langusta 0.1.1 → 0.2.0
Sign up to get free protection for your applications and to get access to all the features.
- data/.travis.yml +7 -0
- data/Gemfile +10 -7
- data/Gemfile.lock +12 -16
- data/{README.rdoc → README.md} +27 -10
- data/Rakefile +3 -10
- data/VERSION +1 -1
- data/langusta.gemspec +23 -47
- data/lib/langusta.rb +36 -10
- data/lib/langusta/codepoints.rb +19 -0
- data/lib/langusta/command.rb +3 -3
- data/lib/langusta/detector.rb +16 -13
- data/lib/langusta/detector_factory.rb +11 -5
- data/lib/langusta/guard.rb +22 -0
- data/lib/langusta/inspector.rb +7 -0
- data/lib/langusta/java_property_reader.rb +2 -3
- data/lib/langusta/lang_profile.rb +12 -18
- data/lib/langusta/language_detection_facade.rb +2 -2
- data/lib/langusta/n_gram.rb +20 -25
- data/lib/langusta/regex_helper.rb +15 -10
- data/lib/langusta/tag_extractor.rb +5 -5
- data/lib/langusta/unicode_block.rb +34 -34
- data/test/helper.rb +12 -3
- data/test/quality/test_falsified.rb +3 -3
- data/test/test_command.rb +1 -0
- data/test/test_detector.rb +18 -17
- data/test/test_detector_factory.rb +17 -5
- data/test/test_java_property_reader.rb +2 -1
- data/test/test_lang_profile.rb +37 -31
- data/test/test_language.rb +1 -0
- data/test/test_language_detection_facade.rb +1 -1
- data/test/test_langusta.rb +6 -6
- data/test/test_n_gram.rb +87 -75
- data/test/test_tag_extractor.rb +19 -18
- data/test/test_unicode_block.rb +2 -1
- metadata +54 -156
- data/lib/langusta/ucs2_string.rb +0 -70
- data/test/test_ucs2_string.rb +0 -9
data/.travis.yml
ADDED
data/Gemfile
CHANGED
@@ -1,11 +1,14 @@
|
|
1
1
|
source :gemcutter
|
2
|
-
gem "oniguruma", "1.1.0"
|
3
|
-
gem "yajl-ruby", "0.8.2"
|
4
2
|
|
5
|
-
|
6
|
-
gem "
|
7
|
-
|
8
|
-
|
3
|
+
platform :mri_18 do
|
4
|
+
gem "oniguruma", "1.1.0"
|
5
|
+
end
|
6
|
+
|
7
|
+
gem "yajl-ruby", "0.8.2", :require => 'yajl'
|
8
|
+
|
9
|
+
gem "bundler"
|
10
|
+
gem "jeweler"
|
11
|
+
|
12
|
+
group :test do
|
9
13
|
gem "mocha"
|
10
|
-
gem "ruby-debug"
|
11
14
|
end
|
data/Gemfile.lock
CHANGED
@@ -1,32 +1,28 @@
|
|
1
1
|
GEM
|
2
2
|
remote: http://rubygems.org/
|
3
3
|
specs:
|
4
|
-
columnize (0.3.2)
|
5
4
|
git (1.2.5)
|
6
|
-
jeweler (1.
|
7
|
-
bundler (~> 1.0
|
5
|
+
jeweler (1.8.3)
|
6
|
+
bundler (~> 1.0)
|
8
7
|
git (>= 1.2.5)
|
9
8
|
rake
|
10
|
-
|
11
|
-
|
9
|
+
rdoc
|
10
|
+
json (1.6.5)
|
11
|
+
metaclass (0.0.1)
|
12
|
+
mocha (0.10.5)
|
13
|
+
metaclass (~> 0.0.1)
|
12
14
|
oniguruma (1.1.0)
|
13
|
-
rake (0.
|
14
|
-
|
15
|
-
|
16
|
-
columnize (>= 0.1)
|
17
|
-
ruby-debug-base (~> 0.10.4.0)
|
18
|
-
ruby-debug-base (0.10.4)
|
19
|
-
linecache (>= 0.3)
|
15
|
+
rake (0.9.2.2)
|
16
|
+
rdoc (3.12)
|
17
|
+
json (~> 1.4)
|
20
18
|
yajl-ruby (0.8.2)
|
21
19
|
|
22
20
|
PLATFORMS
|
23
21
|
ruby
|
24
22
|
|
25
23
|
DEPENDENCIES
|
26
|
-
bundler
|
27
|
-
jeweler
|
24
|
+
bundler
|
25
|
+
jeweler
|
28
26
|
mocha
|
29
27
|
oniguruma (= 1.1.0)
|
30
|
-
rcov
|
31
|
-
ruby-debug
|
32
28
|
yajl-ruby (= 0.8.2)
|
data/{README.rdoc → README.md}
RENAMED
@@ -1,25 +1,41 @@
|
|
1
|
-
|
1
|
+
# langusta
|
2
2
|
|
3
3
|
Langusta is a language detection library based on a method designed and implemented by Nakatani Shuyo. This work is almost a direct 1-to-1 port of the original Java library which can be found at: http://code.google.com/p/language-detection.
|
4
4
|
|
5
5
|
For more information about the method (naive bayesian classification), have a look at this presentation: http://www.slideshare.net/shuyo/language-detection-library-for-java. This implementation uses some resources from the original library, specifically the language profiles.
|
6
6
|
|
7
|
-
|
7
|
+
## Build status
|
8
8
|
|
9
|
-
|
9
|
+
[![Build Status](https://secure.travis-ci.org/jasiek/langusta.png?branch=master)](http://travis-ci.org/jasiek/langusta)
|
10
|
+
|
11
|
+
## Runtime dependencies
|
12
|
+
|
13
|
+
* oniguruma - regular expressions swiss army knife (only required for 1.8.7)
|
10
14
|
* yajl-ruby - a quick and elegant JSON parser
|
11
15
|
|
12
|
-
|
16
|
+
## Usage
|
17
|
+
|
18
|
+
The simplest way to use this library is to use the facade provided with this package.
|
13
19
|
|
14
|
-
|
20
|
+
```ruby
|
21
|
+
require 'langusta'
|
22
|
+
facade = Langusta::LanguageDetectionFacade.new
|
23
|
+
facade.detect('zażółć gęślą jaźń') #=> 'pl'
|
24
|
+
```
|
15
25
|
|
16
|
-
|
26
|
+
If you don't need all 49 profiles, you can boost your detection speed and reduce memory consumption by writing your own facade-like class.
|
27
|
+
|
28
|
+
## Compatibility
|
17
29
|
|
18
30
|
* Ruby 1.8.7
|
31
|
+
* Ruby 1.9.2
|
32
|
+
* Ruby 1.9.3
|
33
|
+
|
34
|
+
## Caveats
|
19
35
|
|
20
|
-
|
36
|
+
Langusta is a memory hog - 49 profiles will take up about 80MB of RAM.
|
21
37
|
|
22
|
-
|
38
|
+
## Contributing to langusta
|
23
39
|
|
24
40
|
* Check out the latest master to make sure the feature hasn't been implemented or the bug hasn't been fixed yet
|
25
41
|
* Check out the issue tracker to make sure someone already hasn't requested it and/or contributed it
|
@@ -29,6 +45,7 @@ A version for Ruby 1.9 is in the works.
|
|
29
45
|
* Make sure to add tests for it. This is important so I don't break it in a future version unintentionally.
|
30
46
|
* Please try not to mess with the Rakefile, version, or history. If you want to have your own version, or is otherwise necessary, that is fine, but please isolate to its own commit so I can cherry-pick around it.
|
31
47
|
|
32
|
-
|
48
|
+
## Copyright
|
49
|
+
|
50
|
+
Copyright (c) 2011, 2012 Jan Szumiec. See LICENSE.txt for further details.
|
33
51
|
|
34
|
-
Copyright (c) 2011 Jan Szumiec. See LICENSE.txt for further details.
|
data/Rakefile
CHANGED
@@ -24,21 +24,14 @@ Jeweler::RubygemsDotOrgTasks.new
|
|
24
24
|
|
25
25
|
require 'rake/testtask'
|
26
26
|
Rake::TestTask.new(:test) do |test|
|
27
|
-
test.libs << 'lib' << 'test'
|
28
|
-
test.pattern = 'test/test_*.rb'
|
29
|
-
test.verbose = true
|
30
|
-
end
|
31
|
-
|
32
|
-
require 'rcov/rcovtask'
|
33
|
-
Rcov::RcovTask.new do |test|
|
34
|
-
test.libs << 'test'
|
27
|
+
test.libs << 'lib' << 'test' << '.'
|
35
28
|
test.pattern = 'test/test_*.rb'
|
36
29
|
test.verbose = true
|
37
30
|
end
|
38
31
|
|
39
32
|
task :default => :test
|
40
33
|
|
41
|
-
require '
|
34
|
+
require 'rdoc/task'
|
42
35
|
Rake::RDocTask.new do |rdoc|
|
43
36
|
version = File.exist?('VERSION') ? File.read('VERSION') : ""
|
44
37
|
|
@@ -49,7 +42,7 @@ Rake::RDocTask.new do |rdoc|
|
|
49
42
|
end
|
50
43
|
|
51
44
|
Rake::TestTask.new('test:quality') do |test|
|
52
|
-
test.libs << 'test/quality'
|
45
|
+
test.libs << 'test/quality' << 'lib' << '.'
|
53
46
|
test.pattern = 'test/quality/test_*.rb'
|
54
47
|
test.verbose = true
|
55
48
|
end
|
data/VERSION
CHANGED
@@ -1 +1 @@
|
|
1
|
-
0.
|
1
|
+
0.2.0
|
data/langusta.gemspec
CHANGED
@@ -4,26 +4,26 @@
|
|
4
4
|
# -*- encoding: utf-8 -*-
|
5
5
|
|
6
6
|
Gem::Specification.new do |s|
|
7
|
-
s.name =
|
8
|
-
s.version = "0.
|
7
|
+
s.name = "langusta"
|
8
|
+
s.version = "0.2.0"
|
9
9
|
|
10
10
|
s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
|
11
11
|
s.authors = ["Jan Szumiec"]
|
12
|
-
s.date =
|
13
|
-
s.
|
14
|
-
s.
|
15
|
-
s.email = %q{jan.szumiec@gmail.com}
|
12
|
+
s.date = "2012-03-04"
|
13
|
+
s.description = "Highly accurate language detection library, uses naive bayesian filter."
|
14
|
+
s.email = "jan.szumiec@gmail.com"
|
16
15
|
s.executables = ["langusta"]
|
17
16
|
s.extra_rdoc_files = [
|
18
17
|
"LICENSE.txt",
|
19
|
-
"README.
|
18
|
+
"README.md"
|
20
19
|
]
|
21
20
|
s.files = [
|
22
21
|
".document",
|
22
|
+
".travis.yml",
|
23
23
|
"Gemfile",
|
24
24
|
"Gemfile.lock",
|
25
25
|
"LICENSE.txt",
|
26
|
-
"README.
|
26
|
+
"README.md",
|
27
27
|
"Rakefile",
|
28
28
|
"VERSION",
|
29
29
|
"bin/langusta",
|
@@ -31,9 +31,12 @@ Gem::Specification.new do |s|
|
|
31
31
|
"data/uppercase.bin",
|
32
32
|
"langusta.gemspec",
|
33
33
|
"lib/langusta.rb",
|
34
|
+
"lib/langusta/codepoints.rb",
|
34
35
|
"lib/langusta/command.rb",
|
35
36
|
"lib/langusta/detector.rb",
|
36
37
|
"lib/langusta/detector_factory.rb",
|
38
|
+
"lib/langusta/guard.rb",
|
39
|
+
"lib/langusta/inspector.rb",
|
37
40
|
"lib/langusta/java_property_reader.rb",
|
38
41
|
"lib/langusta/lang_profile.rb",
|
39
42
|
"lib/langusta/language.rb",
|
@@ -41,7 +44,6 @@ Gem::Specification.new do |s|
|
|
41
44
|
"lib/langusta/n_gram.rb",
|
42
45
|
"lib/langusta/regex_helper.rb",
|
43
46
|
"lib/langusta/tag_extractor.rb",
|
44
|
-
"lib/langusta/ucs2_string.rb",
|
45
47
|
"lib/langusta/unicode_block.rb",
|
46
48
|
"profiles/af",
|
47
49
|
"profiles/ar",
|
@@ -152,59 +154,33 @@ Gem::Specification.new do |s|
|
|
152
154
|
"test/test_langusta.rb",
|
153
155
|
"test/test_n_gram.rb",
|
154
156
|
"test/test_tag_extractor.rb",
|
155
|
-
"test/test_ucs2_string.rb",
|
156
157
|
"test/test_unicode_block.rb"
|
157
158
|
]
|
158
|
-
s.homepage =
|
159
|
+
s.homepage = "http://github.com/jasiek/langusta"
|
159
160
|
s.licenses = ["Apache 2.0"]
|
160
161
|
s.require_paths = ["lib"]
|
161
|
-
s.rubygems_version =
|
162
|
-
s.summary =
|
163
|
-
s.test_files = [
|
164
|
-
"test/helper.rb",
|
165
|
-
"test/quality/test_falsified.rb",
|
166
|
-
"test/test_command.rb",
|
167
|
-
"test/test_detector.rb",
|
168
|
-
"test/test_detector_factory.rb",
|
169
|
-
"test/test_java_property_reader.rb",
|
170
|
-
"test/test_lang_profile.rb",
|
171
|
-
"test/test_language.rb",
|
172
|
-
"test/test_language_detection_facade.rb",
|
173
|
-
"test/test_langusta.rb",
|
174
|
-
"test/test_n_gram.rb",
|
175
|
-
"test/test_tag_extractor.rb",
|
176
|
-
"test/test_ucs2_string.rb",
|
177
|
-
"test/test_unicode_block.rb"
|
178
|
-
]
|
162
|
+
s.rubygems_version = "1.8.17"
|
163
|
+
s.summary = "Language detection library based on http://code.google.com/p/language-detection/."
|
179
164
|
|
180
165
|
if s.respond_to? :specification_version then
|
181
166
|
s.specification_version = 3
|
182
167
|
|
183
168
|
if Gem::Version.new(Gem::VERSION) >= Gem::Version.new('1.2.0') then
|
184
|
-
s.add_runtime_dependency(%q<oniguruma>, ["= 1.1.0"])
|
169
|
+
s.add_runtime_dependency(%q<oniguruma>, ["= 1.1.0"]) if RUBY_VERSION < "1.9"
|
185
170
|
s.add_runtime_dependency(%q<yajl-ruby>, ["= 0.8.2"])
|
186
|
-
s.
|
187
|
-
s.
|
188
|
-
s.add_development_dependency(%q<rcov>, [">= 0"])
|
189
|
-
s.add_development_dependency(%q<mocha>, [">= 0"])
|
190
|
-
s.add_development_dependency(%q<ruby-debug>, [">= 0"])
|
171
|
+
s.add_runtime_dependency(%q<bundler>, [">= 0"])
|
172
|
+
s.add_runtime_dependency(%q<jeweler>, [">= 0"])
|
191
173
|
else
|
192
|
-
s.add_dependency(%q<oniguruma>, ["= 1.1.0"])
|
174
|
+
s.add_dependency(%q<oniguruma>, ["= 1.1.0"]) if RUBY_VERSION < "1.9"
|
193
175
|
s.add_dependency(%q<yajl-ruby>, ["= 0.8.2"])
|
194
|
-
s.add_dependency(%q<bundler>, ["
|
195
|
-
s.add_dependency(%q<jeweler>, ["
|
196
|
-
s.add_dependency(%q<rcov>, [">= 0"])
|
197
|
-
s.add_dependency(%q<mocha>, [">= 0"])
|
198
|
-
s.add_dependency(%q<ruby-debug>, [">= 0"])
|
176
|
+
s.add_dependency(%q<bundler>, [">= 0"])
|
177
|
+
s.add_dependency(%q<jeweler>, [">= 0"])
|
199
178
|
end
|
200
179
|
else
|
201
|
-
s.add_dependency(%q<oniguruma>, ["= 1.1.0"])
|
180
|
+
s.add_dependency(%q<oniguruma>, ["= 1.1.0"]) if RUBY_VERSION < "1.9"
|
202
181
|
s.add_dependency(%q<yajl-ruby>, ["= 0.8.2"])
|
203
|
-
s.add_dependency(%q<bundler>, ["
|
204
|
-
s.add_dependency(%q<jeweler>, ["
|
205
|
-
s.add_dependency(%q<rcov>, [">= 0"])
|
206
|
-
s.add_dependency(%q<mocha>, [">= 0"])
|
207
|
-
s.add_dependency(%q<ruby-debug>, [">= 0"])
|
182
|
+
s.add_dependency(%q<bundler>, [">= 0"])
|
183
|
+
s.add_dependency(%q<jeweler>, [">= 0"])
|
208
184
|
end
|
209
185
|
end
|
210
186
|
|
data/lib/langusta.rb
CHANGED
@@ -2,20 +2,18 @@ $: << File.expand_path(File.dirname(__FILE__))
|
|
2
2
|
|
3
3
|
require 'rubygems'
|
4
4
|
require 'bundler'
|
5
|
-
Bundler.
|
5
|
+
Bundler.require
|
6
6
|
|
7
7
|
require 'optparse'
|
8
|
-
require 'iconv'
|
9
|
-
|
10
|
-
# Required gems
|
11
|
-
require 'oniguruma'
|
12
|
-
require 'yajl'
|
8
|
+
require 'iconv' if RUBY_VERSION < "1.9"
|
13
9
|
|
14
10
|
module Langusta
|
15
11
|
VERSION = '0.1.1'
|
16
12
|
|
13
|
+
autoload :Guard, 'langusta/guard'
|
14
|
+
autoload :Inspector, 'langusta/inspector'
|
17
15
|
autoload :RegexHelper, 'langusta/regex_helper'
|
18
|
-
autoload :
|
16
|
+
autoload :Codepoints, 'langusta/codepoints'
|
19
17
|
autoload :Language, 'langusta/language'
|
20
18
|
autoload :LangProfile, 'langusta/lang_profile'
|
21
19
|
autoload :Detector, 'langusta/detector'
|
@@ -33,8 +31,36 @@ module Langusta
|
|
33
31
|
UPPERCASE_BIN = File.join(ABSOLUTE_PATH, 'data/uppercase.bin')
|
34
32
|
MESSAGES_PROPERTIES = File.join(ABSOLUTE_PATH, 'data/messages.properties')
|
35
33
|
|
36
|
-
class
|
37
|
-
class
|
38
|
-
class
|
34
|
+
class Error < StandardError; end
|
35
|
+
class DuplicateProfilesError < Error; end
|
36
|
+
class NoProfilesLoadedError < Error; end
|
37
|
+
class NoFeaturesInTextError < Error; end
|
38
|
+
|
39
|
+
UTF82CP_SELECTOR = RUBY_VERSION < "1.9" ? :utf82cp_18 : :utf82cp_19
|
40
|
+
CP2UTF8_SELECTOR = RUBY_VERSION < "1.9" ? :cp2utf8_18 : :cp2utf8_19
|
41
|
+
|
42
|
+
def self.utf82cp(utf8_string)
|
43
|
+
send(UTF82CP_SELECTOR, utf8_string)
|
44
|
+
end
|
45
|
+
|
46
|
+
def self.utf82cp_18(utf8_string)
|
47
|
+
Iconv.conv('ucs-2be', 'utf-8', utf8_string).unpack('n*')
|
48
|
+
end
|
49
|
+
|
50
|
+
def self.utf82cp_19(utf8_string)
|
51
|
+
utf8_string.encode('ucs-2be').unpack('n*')
|
52
|
+
end
|
53
|
+
|
54
|
+
def self.cp2utf8(cp_array)
|
55
|
+
send(CP2UTF8_SELECTOR, cp_array)
|
56
|
+
end
|
57
|
+
|
58
|
+
def self.cp2utf8_18(cp_array)
|
59
|
+
Iconv.conv('utf-8', 'ucs-2be', cp_array.pack('n*'))
|
60
|
+
end
|
61
|
+
|
62
|
+
def self.cp2utf8_19(cp_array)
|
63
|
+
cp_array.pack('n*').force_encoding('ucs-2be').encode('utf-8')
|
64
|
+
end
|
39
65
|
end
|
40
66
|
|
@@ -0,0 +1,19 @@
|
|
1
|
+
module Langusta
|
2
|
+
module Codepoints
|
3
|
+
GSUB_SELECTOR = RUBY_VERSION < "1.9" ? :gsub18 : :gsub19
|
4
|
+
|
5
|
+
def self.gsub!(codepoint_array, regex, replacement)
|
6
|
+
string = Langusta.cp2utf8(codepoint_array)
|
7
|
+
string = send(GSUB_SELECTOR, string, regex, replacement)
|
8
|
+
codepoint_array.replace(Langusta.utf82cp(string))
|
9
|
+
end
|
10
|
+
|
11
|
+
def self.gsub18(string, oregex, replacement)
|
12
|
+
oregex.gsub(string, replacement)
|
13
|
+
end
|
14
|
+
|
15
|
+
def self.gsub19(string, regex, replacement)
|
16
|
+
string.gsub(regex, replacement)
|
17
|
+
end
|
18
|
+
end
|
19
|
+
end
|
data/lib/langusta/command.rb
CHANGED
@@ -55,7 +55,7 @@ EOF
|
|
55
55
|
end
|
56
56
|
|
57
57
|
def detect_single_lang(filename, alpha)
|
58
|
-
ucs2_content =
|
58
|
+
ucs2_content = Langusta.utf82cp(File.open(filename).read)
|
59
59
|
detector = @detector_factory.create(alpha)
|
60
60
|
detector.append(ucs2_content)
|
61
61
|
|
@@ -64,8 +64,8 @@ EOF
|
|
64
64
|
|
65
65
|
def initialize_factory(profile_directory)
|
66
66
|
profiles = load_profiles(profile_directory)
|
67
|
-
profiles.
|
68
|
-
@detector_factory.add_profile(profile
|
67
|
+
profiles.each do |profile|
|
68
|
+
@detector_factory.add_profile(profile)
|
69
69
|
end
|
70
70
|
end
|
71
71
|
|
data/lib/langusta/detector.rb
CHANGED
@@ -13,7 +13,7 @@ module Langusta
|
|
13
13
|
def initialize(factory)
|
14
14
|
@word_lang_prob_map = factory.word_lang_prob_map
|
15
15
|
@lang_list = factory.lang_list
|
16
|
-
@text =
|
16
|
+
@text = []
|
17
17
|
@langprob = nil
|
18
18
|
@alpha = ALPHA_DEFAULT
|
19
19
|
@n_trial = 7
|
@@ -25,13 +25,15 @@ module Langusta
|
|
25
25
|
# Append more text to be recognized.
|
26
26
|
# @param text [UCS2String] text to be recognized
|
27
27
|
def append(text)
|
28
|
-
|
29
|
-
|
30
|
-
text.gsub!(RegexHelper::
|
28
|
+
Guard.klass(text, Array, __method__)
|
29
|
+
|
30
|
+
text = Codepoints.gsub!(text, RegexHelper::URL_REGEX, "\x00\x20")
|
31
|
+
text = Codepoints.gsub!(text, RegexHelper::MAIL_REGEX, "\x00\x20")
|
32
|
+
|
31
33
|
text = text.map do |c|
|
32
34
|
NGram.normalize(c)
|
33
35
|
end
|
34
|
-
@text =
|
36
|
+
@text = Codepoints.gsub!(text, RegexHelper::SPACE_REGEX, "\x00\x20")
|
35
37
|
end
|
36
38
|
|
37
39
|
# Detect the language.
|
@@ -102,17 +104,17 @@ module Langusta
|
|
102
104
|
|
103
105
|
def cleaning_text
|
104
106
|
non_latin_count = latin_count = 0
|
105
|
-
@text.
|
106
|
-
if c < "
|
107
|
+
@text.each do |c|
|
108
|
+
if c < 0x007a && c > 0x0041 # c > "z" && c < "A"
|
107
109
|
latin_count += 1
|
108
|
-
elsif c >=
|
110
|
+
elsif c >= 0x3000 && UnicodeBlock.of(c) != UnicodeBlock::LATIN_EXTENDED_ADDITIONAL
|
109
111
|
non_latin_count += 1
|
110
112
|
end
|
111
113
|
end
|
112
114
|
if latin_count * 2 < non_latin_count
|
113
|
-
text_without_latin =
|
114
|
-
@text.
|
115
|
-
text_without_latin << c if c > "
|
115
|
+
text_without_latin = []
|
116
|
+
@text.each do |c|
|
117
|
+
text_without_latin << c if c > 0x007a || c < 0x0041 # c > "z" || c < "A"
|
116
118
|
end
|
117
119
|
@text = text_without_latin
|
118
120
|
end
|
@@ -121,7 +123,7 @@ module Langusta
|
|
121
123
|
def extract_ngrams
|
122
124
|
list = []
|
123
125
|
ngram = NGram.new
|
124
|
-
@text.
|
126
|
+
@text.each do |char|
|
125
127
|
ngram.add_char(char)
|
126
128
|
(1..NGram::N_GRAM).each do |n|
|
127
129
|
w = ngram.get(n)
|
@@ -169,7 +171,8 @@ module Langusta
|
|
169
171
|
# verbose
|
170
172
|
weight = alpha / BASE_FREQ
|
171
173
|
prob.length.times do |i|
|
172
|
-
|
174
|
+
# tiny workaround for nil values in word freq array
|
175
|
+
prob[i] *= weight + (lang_prob_map[i] || 0.0)
|
173
176
|
end
|
174
177
|
true
|
175
178
|
end
|