simplificator-babel 0.1.0 → 0.2.0
Sign up to get free protection for your applications and to get access to all the features.
- data/README.markdown +16 -8
- data/Rakefile +20 -2
- data/VERSION.yml +1 -1
- data/babel.gemspec +2 -2
- data/lib/babel/babel.rb +0 -1
- data/lib/babel/profile.rb +43 -36
- data/lib/babel/string_extensions.rb +2 -1
- data/lib/profiles/profile_deu.yml +13431 -11455
- data/lib/profiles/profile_eng.yml +12586 -8998
- data/lib/profiles/profile_fra.yml +13437 -10092
- data/lib/profiles/profile_ita.yml +13388 -9489
- data/lib/profiles/profile_spa.yml +13367 -9617
- data/samples/guessing.rb +3 -8
- data/test/babel_test.rb +22 -2
- data/test/profile_test.rb +26 -15
- metadata +2 -2
data/samples/guessing.rb
CHANGED
@@ -9,20 +9,15 @@ def show_languages(s)
|
|
9
9
|
puts "Possible Languages for '#{s}': \n#{s.languages[0..2].join(', ')}"
|
10
10
|
end
|
11
11
|
|
12
|
-
en =
|
12
|
+
en = 'Rainy sunday'
|
13
13
|
de = 'Regnerischer Sonntag'
|
14
14
|
fr = 'Je ne regrette rien'
|
15
15
|
sp = 'yo no soy un marinero'
|
16
16
|
|
17
|
-
texts = [en, de, fr, sp]
|
17
|
+
texts = [en, de, fr, sp, 'lago di como']
|
18
18
|
Babel.load_profiles
|
19
19
|
|
20
20
|
texts.each do |text|
|
21
21
|
guess_language text
|
22
22
|
show_languages text
|
23
|
-
end
|
24
|
-
|
25
|
-
|
26
|
-
|
27
|
-
|
28
|
-
|
23
|
+
end
|
data/test/babel_test.rb
CHANGED
@@ -1,13 +1,33 @@
|
|
1
1
|
require 'test_helper'
|
2
2
|
|
3
3
|
class BabelTest < Test::Unit::TestCase
|
4
|
-
|
4
|
+
should 'identify the language' do
|
5
|
+
Babel.load_profiles
|
6
|
+
assert_equal 'deu', 'ich gehe essen'.language
|
7
|
+
assert_equal 'deu', 'guten morgen'.language
|
8
|
+
assert_equal 'deu', 'öäü'.language # :-)
|
9
|
+
assert_equal 'deu', 'du bist der jackpot meines lebens'.language
|
10
|
+
|
11
|
+
assert_equal 'eng', 'imagine there\'s no heaven'.language
|
12
|
+
assert_equal 'eng', 'then there was you'.language
|
13
|
+
assert_equal 'eng', 'small step for me'.language
|
14
|
+
assert_equal 'eng', 'let there'.language
|
15
|
+
|
16
|
+
assert_equal 'fra', 'le coq est mort'.language
|
17
|
+
assert_equal 'fra', 'garçon'.language
|
18
|
+
assert_equal 'fra', 'je ne sais pas'.language
|
19
|
+
|
20
|
+
assert_equal 'spa', 'yo no soy un marinero'.language
|
21
|
+
assert_equal 'spa', 'mucho gusto'.language
|
22
|
+
assert_equal 'ita', 'lago di como'.language
|
23
|
+
end
|
5
24
|
|
6
25
|
should 'have a test' do
|
7
|
-
fail
|
26
|
+
#fail
|
8
27
|
end
|
9
28
|
|
10
29
|
should 'not choke when registering the first' do
|
30
|
+
#puts "abc".n_grams(:pad => true)
|
11
31
|
end
|
12
32
|
|
13
33
|
end
|
data/test/profile_test.rb
CHANGED
@@ -30,13 +30,13 @@ class ProfileTest < Test::Unit::TestCase
|
|
30
30
|
profile.occured('c', 5)
|
31
31
|
profile.occured('d', 3)
|
32
32
|
|
33
|
-
profile.
|
33
|
+
profile.weigh_and_rank
|
34
34
|
|
35
|
-
assert_equal 1, profile.
|
36
|
-
assert_equal 2, profile.
|
37
|
-
assert_equal 3, profile.
|
38
|
-
assert_equal 4, profile.
|
39
|
-
assert_equal 5, profile.
|
35
|
+
assert_equal 1, profile.rank('a')
|
36
|
+
assert_equal 2, profile.rank('b')
|
37
|
+
assert_equal 3, profile.rank('c')
|
38
|
+
assert_equal 4, profile.rank('d')
|
39
|
+
assert_equal 5, profile.rank('e')
|
40
40
|
end
|
41
41
|
|
42
42
|
should 'limit' do
|
@@ -47,7 +47,7 @@ class ProfileTest < Test::Unit::TestCase
|
|
47
47
|
profile.occured('d', 3)
|
48
48
|
profile.occured('e', 1)
|
49
49
|
|
50
|
-
profile.
|
50
|
+
profile.weigh_and_rank
|
51
51
|
profile.limit(3)
|
52
52
|
|
53
53
|
assert_equal 10, profile.occurence('a')
|
@@ -58,11 +58,13 @@ class ProfileTest < Test::Unit::TestCase
|
|
58
58
|
end
|
59
59
|
|
60
60
|
should 'find the distance to another profile' do
|
61
|
+
# distance is based on weigh_and_rank which is a floating point number
|
62
|
+
# hard to test. just make some comparisons with <>
|
61
63
|
a = Profile.new
|
62
64
|
b = Profile.new
|
63
65
|
|
64
66
|
a.occured('a')
|
65
|
-
a.
|
67
|
+
a.weigh_and_rank # always rank when finished with occured(). distance is baed on rank
|
66
68
|
|
67
69
|
assert_equal 0, a.distance(a), 'distance to self is 0'
|
68
70
|
assert_equal 1, a.distance(b)
|
@@ -70,7 +72,7 @@ class ProfileTest < Test::Unit::TestCase
|
|
70
72
|
assert_equal 0, b.distance(b)
|
71
73
|
|
72
74
|
b.occured('a')
|
73
|
-
b.
|
75
|
+
b.weigh_and_rank
|
74
76
|
|
75
77
|
assert_equal 0, a.distance(a)
|
76
78
|
assert_equal 0, a.distance(b)
|
@@ -79,15 +81,14 @@ class ProfileTest < Test::Unit::TestCase
|
|
79
81
|
|
80
82
|
a.occured('b')
|
81
83
|
a.occured('c')
|
82
|
-
a.
|
84
|
+
a.weigh_and_rank
|
83
85
|
|
84
|
-
|
86
|
+
assert 2 < a.distance(b)
|
85
87
|
|
86
88
|
a.occured('d')
|
87
|
-
a.
|
88
|
-
|
89
|
-
|
90
|
-
assert_equal 8, a.distance(b)
|
89
|
+
a.weigh_and_rank
|
90
|
+
|
91
|
+
assert 3 < a.distance(b)
|
91
92
|
end
|
92
93
|
|
93
94
|
should 'learn a text' do
|
@@ -102,4 +103,14 @@ class ProfileTest < Test::Unit::TestCase
|
|
102
103
|
assert_equal 1, profile.occurence('meme')
|
103
104
|
end
|
104
105
|
|
106
|
+
should 'clean a text' do
|
107
|
+
profile = Profile.new
|
108
|
+
assert_equal '', profile.clean_text('')
|
109
|
+
assert_equal ' ', profile.clean_text(' ')
|
110
|
+
assert_equal ' ', profile.clean_text(' ')
|
111
|
+
assert_equal 'hallo', profile.clean_text('1h2a3l4l5o')
|
112
|
+
assert_equal 'no braces and brackets and curly braces', profile.clean_text('(no ({braces) [] and brackets} and curly braces[]')
|
113
|
+
|
114
|
+
end
|
115
|
+
|
105
116
|
end
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: simplificator-babel
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.
|
4
|
+
version: 0.2.0
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- simplificator
|
@@ -9,7 +9,7 @@ autorequire:
|
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
11
|
|
12
|
-
date: 2009-07-
|
12
|
+
date: 2009-07-14 00:00:00 -07:00
|
13
13
|
default_executable:
|
14
14
|
dependencies:
|
15
15
|
- !ruby/object:Gem::Dependency
|