timetwister 0.2.4 → 0.2.5

checksums.yaml

@@ -1,15 +1,15 @@
 ---
 !binary "U0hBMQ==":
   metadata.gz: !binary |-
-    YzlhYjlmZjMwMmI3YjBhNjBiYTYxYTZjMzk0OGQwZGRjYjUwODNkZg==
+    NWIyYWYyOTcwYzY4NjExNWI0NmE1MTU0NjE4ZjMwZWU1MTI4OTlmOA==
   data.tar.gz: !binary |-
-    ZjliMGUwMTQyY2IzZmQzY2E0NjJjMmM2YjViMGM4N2M2MmNiYTdhOQ==
+    OGZlMDhlYTg3MzZkY2Y2Mzg4YTY0ZmZjMWQ3MjI1ODliNTU3OTA2Mg==
 SHA512:
   metadata.gz: !binary |-
-    YTJmOWZjOTdmNWYwNDY0OWNjZGE5OWQ1NThhMjk5YzgwNjI5NDgyYjE3NTU2
-    N2QxMjkxN2ViMmQ0OGI1YWNjMDczM2EzNTQ3MmUyOTg5MzIyNTJmYmNjYWU2
-    Mjc5M2I3MThjZjc5N2ZlNzFhMTc5YmVkNTJlZjZlNjViZjk3ODQ=
+    NzZlOTJkMjY1Njc4Mjc4Yzg5YmU1YTM1YTIyYTA5OTlhOTU3NWU0YTQ1Y2Nj
+    MTQ2MWRkMDJhYTQzODA1YTNkYTA3MDYxZTQyYzczNjIyYThjNTM0MzYzOGQ4
+    NGM4NWY4NjE3MGY1MWI1YTA5Y2NmNmY2ZDAzZTFkMmNjMWY4OTk=
   data.tar.gz: !binary |-
-    Y2EyMWI3N2I0MWM3YTRiYzdiYjFhMWQ4NTU2OGVmZmU3ZDRmY2E2MWE3ZTI3
-    NTY0MzZhZGI4NzNhMDc0ZWE2MGRkOTRhMDEzNTAyYzFmNTNhZTA5NWY5NjA3
-    MGYzMTg1ZTZmZDgwNzY5NzBhZDk4Zjg3MmNiYzVhOGFhMjNkNmI=
+    YWI3ZWI3ZmE4M2NkNTBjNjQ4YjA5ZTgyNWUyOWI2YTU3MzU0MTY3YzgyNTE3
+    YjM4ZTRhZjJmMTZkNjg0YWIxMmJjZWM2Yzg1MGVkMWRmMmJkZjA3ZjcxZmFj
+    Yzc2NDQxYWNmZGQ0YmY1NjJiMWVkM2EwNzE1NjA1ZjUwMzI5MTE=

data/lib/timetwister/parser.rb

@@ -296,8 +296,6 @@ class Parser
 		}
 
 
-
-
 		# 19--, 18--, 18--?, etc.
 		match_replace << {
 			:match => "(#{r[:circa]})?[1-2][0-9]\-{2}",
@@ -352,6 +350,13 @@ class Parser
 			:id => 350
 		}
 
+		# 19th century
+		match_replace << {
+			:match => "(#{r[:decade_qualifier]})?\s?[0-2]?[0-9]\s[Cc][Ee][Nn][Tt][Uu][Rr][Yy]",
+			:proc => proc_century_with_qualifiers,
+			:id => 360
+		}
+
 		match_replace
 	end
 
@@ -487,6 +492,34 @@ class Parser
 		end
 	end
 
+	# mid 19th century
+	def self.proc_century_with_qualifiers
+		proc = Proc.new do |string|
+			century = string.match(/[0-9]{2}/).to_s
+
+			if string.match(/[Ee]arly/)
+				range_start = '00'
+				range_end = '40'
+			elsif string.match(/[Mm]id(dle)?/)
+				range_start = '30'
+				range_end = '80'
+			elsif string.match(/[Ll]ate/)
+				range_start = '70'
+				range_end = '00'
+			else
+				range_start = '00'
+				range_end = '99'
+			end
+
+			century_start = (century + range_start).to_i - 100
+			century_end = (century + range_end).to_i - 100
+
+			@dates[:index_dates] = (century_start..century_end).to_a
+			@dates[:inclusive_range] = true
+			process_year_range()
+		end
+	end
+
 	# early 1990s
 	def self.proc_decade_s_qualified
 		proc = Proc.new do |string|
@@ -792,7 +825,7 @@ class Parser
 						month_date_start = datetime_start.strftime('%Y-%m')
 						month_date_end = datetime_end.strftime('%Y-%m-%d')
 					end
-		end
+				end
 
 				if datetime_start && datetime_end
 					process_date_range(datetime_start,datetime_end)
@@ -821,7 +854,6 @@ class Parser
 		end
 	end
 
-
 	# Transform full date strings into parsed datetime objects
 	# e.g. "September 9, 1999" -> datetime
 	def self.full_date_single_to_datetime(string)

data/lib/timetwister/utilities.rb

@@ -291,7 +291,7 @@ class Utilities
 			# circa, ca. - also matches 'c.', which is actually 'copyright', but is still not something we need to deal with
 			:circa => '\s*[Cc](irc)?a?\.?\s*',
 			# early, late, mid-
-			:decade_qualifier => '([Ee]arly)|([Mm]id)|([Ll]ate)\-?',
+			:decade_qualifier => '(([Ee]arly)|([Mm]id)|([Ll]ate))\-?',
 			# 06-16-1972, 6-16-1972
 			:numeric_date_us => '(0?1)|(0?2)|(0?3)|(0?4)|(0?5)|(0?6)|(0?7)|(0?8)|(0?9)|1[0-2][\-\/](([0-2]?[0-9])|3[01])[\-\/])?[12][0-9]{3}',
 			# 1972-06-16

data/lib/timetwister/version.rb

@@ -1,3 +1,3 @@
 module Timetwister
-  VERSION = "0.2.4"
+  VERSION = "0.2.5"
 end

data/lib/timetwister.rb

@@ -4,29 +4,28 @@ require "timetwister/utilities"
 
 module Timetwister
 
-  	def self.parse(str, options={})
-
-  		out = []
-      str = rearrange_conjunctions(str)
-
-  		str.split(';').each do |semi|
-   			semi.split(/\sand\s|\s\&\s/i).each do |conj|
-
-   				# check for dates of form "Month Day(-Day), Year" before splitting on commas
-   				# (removes certainty markers as to not jam the regex)
-   				if Utilities.replace_ordinals(conj).gsub(/[\?\[\]]/, '').match(/[a-z]*\.?\s[0-9]{1,2}(\s?-[0-9]{1,2})?\,\s[0-9]{4}/i)
-   					out << Parser.string_to_dates(conj, options)
-   				else
-   					conj.split(',').each do |comma|
-   						out << Parser.string_to_dates(comma, options)
-   					end
-   				end
-   			end
+  def self.parse(str, options={})
+
+    out = []
+    str = rearrange_conjunctions(str)
+
+    str.split(';').each do |semi|
+      semi.split(/\sand\s|\s\&\s/i).each do |conj|
+
+        # check for dates of form "Month Day(-Day), Year" before splitting on commas
+        # (removes certainty markers as to not jam the regex)
+        if Utilities.replace_ordinals(conj).gsub(/[\?\[\]]/, '').match(/[a-z]*\.?\s[0-9]{1,2}(\s?-[0-9]{1,2})?\,\s[0-9]{4}/i)
+          out << Parser.string_to_dates(conj, options)
+        else
+          conj.split(',').each do |comma|
+            out << Parser.string_to_dates(comma, options)
+          end
+        end
       end
+    end
 
-		return out
-
-	end
+    return out
+  end
 
   # sometimes years are considered implicit in complex dates
   # e.g. "1900 January & February"

data/spec/dates_spec.rb

@@ -647,7 +647,21 @@ describe Timetwister do
 			expect(date[1][:index_dates]).to eq([1776])
 			expect(date[1][:test_data]).to eq("220")
 		end
+	end
 
+	it "parses dates with qualifiers, certainty, and very vague dates" do
+		date = Timetwister.parse("ca. mid 19th century")
+		expect(date[0][:date_start]).to eq("1830")
+		expect(date[0][:date_start_full]).to eq("1830-01-01")
+		expect(date[0][:date_end]).to eq("1880")
+		expect(date[0][:date_end_full]).to eq("1880-12-31")
 	end
 
+	it "parses centuries" do
+		date = Timetwister.parse("17th century")
+		expect(date[0][:date_start]).to eq('1600')
+		expect(date[0][:date_start_full]).to eq("1600-01-01")
+		expect(date[0][:date_end]).to eq("1699")
+		expect(date[0][:date_end_full]).to eq("1699-12-31")
+	end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: timetwister
 version: !ruby/object:Gem::Version
-  version: 0.2.4
+  version: 0.2.5
 platform: ruby
 authors:
 - Alex Duryee
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-01-08 00:00:00.000000000 Z
+date: 2016-04-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler