parse_date 0.3.3 → 0.3.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 42b7b5f7405ee4a79e7eece920b10fe2852a03c996b32da8c2b8edc58a6e3ede
-  data.tar.gz: eeba668fd1ee75994f743a010dbe37af4ffdcb8ca0c4143ca329c811739c4b28
+  metadata.gz: c818fd11526f5dbddb22a29801089a6ff9946d0eb23a208af5316a34c2201836
+  data.tar.gz: bb5b1f93425fe42659b144554b8a21dc95615a7b7e35ac1a361d7696869367f1
 SHA512:
-  metadata.gz: a22f1e23899efc15e5fd13ab1aac724cddc39ddf9986d0295e2f4360b639483da116ef4540bb35fc0c03cf201c0d29f4e0ac6729339140663ff7622ed8a7a623
-  data.tar.gz: df518ca44ed63ea2925976fe49a11e28cd4ecc40b66752c7db11934d4bba6fe7ade795aca93402cb0f556816cf076e40006e08294162b8f01c527377a4603e51
+  metadata.gz: 5b8f75b733e86e4f6f6178279a7959a6693443ddd2bb0c1315957b3d197993c2bd19790516f5170b6458f0ae4209ce3bf8b9e83fdb9b30b373e0bf2896aed93d
+  data.tar.gz: a3adec7a2979f591386a9674f7d096f70dc9146af27ed3b734404f195fa35c820b5218a850365ac386a6d478cc71bf7b7f1d78ffcdf523d6d4547b3d52caeff3

data/.rubocop_todo.yml

@@ -1,6 +1,6 @@
 # This configuration was generated by
 # `rubocop --auto-gen-config`
-# on 2019-10-28 09:19:23 -0700 using RuboCop version 0.74.0.
+# on 2019-10-28 20:35:55 -0700 using RuboCop version 0.74.0.
 # The point is for the user to remove these configuration records
 # one by one as the offenses are removed from the code base.
 # Note that changes in the inspected code, or installation of new
@@ -17,7 +17,7 @@ Metrics/CyclomaticComplexity:
 # Offense count: 1
 # Configuration parameters: CountComments.
 Metrics/ModuleLength:
-  Max: 178
+  Max: 183
 
 # Offense count: 4
 Metrics/PerceivedComplexity:

data/README.md

@@ -52,13 +52,18 @@ ParseDate.parse_range('17--?-18--?')                    # (1700..1899).to_a
 ParseDate.parse_range('1835 or 1836')                   # [1835, 1836]
 ParseDate.parse_range('17-- or 18--?')                  # (1700..1899).to_a
 ParseDate.parse_range('-2 or 1?')                       # (-2..1).to_a
+ParseDate.parse_range('1500? to 1582')                  # (1500..1582).to_a
 ParseDate.parse_range('17th or 18th century?')          # (1600..1799).to_a
 ParseDate.parse_range('ca. 5th–6th century A.D.')       # (400..599).to_a
 ParseDate.parse_range('ca. 9th–8th century B.C.')       # (-999..-800).to_a
 ParseDate.parse_range('ca. 13th–12th century B.C.')     # (-1399..-1200).to_a
 ParseDate.parse_range('5th century B.C.')               # (-599..-500).to_a
 ParseDate.parse_range('502-504')                        # [502, 503, 504]
+ParseDate.parse_range('950-60')                         # (950..960).to_a
+ParseDate.parse_range('-0150 - -0100')                  # (-150..-100).to_a
 ParseDate.parse_range('-2100 - -2000')                  # (-2100..-2000).to_a
+ParseDate.parse_range('1230—1239 CE')                   # (1230..1239).to_a (alternate hyphen char)
+ParseDate.parse_range('996–1021 CE')                    # (996..1021).to_a (diff alternate hyphen char)
 ParseDate.parse_range('1975 - 1905')                    # last year > first year, raises error
 ParseDate.parse_range('-100 - -150')                    # last year > first year, raises error
 ParseDate.parse_range('1975 or 1905')                   # last year > first year, raises error
@@ -82,7 +87,10 @@ ParseDate.earliest_year('17uu')                         # 1700
 ParseDate.earliest_year('between 1694 and 1799')        # 1694
 ParseDate.earliest_year('between 1 and 5')              # 1
 ParseDate.earliest_year('between 300 and 150 B.C.')     # -300
+ParseDate.earliest_year('1500? to 1582')                # 1500
 ParseDate.earliest_year('1496-1499')                    # 1496
+ParseDate.earliest_year('1230—1239 CE')                 # 1230 (alternate hyphen char)
+ParseDate.earliest_year('996–1021 CE')                  # 996 (diff alternate hyphen char)
 ParseDate.earliest_year('1750?-1867')                   # 1750
 ParseDate.earliest_year('17--?-18--?')                  # 1700
 ParseDate.earliest_year('1835 or 1836')                 # 1835
@@ -93,6 +101,8 @@ ParseDate.earliest_year('ca. 9th–8th century B.C.')     # -999
 ParseDate.earliest_year('ca. 13th–12th century B.C.')   # -1399
 ParseDate.earliest_year('5th century B.C.')             # -599
 ParseDate.earliest_year('502-504')                      # 502
+ParseDate.earliest_year('950-60')                       # 950
+ParseDate.earliest_year('-0150 - -0100')                # -150
 ParseDate.earliest_year('-2100 - -2000')                # -2100
 
 ParseDate.latest_year('20000222')                       # 2000
@@ -105,7 +115,10 @@ ParseDate.latest_year('17uu')                           # 1799
 ParseDate.latest_year('between 1694 and 1799')          # 1799
 ParseDate.latest_year('between 1 and 5')                # 5
 ParseDate.latest_year('between 300 and 150 B.C.')       # -150
+ParseDate.latest_year('1500? to 1582')                  # 1582
 ParseDate.latest_year('1496-1499')                      # 1499
+ParseDate.latest_year('1230—1239 CE')                   # 1239 (alternate hyphen char)
+ParseDate.latest_year('996–1021 CE')                    # 1021 (diff alternate hyphen char)
 ParseDate.latest_year('1750?-1867')                     # 1867
 ParseDate.latest_year('17--?-18--?')                    # 1899
 ParseDate.latest_year('1757-58')                        # 1758
@@ -119,6 +132,8 @@ ParseDate.latest_year('ca. 13th–12th century B.C.')     # -1200
 ParseDate.latest_year('5th century B.C.')               # -500
 ParseDate.latest_year('-5 - 3')                         # 3
 ParseDate.latest_year('502-504')                        # 504
+ParseDate.latest_year('950-60')                         # 960
+ParseDate.latest_year('-0150 - -0100')                  # -100
 ParseDate.latest_year('-2100 - -2000')                  # -2000
 
 ParseDate.range_array('1993', '1995')                   # [1993, 1994, 1995]

data/lib/parse_date/int_from_string.rb

@@ -25,6 +25,7 @@ class ParseDate
       return ParseDate.send(:year_int_for_bc, date_str) if date_str.match(YEAR_BC_REGEX)
 
       result ||= ParseDate.send(:between_earliest_year, date_str)
+      result ||= ParseDate.send(:hyphen_4digit_earliest_year, date_str)
       result ||= ParseDate.send(:negative_first_four_digits, date_str)
       result ||= ParseDate.send(:first_four_digits, date_str)
       result ||= ParseDate.send(:year_from_mm_dd_yy, date_str)
@@ -97,7 +98,13 @@ class ParseDate
       date_str.delete('[]') if date_str.match(BRACKETS_BETWEEN_DIGITS_REGEX)
     end
 
-    YYYY_HYPHEN_YYYY_REGEX = Regexp.new(/(?<first>\d{4})\??\s*[-—]\s*(?<last>\d{4})\??/m)
+    YYYY_HYPHEN_YYYY_REGEX = Regexp.new(/(?<first>\d{3,4})\??\s*(-|—|–|to)\s*(?<last>\d{4})\??/m)
+
+    # Integer value for latest year if we have "yyyy-yyyy" pattern
+    # @return [Integer, nil] yyyy if date_str matches pattern; nil otherwise
+    def hyphen_4digit_earliest_year(date_str)
+      Regexp.last_match(:first).to_i if date_str.match(YYYY_HYPHEN_YYYY_REGEX)
+    end
 
     # Integer value for latest year if we have "yyyy-yyyy" pattern
     # @return [Integer, nil] yyyy if date_str matches pattern; nil otherwise
@@ -105,7 +112,7 @@ class ParseDate
       Regexp.last_match(:last).to_i if date_str.match(YYYY_HYPHEN_YYYY_REGEX)
     end
 
-    YYYY_HYPHEN_YY_REGEX = Regexp.new(/(?<first>\d{4})\??\s*[-—]\s*(?<last>\d{2})\??([^-0-9].*)?$/)
+    YYYY_HYPHEN_YY_REGEX = Regexp.new(/(?<first>\d{3,4})\??\s*(-|—|–|to)\s*(?<last>\d{2})\??([^-0-9].*)?$/)
 
     # Integer value for latest year if we have "yyyy-yy" pattern
     # @return [Integer, nil] yyyy if date_str matches pattern; nil otherwise
@@ -114,14 +121,14 @@ class ParseDate
       return unless matches
 
       first = Regexp.last_match(:first)
-      century = first[0, 2]
+      century = first[0..-3] # whatever is before the last 2 digits
       last = "#{century}#{Regexp.last_match(:last)}"
       last.to_i if ParseDate.year_range_valid?(first.to_i, last.to_i)
     end
 
     YYUU = '\\d{1,2}[u\\-]{2}'
     YYuu_HYPHEN_YYuu_REGEX =
-      Regexp.new("(?<first>#{YYUU})\\??\\s*[-—]\\s*(?<last>#{YYUU})\\??([^u\\-]|$)??", REGEX_OPTS)
+      Regexp.new("(?<first>#{YYUU})\\??\\s*(-|—|–|to)\\s*(?<last>#{YYUU})\\??([^u\\-]|$)??", REGEX_OPTS)
 
     # Integer value for latest year if we have "yyuu-yyuu" pattern
     # @return [Integer, nil] yyyy if date_str matches pattern; nil otherwise
@@ -139,7 +146,8 @@ class ParseDate
     end
 
     # NOTE: some actual data seemed to have a diff hyphen char. (slightly longer)
-    YY_YY_CENTURY_REGEX = Regexp.new(/(?<first>\d{1,2})[a-z]{2}?\s*(-|–|or)\s*(?<last>\d{1,2})[a-z]{2}?\s+centur.*/im)
+    YY_YY_CENTURY_REGEX =
+      Regexp.new(/(?<first>\d{1,2})[a-z]{2}?\s*(-|–|–|or|to)\s*(?<last>\d{1,2})[a-z]{2}?\s+centur.*/im)
 
     # Integer value for latest year if we have nth-nth century pattern
     # @return [Integer, nil] yy99 if date_str matches pattern; nil otherwise
@@ -182,7 +190,7 @@ class ParseDate
     # looks for -yyyy after hyphen and returns if found
     # @return [String, nil] negative 4 digit year (e.g. -1865) if date_str has -yyyy - -yyyy, nil otherwise
     def negative_4digits_after_hyphen(date_str)
-      Regexp.last_match(1) if date_str.match(/\-\d{4}\s*\-\s*(\-\d{4})/)
+      Regexp.last_match(1) if date_str.match(/\-\d{4}\s*(?:-|–|–|or|to)\s*(\-\d{4})/)
     end
 
     # looks for 4 consecutive digits in date_str and returns first occurrence if found
@@ -304,7 +312,7 @@ class ParseDate
     end
 
     FIRST_LAST_EARLY_NUMERIC_REGEX =
-      Regexp.new(/^(?<first>\-?\d{1,3})\??\s*(-|–|or)\s*(?<last>\-?\d{1,4})\??([^\du\-\[]|$)/im)
+      Regexp.new(/^(?<first>\-?\d{1,3})\??\s*(-|–|–|or|to)\s*(?<last>\-?\d{1,4})\??([^\du\-\[]|$)/im)
 
     # Integer value for latest year if we have early numeric year range or single early numeric year
     # @return [Integer, nil] year if date_str matches pattern; nil otherwise

data/lib/parse_date/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 class ParseDate
-  VERSION = '0.3.3'
+  VERSION = '0.3.4'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: parse_date
 version: !ruby/object:Gem::Version
-  version: 0.3.3
+  version: 0.3.4
 platform: ruby
 authors:
 - Naomi Dushay
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2019-10-28 00:00:00.000000000 Z
+date: 2019-10-29 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: zeitwerk