Class: Cass::Contrast

Inherits:
Object
  • Object
show all
Defined in:
lib/cass/contrast.rb

Overview

Defines a single contrast on a document or documents. Currently, only comparisons between two pairs of words (i.e., computation of the interaction term; see Holtzman et al., submitted for details) is implemented.

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(words) ⇒ Contrast 



10
11
12
13
 
# File 'lib/cass/contrast.rb', line 10

def initialize(words)
  words = words.split(/,*\s+/) if words.class == String
  @words = words
end

Instance Attribute Details

#words ⇒ Object

Returns the value of attribute words.



8
9
10
 
# File 'lib/cass/contrast.rb', line 8

def words
  @words
end

Class Method Details

.parse(contrast) ⇒ Object

Initialize a contrast from a string representation.



16
17
18
19
20
 
# File 'lib/cass/contrast.rb', line 16

def self.parse(contrast)
  words = contrast.split(/,*\s+/)
  words = [words[0,2], words[2,2]] if (words.size == 4)
  Contrast.new(words)
end

Instance Method Details

#apply(doc) ⇒ Object

Apply the contrast to a document and return result as a string. See to_s method for format.



24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
 
# File 'lib/cass/contrast.rb', line 24

def apply(doc)
  sim = doc.similarity(@words.flatten)
  if sim.class == Array
    puts "Error: #{doc.name} is missing the following operands: #{sim.join(", ")}"
    return false
  end
  # Compute interaction term
  if @words.flatten.size == 4
    s = [doc.name, sim[0,2], sim[0,3], sim[1,2], sim[1,3]]
    s[1,4] = s[1,4].map { |f| format("%0.4f", f).to_f }
    s << s[1] - s[2] - s[3] + s[4]
    s.join("\t")
  else
    [doc.name, sim[0,1]].join("\t")
  end
end

#to_s ⇒ Object

Returns a string representation of the contrast. For interaction terms, returns five columns (the four pairs and the interaction term). For pairwise contrasts, just returns the similarity.



45
46
47
48
49
50
51
52
 
# File 'lib/cass/contrast.rb', line 45

def to_s
  if @words.flatten.size == 4
    o = @words.flatten
    "document\t#{o[0]}.#{o[2]}\t#{o[0]}.#{o[3]}\t#{o[1]}.#{o[2]}\t#{o[1]}.#{o[3]}\tinteraction"
  else
    "document\t#{@words.join(".")}"
  end
end