chronic 0.2.0 → 0.3.0

data/lib/chronic/handlers.rb

@@ -1,80 +1,133 @@
 module Chronic
 
-	class << self
-	  
-	  def definitions #:nodoc:
-	    @definitions ||= 
+  class << self
+
+    def definitions(options={}) #:nodoc:
+      options[:endian_precedence] = [:middle, :little] if options[:endian_precedence].nil?
+
+      # ensure the endian precedence is exactly two elements long
+      raise ChronicPain, "More than two elements specified for endian precedence array" unless options[:endian_precedence].length == 2
+
+      # handler for dd/mm/yyyy
+      @little_endian_handler ||= Handler.new([:scalar_day, :separator_slash_or_dash, :scalar_month, :separator_slash_or_dash, :scalar_year, :separator_at?, 'time?'], :handle_sd_sm_sy)
+
+      # handler for mm/dd/yyyy
+      @middle_endian_handler ||= Handler.new([:scalar_month, :separator_slash_or_dash, :scalar_day, :separator_slash_or_dash, :scalar_year, :separator_at?, 'time?'], :handle_sm_sd_sy)
+
+      # ensure we have valid endian values
+      options[:endian_precedence].each do |e|
+        raise ChronicPain, "Unknown endian type: #{e.to_s}" unless instance_variable_defined?(endian_variable_name_for(e))
+      end
+
+	    @definitions ||=
       {:time => [Handler.new([:repeater_time, :repeater_day_portion?], nil)],
-        
-       :date => [Handler.new([:repeater_month_name, :scalar_day, :scalar_year], :handle_rmn_sd_sy),
+
+       :date => [Handler.new([:repeater_day_name, :repeater_month_name, :scalar_day, :repeater_time, :separator_slash_or_dash?, :time_zone, :scalar_year], :handle_rdn_rmn_sd_t_tz_sy),
+                 Handler.new([:repeater_month_name, :scalar_day, :scalar_year], :handle_rmn_sd_sy),
                  Handler.new([:repeater_month_name, :scalar_day, :scalar_year, :separator_at?, 'time?'], :handle_rmn_sd_sy),
                  Handler.new([:repeater_month_name, :scalar_day, :separator_at?, 'time?'], :handle_rmn_sd),
+                 Handler.new([:repeater_time, :repeater_day_portion?, :separator_on?, :repeater_month_name, :scalar_day], :handle_rmn_sd_on),
                  Handler.new([:repeater_month_name, :ordinal_day, :separator_at?, 'time?'], :handle_rmn_od),
+                 Handler.new([:repeater_time, :repeater_day_portion?, :separator_on?, :repeater_month_name, :ordinal_day], :handle_rmn_od_on),
                  Handler.new([:repeater_month_name, :scalar_year], :handle_rmn_sy),
                  Handler.new([:scalar_day, :repeater_month_name, :scalar_year, :separator_at?, 'time?'], :handle_sd_rmn_sy),
-                 Handler.new([:scalar_month, :separator_slash_or_dash, :scalar_day, :separator_slash_or_dash, :scalar_year, :separator_at?, 'time?'], :handle_sm_sd_sy),
-                 Handler.new([:scalar_day, :separator_slash_or_dash, :scalar_month, :separator_slash_or_dash, :scalar_year, :separator_at?, 'time?'], :handle_sd_sm_sy),
+                 @middle_endian_handler,
+                 @little_endian_handler,
                  Handler.new([:scalar_year, :separator_slash_or_dash, :scalar_month, :separator_slash_or_dash, :scalar_day, :separator_at?, 'time?'], :handle_sy_sm_sd),
                  Handler.new([:scalar_month, :separator_slash_or_dash, :scalar_year], :handle_sm_sy)],
-                 
+
+       # tonight at 7pm
        :anchor => [Handler.new([:grabber?, :repeater, :separator_at?, :repeater?, :repeater?], :handle_r),
+                   Handler.new([:grabber?, :repeater, :repeater, :separator_at?, :repeater?, :repeater?], :handle_r),
                    Handler.new([:repeater, :grabber, :repeater], :handle_r_g_r)],
-                   
+
+       # 3 weeks from now, in 2 months
        :arrow => [Handler.new([:scalar, :repeater, :pointer], :handle_s_r_p),
                   Handler.new([:pointer, :scalar, :repeater], :handle_p_s_r),
                   Handler.new([:scalar, :repeater, :pointer, 'anchor'], :handle_s_r_p_a)],
-                  
+
+       # 3rd week in march
        :narrow => [Handler.new([:ordinal, :repeater, :separator_in, :repeater], :handle_o_r_s_r),
                    Handler.new([:ordinal, :repeater, :grabber, :repeater], :handle_o_r_g_r)]
       }
+
+      apply_endian_precedences(options[:endian_precedence])
+
+      @definitions
     end
-    
-    def tokens_to_span(tokens, options) #:nodoc:                   
+
+    def tokens_to_span(tokens, options) #:nodoc:
       # maybe it's a specific date
-      
-      self.definitions[:date].each do |handler|
-        if handler.match(tokens, self.definitions)
+
+      definitions = self.definitions(options)
+      definitions[:date].each do |handler|
+        if handler.match(tokens, definitions)
+          puts "-date" if Chronic.debug
           good_tokens = tokens.select { |o| !o.get_tag Separator }
           return self.send(handler.handler_method, good_tokens, options)
         end
       end
-            
+
       # I guess it's not a specific date, maybe it's just an anchor
-            
-      self.definitions[:anchor].each do |handler|
-        if handler.match(tokens, self.definitions)
+
+      definitions[:anchor].each do |handler|
+        if handler.match(tokens, definitions)
+          puts "-anchor" if Chronic.debug
           good_tokens = tokens.select { |o| !o.get_tag Separator }
           return self.send(handler.handler_method, good_tokens, options)
         end
       end
-            
+
       # not an anchor, perhaps it's an arrow
-      
-      self.definitions[:arrow].each do |handler|
-        if handler.match(tokens, self.definitions)
+
+      definitions[:arrow].each do |handler|
+        if handler.match(tokens, definitions)
+          puts "-arrow" if Chronic.debug
           good_tokens = tokens.reject { |o| o.get_tag(SeparatorAt) || o.get_tag(SeparatorSlashOrDash) || o.get_tag(SeparatorComma) }
           return self.send(handler.handler_method, good_tokens, options)
         end
       end
-      
-      # not an arrow, let's hope it's an narrow
-      
-      self.definitions[:narrow].each do |handler|
-        if handler.match(tokens, self.definitions)
+
+      # not an arrow, let's hope it's a narrow
+
+      definitions[:narrow].each do |handler|
+        if handler.match(tokens, definitions)
+          puts "-narrow" if Chronic.debug
           #good_tokens = tokens.select { |o| !o.get_tag Separator }
           return self.send(handler.handler_method, tokens, options)
         end
       end
-      
+
       # I guess you're out of luck!
+      puts "-none" if Chronic.debug
       return nil
     end
-    
+
     #--------------
-    
+
+    def apply_endian_precedences(precedences)
+      date_defs = @definitions[:date]
+
+      # map the precedence array to indices on @definitions[:date]
+      indices = precedences.map { |e|
+        handler = instance_variable_get(endian_variable_name_for(e))
+        date_defs.index(handler)
+      }
+
+      # swap the handlers if we discover they are at odds with the desired preferences
+      swap(date_defs, indices.first, indices.last) if indices.first > indices.last
+    end
+
+    def endian_variable_name_for(e)
+      "@#{e.to_s}_endian_handler".to_sym
+    end
+
+    # exchange two elements in an array
+    def swap(arr, a, b); arr[a], arr[b] = arr[b], arr[a]; end
+
     def day_or_time(day_start, time_tokens, options)
       outer_span = Span.new(day_start, day_start + (24 * 60 * 60))
-      
+
       if !time_tokens.empty?
         @now = outer_span.begin
         time = get_anchor(dealias_and_disambiguate_times(time_tokens, options), options)
@@ -83,30 +136,46 @@ module Chronic
         return outer_span
       end
     end
-    
+
     #--------------
-    
+
     def handle_m_d(month, day, time_tokens, options) #:nodoc:
       month.start = @now
       span = month.this(options[:context])
-      
-      day_start = Time.local(span.begin.year, span.begin.month, day)
-      
+
+      day_start = Chronic.time_class.local(span.begin.year, span.begin.month, day)
+
       day_or_time(day_start, time_tokens, options)
     end
-    
+
     def handle_rmn_sd(tokens, options) #:nodoc:
       handle_m_d(tokens[0].get_tag(RepeaterMonthName), tokens[1].get_tag(ScalarDay).type, tokens[2..tokens.size], options)
     end
-    
+
+    def handle_rmn_sd_on(tokens, options) #:nodoc:
+      if tokens.size > 3
+        handle_m_d(tokens[2].get_tag(RepeaterMonthName), tokens[3].get_tag(ScalarDay).type, tokens[0..1], options)
+      else
+        handle_m_d(tokens[1].get_tag(RepeaterMonthName), tokens[2].get_tag(ScalarDay).type, tokens[0..0], options)
+      end
+    end
+
     def handle_rmn_od(tokens, options) #:nodoc:
       handle_m_d(tokens[0].get_tag(RepeaterMonthName), tokens[1].get_tag(OrdinalDay).type, tokens[2..tokens.size], options)
     end
-    
+
+    def handle_rmn_od_on(tokens, options) #:nodoc:
+      if tokens.size > 3
+        handle_m_d(tokens[2].get_tag(RepeaterMonthName), tokens[3].get_tag(OrdinalDay).type, tokens[0..1], options)
+      else
+        handle_m_d(tokens[1].get_tag(RepeaterMonthName), tokens[2].get_tag(OrdinalDay).type, tokens[0..0], options)
+      end
+    end
+
     def handle_rmn_sy(tokens, options) #:nodoc:
       month = tokens[0].get_tag(RepeaterMonthName).index
       year = tokens[1].get_tag(ScalarYear).type
-      
+
       if month == 12
         next_month_year = year + 1
         next_month_month = 1
@@ -114,66 +183,71 @@ module Chronic
         next_month_year = year
         next_month_month = month + 1
       end
-      
+
       begin
-        Span.new(Time.local(year, month), Time.local(next_month_year, next_month_month))
+        Span.new(Chronic.time_class.local(year, month), Chronic.time_class.local(next_month_year, next_month_month))
       rescue ArgumentError
         nil
       end
     end
-    
+
+    def handle_rdn_rmn_sd_t_tz_sy(tokens, options) #:nodoc:
+      t = Chronic.time_class.parse(@text)
+      Span.new(t, t + 1)
+    end
+
     def handle_rmn_sd_sy(tokens, options) #:nodoc:
       month = tokens[0].get_tag(RepeaterMonthName).index
       day = tokens[1].get_tag(ScalarDay).type
       year = tokens[2].get_tag(ScalarYear).type
-      
+
       time_tokens = tokens.last(tokens.size - 3)
-      
+
       begin
-        day_start = Time.local(year, month, day)
+        day_start = Chronic.time_class.local(year, month, day)
         day_or_time(day_start, time_tokens, options)
       rescue ArgumentError
         nil
       end
     end
-    
+
     def handle_sd_rmn_sy(tokens, options) #:nodoc:
       new_tokens = [tokens[1], tokens[0], tokens[2]]
       time_tokens = tokens.last(tokens.size - 3)
       self.handle_rmn_sd_sy(new_tokens + time_tokens, options)
     end
-    
+
     def handle_sm_sd_sy(tokens, options) #:nodoc:
       month = tokens[0].get_tag(ScalarMonth).type
       day = tokens[1].get_tag(ScalarDay).type
       year = tokens[2].get_tag(ScalarYear).type
-      
+
       time_tokens = tokens.last(tokens.size - 3)
-      
+
       begin
-        day_start = Time.local(year, month, day) #:nodoc:
+        day_start = Chronic.time_class.local(year, month, day) #:nodoc:
         day_or_time(day_start, time_tokens, options)
       rescue ArgumentError
         nil
       end
     end
-    
+
     def handle_sd_sm_sy(tokens, options) #:nodoc:
       new_tokens = [tokens[1], tokens[0], tokens[2]]
       time_tokens = tokens.last(tokens.size - 3)
       self.handle_sm_sd_sy(new_tokens + time_tokens, options)
     end
-    
+
     def handle_sy_sm_sd(tokens, options) #:nodoc:
       new_tokens = [tokens[1], tokens[2], tokens[0]]
       time_tokens = tokens.last(tokens.size - 3)
       self.handle_sm_sd_sy(new_tokens + time_tokens, options)
     end
-    
+
     def handle_sm_sy(tokens, options) #:nodoc:
       month = tokens[0].get_tag(ScalarMonth).type
       year = tokens[1].get_tag(ScalarYear).type
-      
+
       if month == 12
         next_month_year = year + 1
         next_month_month = 1
@@ -181,40 +255,40 @@ module Chronic
         next_month_year = year
         next_month_month = month + 1
       end
-      
+
       begin
-        Span.new(Time.local(year, month), Time.local(next_month_year, next_month_month))
+        Span.new(Chronic.time_class.local(year, month), Chronic.time_class.local(next_month_year, next_month_month))
       rescue ArgumentError
         nil
       end
     end
-    
+
     # anchors
-    
+
     def handle_r(tokens, options) #:nodoc:
       dd_tokens = dealias_and_disambiguate_times(tokens, options)
       self.get_anchor(dd_tokens, options)
     end
-    
+
     def handle_r_g_r(tokens, options) #:nodoc:
       new_tokens = [tokens[1], tokens[0], tokens[2]]
       self.handle_r(new_tokens, options)
     end
-    
+
     # arrows
-    
+
     def handle_srp(tokens, span, options) #:nodoc:
       distance = tokens[0].get_tag(Scalar).type
       repeater = tokens[1].get_tag(Repeater)
       pointer = tokens[2].get_tag(Pointer).type
-      
+
       repeater.offset(span, distance, pointer)
     end
-    
+
     def handle_s_r_p(tokens, options) #:nodoc:
       repeater = tokens[1].get_tag(Repeater)
-            
-      # span = 
+
+      # span =
       # case true
       # when [RepeaterYear, RepeaterSeason, RepeaterSeasonName, RepeaterMonth, RepeaterMonthName, RepeaterFortnight, RepeaterWeek].include?(repeater.class)
       #   self.parse("this hour", :guess => false, :now => @now)
@@ -225,24 +299,24 @@ module Chronic
       # else
       #   raise(ChronicPain, "Invalid repeater: #{repeater.class}")
       # end
-      
+
       span = self.parse("this second", :guess => false, :now => @now)
-      
+
       self.handle_srp(tokens, span, options)
     end
-    
+
     def handle_p_s_r(tokens, options) #:nodoc:
       new_tokens = [tokens[1], tokens[2], tokens[0]]
       self.handle_s_r_p(new_tokens, options)
     end
-    
+
     def handle_s_r_p_a(tokens, options) #:nodoc:
       anchor_span = get_anchor(tokens[3..tokens.size - 1], options)
       self.handle_srp(tokens, anchor_span, options)
     end
-    
+
     # narrows
-    
+
     def handle_orr(tokens, outer_span, options) #:nodoc:
       repeater = tokens[1].get_tag(Repeater)
       repeater.start = outer_span.begin - 1
@@ -257,34 +331,34 @@ module Chronic
       end
       span
     end
-    
+
     def handle_o_r_s_r(tokens, options) #:nodoc:
       outer_span = get_anchor([tokens[3]], options)
       handle_orr(tokens[0..1], outer_span, options)
     end
-    
+
     def handle_o_r_g_r(tokens, options) #:nodoc:
       outer_span = get_anchor(tokens[2..3], options)
       handle_orr(tokens[0..1], outer_span, options)
     end
-    
+
     # support methods
-    
+
     def get_anchor(tokens, options) #:nodoc:
       grabber = Grabber.new(:this)
       pointer = :future
-      
+
       repeaters = self.get_repeaters(tokens)
       repeaters.size.times { tokens.pop }
-                    
+
       if tokens.first && tokens.first.get_tag(Grabber)
         grabber = tokens.first.get_tag(Grabber)
         tokens.pop
       end
-      
+
       head = repeaters.shift
       head.start = @now
-                  
+
       case grabber.type
         when :last
           outer_span = head.next(:past)
@@ -298,11 +372,11 @@ module Chronic
           outer_span = head.next(:future)
         else raise(ChronicPain, "Invalid grabber")
       end
-      
+
       puts "--#{outer_span}" if Chronic.debug
       anchor = find_within(repeaters, outer_span, pointer)
     end
-    
+
     def get_repeaters(tokens) #:nodoc:
       repeaters = []
 		  tokens.each do |token|
@@ -312,44 +386,76 @@ module Chronic
       end
       repeaters.sort.reverse
     end
-    
+
     # Recursively finds repeaters within other repeaters.
     # Returns a Span representing the innermost time span
     # or nil if no repeater union could be found
     def find_within(tags, span, pointer) #:nodoc:
       puts "--#{span}" if Chronic.debug
       return span if tags.empty?
-      
+
       head, *rest = tags
       head.start = pointer == :future ? span.begin : span.end
       h = head.this(:none)
-            
+
       if span.include?(h.begin) || span.include?(h.end)
         return find_within(rest, h, pointer)
       else
         return nil
       end
     end
-    
+
     def dealias_and_disambiguate_times(tokens, options) #:nodoc:
       # handle aliases of am/pm
-      # 5:00 in the morning => 5:00 am
-      # 7:00 in the evening => 7:00 pm
-      #ttokens = []
-      tokens.each_with_index do |t0, i|
-        t1 = tokens[i + 1]
-        if t1 && (t1tag = t1.get_tag(RepeaterDayPortion)) && t0.get_tag(RepeaterTime)
-          if [:morning].include?(t1tag.type)
-            t1.untag(RepeaterDayPortion)
-            t1.tag(RepeaterDayPortion.new(:am))
-          elsif [:afternoon, :evening, :night].include?(t1tag.type)
-            t1.untag(RepeaterDayPortion)
-            t1.tag(RepeaterDayPortion.new(:pm))
-          end
+      # 5:00 in the morning -> 5:00 am
+      # 7:00 in the evening -> 7:00 pm
+
+      day_portion_index = nil
+      tokens.each_with_index do |t, i|
+        if t.get_tag(RepeaterDayPortion)
+          day_portion_index = i
+          break
+        end
+      end
+
+      time_index = nil
+      tokens.each_with_index do |t, i|
+        if t.get_tag(RepeaterTime)
+          time_index = i
+          break
+        end
+      end
+
+      if (day_portion_index && time_index)
+        t1 = tokens[day_portion_index]
+        t1tag = t1.get_tag(RepeaterDayPortion)
+
+        if [:morning].include?(t1tag.type)
+          puts '--morning->am' if Chronic.debug
+          t1.untag(RepeaterDayPortion)
+          t1.tag(RepeaterDayPortion.new(:am))
+        elsif [:afternoon, :evening, :night].include?(t1tag.type)
+          puts "--#{t1tag.type}->pm" if Chronic.debug
+          t1.untag(RepeaterDayPortion)
+          t1.tag(RepeaterDayPortion.new(:pm))
         end
       end
-      #tokens = ttokens
-            
+
+      # tokens.each_with_index do |t0, i|
+      #   t1 = tokens[i + 1]
+      #   if t1 && (t1tag = t1.get_tag(RepeaterDayPortion)) && t0.get_tag(RepeaterTime)
+      #     if [:morning].include?(t1tag.type)
+      #       puts '--morning->am' if Chronic.debug
+      #       t1.untag(RepeaterDayPortion)
+      #       t1.tag(RepeaterDayPortion.new(:am))
+      #     elsif [:afternoon, :evening, :night].include?(t1tag.type)
+      #       puts "--#{t1tag.type}->pm" if Chronic.debug
+      #       t1.untag(RepeaterDayPortion)
+      #       t1.tag(RepeaterDayPortion.new(:pm))
+      #     end
+      #   end
+      # end
+
       # handle ambiguous times if :ambiguous_time_range is specified
       if options[:ambiguous_time_range] != :none
         ttokens = []
@@ -364,25 +470,25 @@ module Chronic
         end
         tokens = ttokens
       end
-      
+
       tokens
     end
-    
+
   end
-  
+
   class Handler #:nodoc:
     attr_accessor :pattern, :handler_method
-    
+
     def initialize(pattern, handler_method)
       @pattern = pattern
       @handler_method = handler_method
     end
-    
+
     def constantize(name)
       camel = name.to_s.gsub(/(^|_)(.)/) { $2.upcase }
       ::Chronic.module_eval(camel, __FILE__, __LINE__)
     end
-    
+
     def match(tokens, definitions)
       token_index = 0
       @pattern.each do |element|
@@ -409,6 +515,10 @@ module Chronic
       return false if token_index != tokens.size
       return true
     end
+
+    def ==(other)
+      self.pattern == other.pattern
+    end
   end
-  
-end
+
+end

data/lib/{numerizer → chronic/numerizer}/numerizer.rb

@@ -44,32 +44,34 @@ class Numerizer
                     ['trillion', 1_000_000_000_000],
                   ]
 
-class << self
-  def numerize(string)
+  def self.numerize(string)
     string = string.dup
-  
+
     # preprocess
-    string.gsub!(/ +|([^\d])-([^d])/, '\1 \2') # will mutilate hyphenated-words but shouldn't matter for date extraction
+    string.gsub!(/ +|([^\d])-([^\d])/, '\1 \2') # will mutilate hyphenated-words but shouldn't matter for date extraction
     string.gsub!(/a half/, 'haAlf') # take the 'a' out so it doesn't turn into a 1, save the half for the end
 
     # easy/direct replacements
-  
+
     DIRECT_NUMS.each do |dn|
-      string.gsub!(/#{dn[0]}/i, dn[1])
+      string.gsub!(/#{dn[0]}/i, '<num>' + dn[1])
     end
 
     # ten, twenty, etc.
 
     TEN_PREFIXES.each do |tp|
-      string.gsub!(/(?:#{tp[0]})( *\d(?=[^\d]|$))*/i) { (tp[1] + $1.to_i).to_s }
+      string.gsub!(/(?:#{tp[0]}) *<num>(\d(?=[^\d]|$))*/i) { '<num>' + (tp[1] + $1.to_i).to_s }
+    end
+
+    TEN_PREFIXES.each do |tp|
+      string.gsub!(/#{tp[0]}/i) { '<num>' + tp[1].to_s }
     end
 
     # hundreds, thousands, millions, etc.
 
     BIG_PREFIXES.each do |bp|
-      string.gsub!(/(\d*) *#{bp[0]}/i) { (bp[1] * $1.to_i).to_s}
+      string.gsub!(/(?:<num>)?(\d*) *#{bp[0]}/i) { '<num>' + (bp[1] * $1.to_i).to_s}
       andition(string)
-      #combine_numbers(string) # Should to be more efficient way to do this
     end
 
     # fractional addition
@@ -77,27 +79,19 @@ class << self
     # (with extraneous .0's and such )
     string.gsub!(/(\d+)(?: | and |-)*haAlf/i) { ($1.to_f + 0.5).to_s }
 
-    string
+    string.gsub(/<num>/, '')
   end
 
-private
-  def andition(string)
+  private
+
+  def self.andition(string)
     sc = StringScanner.new(string)
-    while(sc.scan_until(/(\d+)( | and )(\d+)(?=[^\w]|$)/i))
+    while(sc.scan_until(/<num>(\d+)( | and )<num>(\d+)(?=[^\w]|$)/i))
       if sc[2] =~ /and/ || sc[1].size > sc[3].size
-        string[(sc.pos - sc.matched_size)..(sc.pos-1)] = (sc[1].to_i + sc[3].to_i).to_s
+        string[(sc.pos - sc.matched_size)..(sc.pos-1)] = '<num>' + (sc[1].to_i + sc[3].to_i).to_s
         sc.reset
       end
     end
   end
 
-#  def combine_numbers(string)
-#    sc = StringScanner.new(string)
-#    while(sc.scan_until(/(\d+)(?: | and |-)(\d+)(?=[^\w]|$)/i))
-#      string[(sc.pos - sc.matched_size)..(sc.pos-1)] = (sc[1].to_i + sc[2].to_i).to_s
-#      sc.reset
-#    end
-#  end
-
 end
-end