Merge pull request #32 from sul-dlss/earlier-ranges

handle earlier dates correctly (negative 4 digits;  ranges of 3 digits)

.rubocop.yml

@@ -17,7 +17,7 @@ Metrics/LineLength:
   Max: 120
 
 Metrics/MethodLength:
-  Max: 22
+  Max: 25
 
 Style/NumericLiterals:
   Enabled: false

.rubocop_todo.yml

@@ -1,14 +1,14 @@
 # This configuration was generated by
 # `rubocop --auto-gen-config`
-# on 2019-10-22 09:04:50 -0700 using RuboCop version 0.74.0.
+# on 2019-10-24 13:45:33 -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
 # versions of RuboCop, may require this file to be generated again.
 
 # Offense count: 5
 Metrics/AbcSize:
-  Max: 37
+  Max: 39
 
 # Offense count: 4
 Metrics/CyclomaticComplexity:
@@ -17,7 +17,7 @@ Metrics/CyclomaticComplexity:
 # Offense count: 1
 # Configuration parameters: CountComments.
 Metrics/ModuleLength:
-  Max: 168
+  Max: 176
 
 # Offense count: 4
 Metrics/PerceivedComplexity:

README.md

@@ -57,6 +57,8 @@ 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('-2100 - -2000')                  # (-2100..-2000).to_a
 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
@@ -90,6 +92,8 @@ ParseDate.earliest_year('ca. 5th–6th century A.D.')     # 400
 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('-2100 - -2000')                # -2100
 
 ParseDate.latest_year('20000222')                       # 2000
 ParseDate.latest_year('195-')                           # 1959
@@ -113,7 +117,9 @@ ParseDate.latest_year('ca. 5th–6th century A.D.')       # 599
 ParseDate.latest_year('ca. 9th–8th century B.C.')       # -800
 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('-5 - 3')                         # 3
+ParseDate.latest_year('502-504')                        # 504
+ParseDate.latest_year('-2100 - -2000')                  # -2000
 
 ParseDate.range_array('1993', '1995')                   # [1993, 1994, 1995]
 ParseDate.range_array(1993, 1995)                       # [1993, 1994, 1995]

lib/parse_date/int_from_string.rb

@@ -25,6 +25,7 @@ def self.earliest_year(date_str)
       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(:negative_first_four_digits, date_str)
       result ||= ParseDate.send(:first_four_digits, date_str)
       result ||= ParseDate.send(:year_from_mm_dd_yy, date_str)
       result ||= ParseDate.send(:first_year_for_decade, date_str) # 198x or 201x
@@ -61,6 +62,7 @@ def self.latest_year(date_str)
       result ||= ParseDate.send(:hyphen_2digit_latest_year, date_str)
       result ||= ParseDate.send(:yyuu_after_hyphen, date_str)
       result ||= ParseDate.send(:year_after_or, date_str)
+      result ||= ParseDate.send(:negative_4digits_after_hyphen, date_str)
       result ||= ParseDate.send(:first_four_digits, date_str)
       result ||= ParseDate.send(:year_from_mm_dd_yy, date_str)
       result ||= ParseDate.send(:last_year_for_decade, date_str) # 198x or 201x
@@ -75,12 +77,12 @@ def self.latest_year(date_str)
       result.to_i if result && year_int_valid?(result.to_i)
     end
 
-    # true if the year is between -999 and (current year + 1), inclusive
+    # true if the year is between -9999 and (current year + 1), inclusive
     # @return [Boolean] true if the year is between -999 and (current year + 1); false otherwise
     def self.year_int_valid?(year)
       return false unless year.is_a? Integer
 
-      (-1000 < year.to_i) && (year < Date.today.year + 2)
+      (-10000 < year.to_i) && (year < Date.today.year + 2)
     end
 
     protected
@@ -169,6 +171,18 @@ def last_year_mult_centuries_bc(date_str)
       nth * -100
     end
 
+    # looks for -yyyy at beginning of date_str and returns if found
+    # @return [String, nil] negative 4 digit year (e.g. -1865) if date_str has -yyyy, nil otherwise
+    def negative_first_four_digits(date_str)
+      Regexp.last_match(1) if date_str.match(/^(\-\d{4})/)
+    end
+
+    # 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})/)
+    end
+
     # looks for 4 consecutive digits in date_str and returns first occurrence if found
     # @return [String, nil] 4 digit year (e.g. 1865, 0950) if date_str has yyyy, nil otherwise
     def first_four_digits(date_str)
@@ -195,7 +209,7 @@ def year_from_mm_dd_yy(date_str)
       nil # explicitly want nil if date won't parse
     end
 
-    DECADE_4CHAR_REGEX = Regexp.new('(^|\D)\d{3}[u\-?x]', REGEX_OPTS)
+    DECADE_4CHAR_REGEX = Regexp.new('(^|\D)\d{3}[u\-?x]($|\D)', REGEX_OPTS)
 
     # first year of decade (as String) if we have:  yyyu, yyy-, yyy? or yyyx pattern
     #   note that these are the only decade patterns found in our actual date strings in MODS records
@@ -279,7 +293,7 @@ def between_bc_latest_year(date_str)
       "-#{Regexp.last_match(:last)}".to_i if date_str.match(BETWEEN_Yn_AND_Yn_BC_REGEX)
     end
 
-    EARLY_NUMERIC_REGEX = Regexp.new('^\-?\d{1,3}([^\du\-\[]|$)', REGEX_OPTS)
+    EARLY_NUMERIC_REGEX = Regexp.new('^\-?\d{1,3}([^\du\[]|$)', REGEX_OPTS)
 
     # year if date_str contains yyy, yy, y, -y, -yy, -yyy, -yyyy
     # @return [Integer, nil] year if date_str matches pattern; nil otherwise

spec/parse_date/int_from_string_spec.rb

@@ -420,6 +420,8 @@
       'ca. 5th–6th century A.D.' => 400,
       'ca. 9th–8th century B.C.' => -999,
       'ca. 13th–12th century B.C.' => -1399,
+      '502-504' => 502,
+      '-2100 - -2000' => -2100,
       # '2nd millennium B.C. or ca. 8th century B.C.' => -899, # millennium not yet implemented
     }.each do |example, expected|
       it "#{expected} for #{example}" do
@@ -534,6 +536,8 @@
       'ca. 5th–6th century A.D.' => 599,
       'ca. 9th–8th century B.C.' => -800,
       'ca. 13th–12th century B.C.' => -1200,
+      '502-504' => 504,
+      '-2100 - -2000' => -2000,
       # '2nd millennium B.C. or ca. 8th century B.C.' => -800, # millennium not yet implemented
     }.each do |example, expected|
       it "#{expected} for #{example}" do
@@ -644,7 +648,7 @@
 
   describe '.year_int_valid?' do
     { # example int as key, expected result as value
-      -1666 => false,
+      -1666 => true,
       -999 => true,
       -35 => true,
       -3 => true,

spec/parse_date_spec.rb

@@ -44,6 +44,8 @@
         'ca. 9th–8th century B.C.' => (-999..-800).to_a,
         'ca. 13th–12th century B.C.' => (-1399..-1200).to_a,
         '5th century B.C.' => (-599..-500).to_a,
+        '502-504' => [502, 503, 504],
+        '-2100 - -2000' => (-2100..-2000).to_a,
       }.each do |example, expected|
         it "#{example} returns array from earliest (#{expected.first}) to latest (#{expected.last})" do
           expect(ParseDate.parse_range(example)).to eq expected