-
Notifications
You must be signed in to change notification settings - Fork 49
/
tokenizer.rb
231 lines (206 loc) · 7.28 KB
/
tokenizer.rb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
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
231
# encoding: utf-8
#
module Picky
# Defines tokenizing processes used both in indexing and querying.
#
class Tokenizer
def self.index_default= new_default
@index_default = new_default
end
def self.index_default
@index_default ||= new
end
def self.query_default= new_default
@query_default = new_default
end
def self.query_default
@query_default ||= new
end
# TODO Move EMPTY_STRING top level.
#
EMPTY_STRING = ''.freeze
def to_s
reject_condition_location = @reject_condition.to_s[/:(\d+) \(lambda\)/, 1]
<<-TOKENIZER
Removes characters: #{@removes_characters_regexp ? "/#{@removes_characters_regexp.source}/" : '-'}
Stopwords: #{@remove_stopwords_regexp ? "/#{@remove_stopwords_regexp.source}/" : '-'}
Splits text on: #{@splits_text_on.respond_to?(:source) ? "/#{@splits_text_on.source}/" : (@splits_text_on ? @splits_text_on : '-')}
Normalizes words: #{@normalizes_words_regexp_replaces ? @normalizes_words_regexp_replaces : '-'}
Rejects tokens? #{reject_condition_location ? "Yes, see line #{reject_condition_location} in app/application.rb" : '-'}
Substitutes chars? #{@substituter ? "Yes, using #{@substituter}." : '-' }
Case sensitive? #{@case_sensitive ? "Yes." : "-"}
TOKENIZER
end
# Stopwords.
#
# We only allow regexps (even if string would be okay
# too for gsub! - it's too hard to understand)
#
def stopwords regexp
check_argument_in __method__, Regexp, regexp
@remove_stopwords_regexp = regexp
end
def remove_stopwords text
text.gsub! @remove_stopwords_regexp, EMPTY_STRING if @remove_stopwords_regexp
text
end
@@non_single_stopword_regexp = /^\b[\w:]+?\b[\.\*\~]?\s?$/
def remove_non_single_stopwords text
return text unless @remove_stopwords_regexp
return text if text.match @@non_single_stopword_regexp
remove_stopwords text
end
# Illegals.
#
# We only allow regexps (even if string would be okay
# too for gsub! - it's too hard to understand)
#
def removes_characters regexp
check_argument_in __method__, Regexp, regexp
@removes_characters_regexp = regexp
end
def remove_illegals text
text.gsub! @removes_characters_regexp, EMPTY_STRING if @removes_characters_regexp
text
end
# Splitting.
#
# We allow Strings and Regexps.
# Note: We do not test against to_str since symbols do not work with String#split.
#
def splits_text_on regexp_or_string
raise ArgumentError.new "#{__method__} takes a Regexp or String as argument, not a #{regexp_or_string.class}." unless Regexp === regexp_or_string || String === regexp_or_string
@splits_text_on = regexp_or_string
end
def split text
text.split @splits_text_on
end
# Normalizing.
#
# We only allow arrays.
#
def normalizes_words regexp_replaces
raise ArgumentError.new "#{__method__} takes an Array of replaces as argument, not a #{regexp_replaces.class}." unless regexp_replaces.respond_to?(:to_ary)
@normalizes_words_regexp_replaces = regexp_replaces
end
def normalize_with_patterns text
return text unless @normalizes_words_regexp_replaces
@normalizes_words_regexp_replaces.each do |regex, replace|
# This should be sufficient
#
text.gsub!(regex, replace) and break
end
text
end
def normalize_with_patterns?
@normalizes_words_regexp_replaces
end
# Substitute Characters with this substituter.
#
# Default is European Character substitution.
#
def substitutes_characters_with substituter = CharacterSubstituters::WestEuropean.new
raise ArgumentError.new "The substitutes_characters_with option needs a character substituter, which responds to #substitute." unless substituter.respond_to?(:substitute)
@substituter = substituter
end
def substitute_characters text
substituter?? substituter.substitute(text) : text
end
# Reject tokens after tokenizing based on the given criteria.
#
def rejects_token_if &condition
@reject_condition = condition
end
def reject tokens
tokens.reject! &@reject_condition
end
def case_sensitive case_sensitive
@case_sensitive = case_sensitive
end
def downcase?
!@case_sensitive
end
def maximum_tokens amount
@maximum_tokens = amount
end
def cap words
words.slice!(@maximum_tokens..-1) if cap?(words)
end
def cap? words
@maximum_tokens && words.size > @maximum_tokens
end
# Checks if the right argument type has been given.
#
def check_argument_in method, type, argument, &condition
raise ArgumentError.new "Application##{method} takes a #{type} as argument, not a #{argument.class}." unless type === argument
end
attr_reader :substituter
alias substituter? substituter
def initialize options = {}
substitutes_characters_with options[:substitutes_characters_with] if options[:substitutes_characters_with]
removes_characters options[:removes_characters] if options[:removes_characters]
stopwords options[:stopwords] if options[:stopwords]
splits_text_on options[:splits_text_on] || /\s/
normalizes_words options[:normalizes_words] if options[:normalizes_words]
maximum_tokens options[:maximum_tokens]
rejects_token_if &(options[:rejects_token_if] || :blank?)
case_sensitive options[:case_sensitive] unless options[:case_sensitive].nil?
end
# Returns a number of tokens, generated from the given text,
# based on the parameters given.
#
# Returns:
# [[:token1, :token2], ["Original1", "Original2"]]
#
def tokenize text
text = preprocess text.to_s # processing the text
return empty_tokens if text.blank?
words = pretokenize text # splitting and preparations for tokenizing
return empty_tokens if words.empty?
tokens = tokens_for words # creating tokens / strings
[tokens, words]
end
# Default preprocessing hook.
#
# Does:
# 1. Character substitution.
# 2. Remove illegal expressions.
# 3. Remove non-single stopwords. (Stopwords that occur with other words)
#
def preprocess text
text = substitute_characters text
remove_illegals text
# We do not remove single stopwords e.g. in the indexer for
# an entirely different reason than in the query tokenizer.
# An indexed thing with just name "UND" (a possible stopword)
# should not lose its name.
#
remove_non_single_stopwords text
text
end
# Pretokenizing.
#
# Does:
# * Split the text into words.
# * Cap the amount of tokens if maximum_tokens is set.
#
def pretokenize text
words = split text
words.collect! { |word| normalize_with_patterns word } if normalize_with_patterns?
reject words
cap words if cap?(words)
words
end
# Downcases.
#
def tokens_for words
words.collect! { |word| word.downcase!; word } if downcase?
words
end
# Returns empty tokens.
#
def empty_tokens
[[], []]
end
end
end