xml.attr(:encoding)&.empty?
+
+      xml.attr(:encoding)&.downcase
     end
 
     ##
@@ -359,7 +392,9 @@ module Mods
     #
     # @return [String]
     def point
-      xml.attr(:point)
+      return if xml.attr(:point)&.empty?
+
+      xml.attr(:point)&.downcase
     end
 
     ##
@@ -391,7 +426,7 @@ module Mods
     #
     # @return [String]
     def qualifier
-      xml.attr(:qualifier)
+      xml.attr(:qualifier)&.downcase
     end
 
     ##
@@ -419,6 +454,8 @@ 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/marc_country_codes.rb

@@ -1,8 +1,8 @@
-# encoding: UTF-8
-
 # Represents the Marc Country Codes mapped to names, from http://www.loc.gov/marc/countries/countries_code.html 2013-01-03
-#key - Marc Country code
-#value - Marc Country term
+# key - Marc Country code
+# value - Marc Country term
+# frozen_string_literal: true
+
 MARC_COUNTRY = {
   'aa' => "Albania",
   'abc' => "Alberta",
@@ -11,7 +11,6 @@ MARC_COUNTRY = {
   'ae' => "Algeria",
   'af' => "Afghanistan",
   'ag' => "Argentina",
-  #'ai' => "Anguilla", # discontinued
   'ai' => "Armenia (Republic)",
   'air' => "Armenian S.S.R.", # discontinued
   'aj' => "Azerbaijan",
@@ -40,7 +39,7 @@ MARC_COUNTRY = {
   'bi' => "British Indian Ocean Territory",
   'bl' => "Brazil",
   'bm' => "Bermuda Islands",
-  'bn' => "Bosnia and Hercegovina",
+  'bn' => "Bosnia and Herzegovina",
   'bo' => "Bolivia",
   'bp' => "Solomon Islands",
   'br' => "Burma",
@@ -74,7 +73,7 @@ MARC_COUNTRY = {
   'cs' => "Czechoslovakia", # discontinued
   'ctu' => "Connecticut",
   'cu' => "Cuba",
-  'cv' => "Cape Verde",
+  'cv' => "Cabo Verde",
   'cw' => "Cook Islands",
   'cx' => "Central African Republic",
   'cy' => "Cyprus",
@@ -109,6 +108,7 @@ MARC_COUNTRY = {
   'gb' => "Kiribati",
   'gd' => "Grenada",
   'ge' => "Germany (East)", # discontinued
+  'gg' => "Guernsey",
   'gh' => "Ghana",
   'gi' => "Gibraltar",
   'gl' => "Greenland",
@@ -137,6 +137,7 @@ MARC_COUNTRY = {
   'ie' => "Ireland",
   'ii' => "India",
   'ilu' => "Illinois",
+  'im' => "Isle of Man",
   'inu' => "Indiana",
   'io' => "Indonesia",
   'iq' => "Iraq",
@@ -148,6 +149,7 @@ MARC_COUNTRY = {
   'iw' => "Israel-Jordan Demilitarized Zones", # discontinued
   'iy' => "Iraq-Saudi Arabia Neutral Zone",
   'ja' => "Japan",
+  'je' => "Jersey",
   'ji' => "Johnston Atoll",
   'jm' => "Jamaica",
   'jn' => "Jan Mayen", # discontinued
@@ -284,7 +286,7 @@ MARC_COUNTRY = {
   'snc' => "Saskatchewan",
   'so' => "Somalia",
   'sp' => "Spain",
-  'sq' => "Swaziland",
+  'sq' => "Eswatini",
   'sr' => "Surinam",
   'ss' => "Western Sahara",
   'st' => "Saint-Martin",
@@ -365,7 +367,7 @@ MARC_COUNTRY = {
   'xk' => "Saint Lucia",
   'xl' => "Saint Pierre and Miquelon",
   'xm' => "Saint Vincent and the Grenadines",
-  'xn' => "Macedonia",
+  'xn' => "North Macedonia",
   'xna' => "New South Wales",
   'xo' => "Slovakia",
   'xoa' => "Northern Territory",
@@ -384,4 +386,4 @@ MARC_COUNTRY = {
   'ys' => "Yemen (People's Democratic Republic)", # discontinued
   'yu' => "Serbia and Montenegro", # discontinued
   'za' => "Zambia"
-}
+}.freeze