datte 0.1.0 → 0.3.0

data/lib/datte/date_parser.rb

@@ -19,23 +19,36 @@ module Datte
         end
       end
 
-      return @date.to_datetime
+      ABSOLUTE_TIMES.each do |matcher|
+        if md = @body.match(matcher)
+          @date.update_time(md)
+          p @date
+          break
+        end
+      end
 
-      # MATCHER.each do |matcher|
-      #   @body.match(matcher) do |md|
-      #     y = year(md)
-      #     m = month(md)
-      #     d = day(md)
-      #     #h = hour(md)
-      #     #m = minute(md)
-      #
-      #     @md = md
-      #     return y, m, d
-      #   end
-      # end
+      NOUNS.each do |matcher_s, method|
+        matcher = Regexp.new(matcher_s.to_s)
+        if md = @body.match(matcher)
+          eval(method)
+          break
+        end
+      end
+
+      AFTERS.each do |matcher|
+        if md = @body.match(matcher)
+          @date.after(md)
+          break
+        end
+      end
+
+
+      return @date.to_datetime
     end
 
     private
 
+    def next_day(day)
+    end
   end
 end

data/lib/datte/datetime_table.rb

@@ -37,17 +37,25 @@ module Datte
     '(?<min>\d{1,2})分'
   ].map { |pattern| Regexp.compile(pattern) }.freeze
 
-  # DATES = {
-  #   '明日|あした': 'send(:next_day, 1)',
-  #   '明後日|あさって': 'send(:next_day, 2)',
-  #   '': '',
-  # }
+  NOUNS = {
+    '明日|あした|あす': 'send(:next_day, 1)',
+    '明後日|あさって': 'send(:next_day, 2)',
+    '明々後日|しあさって': 'send(:next_day, 3)',
+    '今日|きょう': 'send(:next_day, 0)'
+  }
 
-  # TIMES = {
-  #   '(?<hour>\d{1,2})時(?<min>\d{1,2})分',
-  #   '(?<hour>\d{1,2})時',
-  #   ''
-  # }
+  AFTERS = [
+    # 何年後
+    '(?<year>\d{1)年後',
+    # 何ヶ月後
+    '(?<month>\d{1,2}ヶ月後)',
+    # 何日後
+    '(?<day>\d{1,2})日後',
+    # 何時間後
+    '(?<hour>\d{1,2})時間後',
+    # 何分後
+    '(?<min>\d{1,2})分後'
+  ].map { |pattern| Regexp.compile(pattern) }.freeze
 
   class DatetimeTable
     def initialize

data/lib/datte/dattetime.rb

@@ -2,22 +2,20 @@ module Datte
   class Dattetime
 
     DEFAULT_OPTIONS = {
-      force_update: false
+      force_update: false,
+      level: 1
     }
 
-    attr_reader :year, :month, :day, :hour, :min, :sec
+    attr_reader :year, :month, :day, :hour, :min
 
     def initialize(options = {})
       @options = DEFAULT_OPTIONS.merge(options)
-      @date = Date.today
-    end
-
-    def to_s
-      @date.to_s
+      @date = DateTime.now
     end
 
     def to_datetime
-      DateTime.new(@year, @month, @day, 12, 0, 0) rescue nil
+      return nil unless check_level?
+      DateTime.new(y, m, d, h, mi, 0) rescue nil
     end
 
     # 年か月か日を更新
@@ -29,18 +27,29 @@ module Datte
     end
 
     # 時か分を更新
-    def update_time(hour, min)
-    end
-
-    # 何年後、何ヶ月後、何日後
-    def after_date(year, month, day)
+    def update_time(md, options = @options)
+      op = @options[:force_update] ? '=' : '||='
+      eval("@hour #{op} hour!(md)")
+      eval("@min #{op} min!(md)")
     end
 
-    # 何時間後、何分後
-    def after_time(hour, min)
+    # 何年後、何ヶ月後、何日後, 何時間後, 何分後
+    def after(md)
+      @date >> (md[:year].to_i * 12) if md.matched?(:year) # 何年後
+      @date >> md[:month].to_i if md.matched?(:month) # 何ヶ月後
+      @date + md[:day].to_i if md.matched?(:day) # 何日後
+      @date + Rational(md[:hour].to_i, 24) # 何時間後
+      @date + Rational(md[:hour].to_i, 24 * 60) # 何分後
     end
 
     private
+
+    def y; @year || now[:year] end
+    def m; @month || now[:month] end
+    def d; @day || now[:day] end
+    def h; @hour || now[:hour] end
+    def mi; @min || 0 end
+
     def now
       d = DateTime.now
       { year: d.year, month: d.month, day: d.day, hour: d.hour, min: d.min }
@@ -57,6 +66,22 @@ module Datte
     def day!(md)
       md.matched?(:day) ? md[:day].to_i : now[:day]
     end
+
+    def hour!(md)
+      md.matched?(:hour) ? md[:hour].to_i : now[:hour]
+    end
+
+    def min!(md)
+      md.matched?(:min) ? md[:min].to_i : 0
+    end
+
+    def check_level?
+      counter = 0
+      [@year, @month, @day, @hour, @min].each do |check|
+        counter += 1 unless check.nil?
+      end
+      @options[:level] < counter
+    end
   end
 end
 

data/lib/datte/train.rb

@@ -0,0 +1,203 @@
+module Datte
+  class Train
+    FNAME = File.join(File.dirname(__FILE__), 'data.txt')
+
+    # SEE: http://qiita.com/Hironsan/items/326b66711eb4196aa9d4
+
+    def initialize(path = FNAME)
+      train_sents = corpus_read
+      x = corpus_read[0]
+      p x
+      p '==='
+      p sent2features(x)
+      #sent2features(corpus_read[0])[0]
+      #sent2features(train_sents[0])[0]
+    end
+
+    def x_train
+      train_sents.each do |s|
+        return sent2features(s)
+      end
+    end
+
+    def y_train
+      train_sents.each do |s|
+        return sent2labels(s)
+      end
+    end
+
+    def x_test
+      test_sents.each do |s|
+        return sent2features(s)
+      end
+    end
+
+    def y_test
+      test_sents.each do |s|
+        return sent2labels(s)
+      end
+    end
+
+    def train()
+    end
+
+    private
+    def hiragana?(ch)
+      0x3040 <= ch.ord && ch.ord <= 0x309F
+    end
+
+    def katakana(ch)
+      0x30A0 <= ch.ord && ch.ord <= 0x30FF
+    end
+
+    def space?(ch)
+      !(ch =~ /^\s*$/).nil?
+    end
+
+    def integer?(ch)
+      Integer(ch)
+      Integer(ch)
+      true
+    rescue ArgumentError
+      false
+    end
+
+    def lower?(ch)
+      ch == ch.downcase
+    end
+
+    def upper?(ch)
+      ch == ch.upcase
+    end
+
+    def chara_type(ch)
+      if space?(ch) then 'ZSPACE'
+      elsif integer?(ch) then 'ZDIGIT'
+      elsif lower?(ch) then 'ZLLET'
+      elsif upper?(ch) then 'ZULET'
+      elsif hiragana?(ch) then 'HIRAG'
+      elsif katakana?(ch) then 'KATAK'
+      else 'OTHER'
+      end
+    end
+
+    def chara_types(str)
+      types = str.each_char.to_a.map do |ch|
+        chara_type(ch)
+      end
+      types.uniq.sort().join('-')
+    end
+
+    def extract_pos(morph)
+      idx = morph.index('*')
+      morph[1, idx-1].join('-')
+    end
+
+    def word2features(sent, i)
+      word = sent[i][0]
+      chtype = chara_types(sent[i][0])
+      postag = extract_pos(sent[i])
+
+      features = [
+        'bias',
+        'word=' + word,
+        'type=' + chtype,
+        'pos_tag=' + postag
+      ]
+
+      if i >= 2
+        word2 = sent[i-2][0]
+        chtype2 = chara_types(sent[i-2][0])
+        postag2 = extract_pos(sent[i-2])
+        iobtag2 = sent[i-2][-1]
+        features.push(*[
+          '-2:word=' + word2,
+          '-2:type=' + chtype2,
+          '-2:postag=' + postag2,
+          '-2:iobtag=' + iobtag2
+        ])
+      else
+        features.push('BOS')
+      end
+
+      if i >= 1
+        word1 = sent[i-1][0]
+        chtype1 = chara_types(sent[i-1][0])
+        postag1 = extract_pos(sent[i-1])
+        iobtag1 = sent[i-1][-1]
+        features.push(*[
+          '-1:word=' + word1,
+          '-1:type=' + chtype1,
+          '-1:postag=' + postag1,
+          '-1:iobtag=' + iobtag1
+        ])
+      else
+        features.push('BOS')
+      end
+
+      if i < sent.length - 1
+        word1 = sent[i+1][0]
+        chtype1 = chara_types(sent[i+1][0])
+        postag1 = extract_pos(sent[i+1])
+        features.push(*[
+          '+1:word=' + word1,
+          '+1:type=' + chtype1,
+          '+1:postag=' + postag1
+        ])
+      else
+        features.push('EOS')
+      end
+
+      if i < sent.length - 2
+        word2 = sent[i+2][0]
+        chtype2 = chara_types(sent[i+2][0])
+        postag2 = extract_pos(sent[i+2])
+        features.push(*[
+          '+2:word=' + word2,
+          '+2:type=' + chtype2,
+          '+2:postag=' + postag2
+        ])
+      else
+        features.push('EOS')
+      end
+
+      return features
+    end
+
+    def sent2features(sent)
+      (0..(sent.length)).to_a.map do |i|
+        return word2features(sent, i)
+      end
+    end
+
+    def sent2labels(sent)
+      sent.map do |morph|
+        morph[-1]
+      end
+    end
+
+    def sent2tokens(sent)
+      sent.map do |morph|
+        morph[0]
+      end
+    end
+
+    def corpus_read
+      sents, sent = [], []
+
+      File.open(FNAME, 'r') do |file|
+        file.each_line do |line|
+          if line == "\n"
+            sents.push(sent)
+            sent = []
+            next
+          end
+          morph_info = line.strip().split(' ')
+          sent.push(morph_info)
+        end
+      end
+      sents
+    end
+
+  end
+end

data/lib/datte/version.rb

@@ -1,3 +1,3 @@
 module Datte
-  VERSION = "0.1.0"
+  VERSION = "0.3.0"
 end

data/lib/datte.rb

@@ -4,3 +4,4 @@ require File.join(File.dirname(__FILE__), 'datte', 'parser')
 require File.join(File.dirname(__FILE__), 'datte', 'date_parser')
 require File.join(File.dirname(__FILE__), 'datte', 'dattetime')
 require File.join(File.dirname(__FILE__), 'datte', 'datetime_table')
+require File.join(File.dirname(__FILE__), 'datte', 'train')

data/spec/datte_spec.rb

@@ -5,7 +5,13 @@ describe Datte do
     expect(Datte::VERSION).not_to be nil
   end
 
-  it "does something useful" do
-    expect(false).to eq(true)
+  describe "::Parser" do
+    datte = Datte::Parser.new
+
+    it "2016/11/1日に遊ぼー" do
+      body = "2016/11/1日に遊ぼー"
+      d = DateTime.now
+      expect(datte.parse_date(body)).to eq(DateTime.new(2016, 11, 1, d.hour, 0, 0))
+    end
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: datte
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.3.0
 platform: ruby
 authors:
 - pokohide
@@ -73,10 +73,12 @@ files:
 - bin/setup
 - datte.gemspec
 - lib/datte.rb
+- lib/datte/data.txt
 - lib/datte/date_parser.rb
 - lib/datte/datetime_table.rb
 - lib/datte/dattetime.rb
 - lib/datte/parser.rb
+- lib/datte/train.rb
 - lib/datte/version.rb
 - spec/datte_spec.rb
 - spec/spec_helper.rb