Class: Natural::Fragment

Inherits:
Tree::TreeNode
  • Object
show all
Defined in:
lib/natural/fragment.rb

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(options = {}) ⇒ Fragment

Returns a new instance of Fragment.



7
8
9
10
11
 
# File 'lib/natural/fragment.rb', line 7

def initialize(options={})
  @ids = options[:ids]
  self.text = options[:text]
  super("#{GREEN}#{options[:text]}#{CLEAR} #{self.class.to_s.split('::').last.underscore} (#{self.id_range})", options[:text])
end

Instance Attribute Details

#aggregator ⇒ Object

Returns the value of attribute aggregator.



5
6
7
 
# File 'lib/natural/fragment.rb', line 5

def aggregator
  @aggregator
end

#filter ⇒ Object

Returns the value of attribute filter.



5
6
7
 
# File 'lib/natural/fragment.rb', line 5

def filter
  @filter
end

#score ⇒ Object

Returns the value of attribute score.



5
6
7
 
# File 'lib/natural/fragment.rb', line 5

def score
  @score
end

#text ⇒ Object

Returns the value of attribute text.



5
6
7
 
# File 'lib/natural/fragment.rb', line 5

def text
  @text
end

Class Method Details

.find(options) ⇒ Object 



88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
 
# File 'lib/natural/fragment.rb', line 88

def self.find(options)
  text_to_search = options[:text]
  looking_for = options[:looking_for]
  old_matches = options[:matches] || {}
  if options[:matches] && (options[:merge_results].class == NilClass || options[:merge_results])
    new_matches = options[:matches]
  else
    new_matches = {}
  end
  match_class = options[:match_class] || self
  words = text_to_search.split(' ')

  case
  when looking_for.class == String || (looking_for.class == Array && looking_for.all? {|a| a.class == String})
    return old_matches if old_matches[match_class]
    looking_for = [looking_for] if looking_for.class != Array
    looking_for = looking_for.map{|a| a.singularize.downcase}

    # look for the longest possible matches and work our way down to the short ones
    0.upto(words.size-1) do |first|
      (words.size-1).downto(first) do |last|
        match = nil
        selection = words[(first..last)].join(' ').strip.downcase

        if looking_for.include?(selection.singularize.downcase)
          match = match_class.new(:ids => (first..last).to_a, :text => selection)
        end

        # didn't find a simple match, try swapping some or all words for alternatives and try again
        if !match && !(match_class < Natural::Alternative)
          fragments = old_matches.select {|k,v| k < Natural::Alternative && !v.blank?}.values.flatten.select {|a| a.ids.first >= first && a.ids.last <= last}

          # assemble a list of all the possible, non-overlapping swaps
          combinations = (1..fragments.size).inject([]) do |memo, i| 
            fragments.combination(i).each do |combo|
              if !combo.combination(2).any? {|a| (a[0].ids.first..a[0].ids.last).overlaps?(a[1].ids.first..a[1].ids.last)}
                memo << combo
              end
            end                
            memo
          end

          combinations.each do |combo|
            alternative_words = words.clone
            alternative_fragments = []

            combo.each do |fragment|
              alternative_words.slice!(fragment.ids.first..fragment.ids.last)
              alternative_words.insert(fragment.ids.first, fragment.to_s)
              alternative_fragments << fragment
            end
            alternative_selection = alternative_words[(first..last)].join(' ').strip.downcase

            if looking_for.include?(alternative_selection.singularize.downcase)
              match = match_class.new(:ids => (first..last).to_a, :text => alternative_selection)
              leftovers = ((first..last).to_a - combo.map {|a| a.ids}.flatten).to_ranges
              leftovers.each do |range|
                alternative_fragments << Fragment.new(:ids => range.to_a, :text => words[range].join(' '))
              end
              alternative_fragments.sort_by {|a| a.ids.first}.each {|a| match << a}
            end

          end
        end

        new_matches[match_class] = [] if !new_matches[match_class]
        if match
          if match_class < Natural::Alternative
            new_matches = recurse_alternatives(match, options)
          else
            new_matches[match_class] << match
          end
        end
      end
    end

  when looking_for.class <= Fragment

    return old_matches if old_matches[looking_for]
    new_matches = klass.find(:text => text_to_search, :matches => old_matches, :spellings => options[:spellings], :synonyms => options[:synonyms], :expansions => options[:expansions])

  when (looking_for.class == Hash && looking_for[:or]) || looking_for.class == Array

    looking_for.each do |term|
      new_matches = Fragment.find(:text => text_to_search, :looking_for => term, :matches => old_matches, :match_class => match_class, :spellings => options[:spellings], :synonyms => options[:synonyms], :expansions => options[:expansions])
    end

  when looking_for.class == Hash && looking_for[:and] # look for a sequence of strings and/or fragments
    looking_for = looking_for[:and]
    # first we find the starting term
    if looking_for.first.class == Class && looking_for.first <= Fragment
      new_matches = looking_for.first.find(:text => text_to_search, :matches => old_matches, :spellings => options[:spellings], :synonyms => options[:synonyms], :expansions => options[:expansions])
      starting_term_matches = old_matches[looking_for.first]
    else
      starting_term_matches = Fragment.find(options.merge(:looking_for => looking_for.first, :merge_results => false)).values.first
    end

    # look for the next string/fragment in the sequence
    (starting_term_matches || []).each do |first_term|
      fragments = [first_term]
      looking_for[1..-1].each do |term|
        if term.class == Class && term <= Fragment
          new_matches = term.find(:text => text_to_search, :matches => old_matches, :spellings => options[:spellings], :synonyms => options[:synonyms], :expansions => options[:expansions]) if !old_matches[term]
          new_matches[term].each do |match|
            if match.ids.first == fragments.select {|a| a}.last.ids.last + 1
              fragments << match
            end
          end
        elsif [Array, Hash, String].include?(term.class)
          term_updated = term.class == String ? [term] : term
          (Fragment.find(:text => text_to_search, :looking_for => term_updated, :spellings => options[:spellings], :synonyms => options[:synonyms], :expansions => options[:expansions]).values.first || []).each do |match|
            if match.ids.first == fragments.select {|a| a}.last.ids.last + 1
              fragments << Fragment.new(:ids => match.ids, :text => match.to_s)
            end
          end
        else # turn nils into fragments
          last_fragment = fragments.select {|a| a}.last
          id = last_fragment.ids.last + 1
          fragments << Fragment.new(:ids => [id], :text => words[id])
        end
      end

      # found a match
      looking_for_updated = looking_for.map{|a| [String, Array, Hash, NilClass].include?(a.class) ? Fragment : a}

      if fragments.map{|a| a.class} == looking_for_updated
        ids = (fragments.first.ids.first..fragments.last.ids.last).to_a
        text = fragments.inject('') {|memo, fragment| memo += fragment.to_s + ' '}.strip
        match = match_class.new(:ids => ids, :text => text)
        fragments.each do |fragment|
          match << fragment
        end

        new_matches[match_class] = [] if !new_matches[match_class]
        if match
          new_matches[match_class] << match
        end
      end
    end
  end

  new_matches
end

.recurse_alternatives(match, options) ⇒ Object 



232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
 
# File 'lib/natural/fragment.rb', line 232

def self.recurse_alternatives(match, options)
  new_matches = options[:matches]

  match.replacements(options).each do |replacement|
    new_matches[match_class] = [] if !new_matches[match.class]
    new_matches[match.class] << replacement

    unused_alternatives = ObjectSpace.each_object(Class).select {|a| a < Natural::Alternative}
    replacement.breadth_each {|node| unused_alternatives -= [node.class]}

    unused_alternatives.each do |alternative|
      next_layer = alternative.find(options.merge(:text => replacement.to_s, :matches => {})).values.first
      next_layer.each do |frag|
        new_frag = frag.class.new(:ids => replacement.ids, :text => frag.text)
        new_frag << replacement.clone
        new_matches = recurse_alternatives(new_frag, options)
      end
    end
  end

  new_matches
end

Instance Method Details

#all_filters ⇒ Object 



26
27
28
29
30
31
32
33
34
 
# File 'lib/natural/fragment.rb', line 26

def all_filters
  if self.is_leaf?
    self.filter
  else
    self.children.inject('') do |result, item| 
      result = [result, self.filter, item.all_filters].select{|a| !a.blank?}.uniq.join('.')
    end
  end
end

#clone(height = nil) ⇒ Object 



78
79
80
81
82
83
84
85
86
 
# File 'lib/natural/fragment.rb', line 78

def clone(height=nil)
  result = self.class.new(:ids => self.ids, :text => self.text)
  if !height || height > 0
    self.children.each do |child|
      result << child.clone(height ? height-1 : nil)
    end
  end
  result
end

#data(context = nil) ⇒ Object 



70
71
72
 
# File 'lib/natural/fragment.rb', line 70

def data(context=nil)
  nil
end

#id_range ⇒ Object 



22
23
24
 
# File 'lib/natural/fragment.rb', line 22

def id_range
  @ids.size > 1 ? @ids.first..@ids.last : @ids.first
end

#ids ⇒ Object

recurse to the leaves and print out the id range of the underyling words



14
15
16
 
# File 'lib/natural/fragment.rb', line 14

def ids
  self.is_leaf? ? [@ids].flatten : self.children.inject([]) {|result, item| result += item.ids}
end

#ids=(values) ⇒ Object 



18
19
20
 
# File 'lib/natural/fragment.rb', line 18

def ids=(values)
  @ids = values
end

#pretty_to_s(level = 0) ⇒ Object 



51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
 
# File 'lib/natural/fragment.rb', line 51

def pretty_to_s(level=0)
  result = ''
  
  if is_root? || level == 0
    result += '*'
  else
    result += "|" unless parent.is_last_sibling?
    result += (' ' * ((level - 1) * 4 + (parent.is_last_sibling? ? 0 : -1)))
    result += is_last_sibling? ? '+' : '|'
    result += '---'
    result += has_children? ? '+' : '>'
  end

  result += " #{name}\n"
  children.each {|child| result += child.pretty_to_s(level+1)}
  
  result
end

#to_s(options = {}) ⇒ Object

recurse to the leaves and print out all the words, applying all edits along the way



37
38
39
40
41
42
43
44
45
46
47
48
49
 
# File 'lib/natural/fragment.rb', line 37

def to_s(options={})
  if self.is_leaf?
    location = self
    if !options[:without_edits]
      while !location.is_root? && location.parent.class < Natural::Alternative do
        location = location.parent
      end
    end
    location.text
  else
    self.children.inject('') {|result, item| (result += item.to_s + ' ')}.strip
  end
end