wikipedia-vandalism_detection 0.1.0-java

Sign up to get free protection for your applications and to get access to all the features.
Files changed (245) hide show
  1. checksums.yaml +7 -0
  2. data/.gitignore +19 -0
  3. data/.rspec +2 -0
  4. data/.rubocop.yml +35 -0
  5. data/.travis.yml +6 -0
  6. data/Gemfile +4 -0
  7. data/LICENSE.txt +4 -0
  8. data/README.md +288 -0
  9. data/Rakefile +11 -0
  10. data/config/wikipedia-vandalism-detection.yml.example +103 -0
  11. data/lib/java/SMOTE.jar +0 -0
  12. data/lib/java/balancedRandomForest.jar +0 -0
  13. data/lib/java/diffutils-1.3.0.jar +0 -0
  14. data/lib/java/oneClassClassifier.jar +0 -0
  15. data/lib/java/realAdaBoost.jar +0 -0
  16. data/lib/java/swc-engine-1.1.0-jar-with-dependencies.jar +0 -0
  17. data/lib/java/sweble-wikitext-extractor.jar +0 -0
  18. data/lib/weka/classifiers/meta/one_class_classifier.rb +21 -0
  19. data/lib/weka/classifiers/meta/real_ada_boost.rb +15 -0
  20. data/lib/weka/classifiers/trees/balanced_random_forest.rb +16 -0
  21. data/lib/weka/filters/supervised/instance/smote.rb +15 -0
  22. data/lib/wikipedia/vandalism_detection/algorithms/kullback_leibler_divergence.rb +103 -0
  23. data/lib/wikipedia/vandalism_detection/algorithms.rb +12 -0
  24. data/lib/wikipedia/vandalism_detection/classifier.rb +202 -0
  25. data/lib/wikipedia/vandalism_detection/configuration.rb +350 -0
  26. data/lib/wikipedia/vandalism_detection/diff.rb +36 -0
  27. data/lib/wikipedia/vandalism_detection/edit.rb +81 -0
  28. data/lib/wikipedia/vandalism_detection/evaluator.rb +640 -0
  29. data/lib/wikipedia/vandalism_detection/exceptions.rb +47 -0
  30. data/lib/wikipedia/vandalism_detection/feature_calculator.rb +94 -0
  31. data/lib/wikipedia/vandalism_detection/features/all_wordlists_frequency.rb +22 -0
  32. data/lib/wikipedia/vandalism_detection/features/all_wordlists_impact.rb +21 -0
  33. data/lib/wikipedia/vandalism_detection/features/anonymity.rb +17 -0
  34. data/lib/wikipedia/vandalism_detection/features/anonymity_previous.rb +32 -0
  35. data/lib/wikipedia/vandalism_detection/features/article_size.rb +17 -0
  36. data/lib/wikipedia/vandalism_detection/features/bad_frequency.rb +21 -0
  37. data/lib/wikipedia/vandalism_detection/features/bad_impact.rb +20 -0
  38. data/lib/wikipedia/vandalism_detection/features/base.rb +61 -0
  39. data/lib/wikipedia/vandalism_detection/features/biased_frequency.rb +21 -0
  40. data/lib/wikipedia/vandalism_detection/features/biased_impact.rb +21 -0
  41. data/lib/wikipedia/vandalism_detection/features/blanking.rb +26 -0
  42. data/lib/wikipedia/vandalism_detection/features/character_diversity.rb +23 -0
  43. data/lib/wikipedia/vandalism_detection/features/character_sequence.rb +19 -0
  44. data/lib/wikipedia/vandalism_detection/features/comment_bad_frequency.rb +21 -0
  45. data/lib/wikipedia/vandalism_detection/features/comment_biased_frequency.rb +21 -0
  46. data/lib/wikipedia/vandalism_detection/features/comment_length.rb +16 -0
  47. data/lib/wikipedia/vandalism_detection/features/comment_markup_frequency.rb +30 -0
  48. data/lib/wikipedia/vandalism_detection/features/comment_pronoun_frequency.rb +21 -0
  49. data/lib/wikipedia/vandalism_detection/features/comment_sex_frequency.rb +21 -0
  50. data/lib/wikipedia/vandalism_detection/features/comment_vulgarism_frequency.rb +21 -0
  51. data/lib/wikipedia/vandalism_detection/features/compressibility.rb +31 -0
  52. data/lib/wikipedia/vandalism_detection/features/contains_base.rb +16 -0
  53. data/lib/wikipedia/vandalism_detection/features/copyedit.rb +17 -0
  54. data/lib/wikipedia/vandalism_detection/features/digit_ratio.rb +23 -0
  55. data/lib/wikipedia/vandalism_detection/features/edits_per_user.rb +72 -0
  56. data/lib/wikipedia/vandalism_detection/features/emoticons_frequency.rb +27 -0
  57. data/lib/wikipedia/vandalism_detection/features/emoticons_impact.rb +29 -0
  58. data/lib/wikipedia/vandalism_detection/features/frequency_base.rb +18 -0
  59. data/lib/wikipedia/vandalism_detection/features/impact_base.rb +25 -0
  60. data/lib/wikipedia/vandalism_detection/features/inserted_character_distribution.rb +23 -0
  61. data/lib/wikipedia/vandalism_detection/features/inserted_external_links.rb +19 -0
  62. data/lib/wikipedia/vandalism_detection/features/inserted_internal_links.rb +18 -0
  63. data/lib/wikipedia/vandalism_detection/features/inserted_size.rb +19 -0
  64. data/lib/wikipedia/vandalism_detection/features/inserted_words.rb +17 -0
  65. data/lib/wikipedia/vandalism_detection/features/longest_word.rb +21 -0
  66. data/lib/wikipedia/vandalism_detection/features/markup_frequency.rb +29 -0
  67. data/lib/wikipedia/vandalism_detection/features/markup_impact.rb +30 -0
  68. data/lib/wikipedia/vandalism_detection/features/non_alphanumeric_ratio.rb +23 -0
  69. data/lib/wikipedia/vandalism_detection/features/personal_life.rb +17 -0
  70. data/lib/wikipedia/vandalism_detection/features/pronoun_frequency.rb +21 -0
  71. data/lib/wikipedia/vandalism_detection/features/pronoun_impact.rb +20 -0
  72. data/lib/wikipedia/vandalism_detection/features/removed_all_wordlists_frequency.rb +22 -0
  73. data/lib/wikipedia/vandalism_detection/features/removed_bad_frequency.rb +21 -0
  74. data/lib/wikipedia/vandalism_detection/features/removed_biased_frequency.rb +21 -0
  75. data/lib/wikipedia/vandalism_detection/features/removed_character_distribution.rb +23 -0
  76. data/lib/wikipedia/vandalism_detection/features/removed_emoticons_frequency.rb +28 -0
  77. data/lib/wikipedia/vandalism_detection/features/removed_markup_frequency.rb +30 -0
  78. data/lib/wikipedia/vandalism_detection/features/removed_pronoun_frequency.rb +21 -0
  79. data/lib/wikipedia/vandalism_detection/features/removed_sex_frequency.rb +21 -0
  80. data/lib/wikipedia/vandalism_detection/features/removed_size.rb +19 -0
  81. data/lib/wikipedia/vandalism_detection/features/removed_vulgarism_frequency.rb +21 -0
  82. data/lib/wikipedia/vandalism_detection/features/removed_words.rb +17 -0
  83. data/lib/wikipedia/vandalism_detection/features/replacement_similarity.rb +20 -0
  84. data/lib/wikipedia/vandalism_detection/features/reverted.rb +16 -0
  85. data/lib/wikipedia/vandalism_detection/features/revisions_character_distribution.rb +25 -0
  86. data/lib/wikipedia/vandalism_detection/features/same_editor.rb +32 -0
  87. data/lib/wikipedia/vandalism_detection/features/sex_frequency.rb +21 -0
  88. data/lib/wikipedia/vandalism_detection/features/sex_impact.rb +20 -0
  89. data/lib/wikipedia/vandalism_detection/features/size_increment.rb +20 -0
  90. data/lib/wikipedia/vandalism_detection/features/size_ratio.rb +28 -0
  91. data/lib/wikipedia/vandalism_detection/features/term_frequency.rb +31 -0
  92. data/lib/wikipedia/vandalism_detection/features/time_interval.rb +38 -0
  93. data/lib/wikipedia/vandalism_detection/features/time_of_day.rb +21 -0
  94. data/lib/wikipedia/vandalism_detection/features/upper_case_ratio.rb +23 -0
  95. data/lib/wikipedia/vandalism_detection/features/upper_case_words_ratio.rb +33 -0
  96. data/lib/wikipedia/vandalism_detection/features/upper_to_lower_case_ratio.rb +23 -0
  97. data/lib/wikipedia/vandalism_detection/features/vulgarism_frequency.rb +21 -0
  98. data/lib/wikipedia/vandalism_detection/features/vulgarism_impact.rb +21 -0
  99. data/lib/wikipedia/vandalism_detection/features/weekday.rb +19 -0
  100. data/lib/wikipedia/vandalism_detection/features/words_increment.rb +20 -0
  101. data/lib/wikipedia/vandalism_detection/features.rb +66 -0
  102. data/lib/wikipedia/vandalism_detection/instances.rb +121 -0
  103. data/lib/wikipedia/vandalism_detection/page.rb +101 -0
  104. data/lib/wikipedia/vandalism_detection/page_parser.rb +63 -0
  105. data/lib/wikipedia/vandalism_detection/revision.rb +75 -0
  106. data/lib/wikipedia/vandalism_detection/revision_parser.rb +67 -0
  107. data/lib/wikipedia/vandalism_detection/test_dataset.rb +374 -0
  108. data/lib/wikipedia/vandalism_detection/text.rb +23 -0
  109. data/lib/wikipedia/vandalism_detection/training_dataset.rb +282 -0
  110. data/lib/wikipedia/vandalism_detection/version.rb +5 -0
  111. data/lib/wikipedia/vandalism_detection/wikitext_extractor.rb +80 -0
  112. data/lib/wikipedia/vandalism_detection/word_lists/bad.rb +11 -0
  113. data/lib/wikipedia/vandalism_detection/word_lists/biased.rb +20 -0
  114. data/lib/wikipedia/vandalism_detection/word_lists/emoticons.rb +26 -0
  115. data/lib/wikipedia/vandalism_detection/word_lists/markup.rb +19 -0
  116. data/lib/wikipedia/vandalism_detection/word_lists/pronouns.rb +11 -0
  117. data/lib/wikipedia/vandalism_detection/word_lists/sex.rb +10 -0
  118. data/lib/wikipedia/vandalism_detection/word_lists/vulgarism.rb +96 -0
  119. data/lib/wikipedia/vandalism_detection/word_lists.rb +17 -0
  120. data/lib/wikipedia/vandalism_detection.rb +29 -0
  121. data/lib/wikipedia.rb +41 -0
  122. data/spec/factories/edit.rb +19 -0
  123. data/spec/factories/page.rb +12 -0
  124. data/spec/factories/revision.rb +51 -0
  125. data/spec/resources/config/wikipedia-vandalism-detection.yml +35 -0
  126. data/spec/resources/corpora/test/edits.csv +8 -0
  127. data/spec/resources/corpora/test/ground-truth.txt +3 -0
  128. data/spec/resources/corpora/test/revisions/part-1/326471754.txt +199 -0
  129. data/spec/resources/corpora/test/revisions/part-1/326873205.txt +137 -0
  130. data/spec/resources/corpora/test/revisions/part-1/328774035.txt +162 -0
  131. data/spec/resources/corpora/test/revisions/part-2/307084144.txt +137 -0
  132. data/spec/resources/corpora/test/revisions/part-2/326978767.txt +199 -0
  133. data/spec/resources/corpora/test/revisions/part-2/328774110.txt +162 -0
  134. data/spec/resources/corpora/training/annotations.csv +7 -0
  135. data/spec/resources/corpora/training/edits.csv +7 -0
  136. data/spec/resources/corpora/training/revisions/part-1/326471754.txt +199 -0
  137. data/spec/resources/corpora/training/revisions/part-1/326873205.txt +137 -0
  138. data/spec/resources/corpora/training/revisions/part-1/328774035.txt +162 -0
  139. data/spec/resources/corpora/training/revisions/part-2/307084144.txt +137 -0
  140. data/spec/resources/corpora/training/revisions/part-2/326978767.txt +199 -0
  141. data/spec/resources/corpora/training/revisions/part-2/328774110.txt +162 -0
  142. data/spec/resources/page_with_redirects.xml +85 -0
  143. data/spec/resources/redirect_page.xml +59 -0
  144. data/spec/resources/revision_simplified.xml +13 -0
  145. data/spec/resources/sample_revision.txt +137 -0
  146. data/spec/resources/sample_revision_clean_text.txt +1 -0
  147. data/spec/resources/sample_revision_plain_text.txt +183 -0
  148. data/spec/resources/vandalism_on_wikipedia.xml +234 -0
  149. data/spec/resources/vandalism_on_wikipedia_simplified.xml +119 -0
  150. data/spec/resources/wikipedia_tokens.txt +30 -0
  151. data/spec/spec_helper.rb +38 -0
  152. data/spec/support/macros/file_reading.rb +6 -0
  153. data/spec/support/macros/test_configuration.rb +81 -0
  154. data/spec/vandalism_detection/algorithms/kullback_leibler_divergence_spec.rb +34 -0
  155. data/spec/vandalism_detection/classifier_spec.rb +330 -0
  156. data/spec/vandalism_detection/configuration_spec.rb +601 -0
  157. data/spec/vandalism_detection/diff_spec.rb +40 -0
  158. data/spec/vandalism_detection/edit_spec.rb +122 -0
  159. data/spec/vandalism_detection/evaluator_spec.rb +711 -0
  160. data/spec/vandalism_detection/feature_calculator_spec.rb +135 -0
  161. data/spec/vandalism_detection/features/all_wordlists_frequency_spec.rb +30 -0
  162. data/spec/vandalism_detection/features/all_wordlists_impact_spec.rb +52 -0
  163. data/spec/vandalism_detection/features/anonymity_previous_spec.rb +67 -0
  164. data/spec/vandalism_detection/features/anonymity_spec.rb +17 -0
  165. data/spec/vandalism_detection/features/article_size_spec.rb +29 -0
  166. data/spec/vandalism_detection/features/bad_frequency_spec.rb +30 -0
  167. data/spec/vandalism_detection/features/bad_impact_spec.rb +53 -0
  168. data/spec/vandalism_detection/features/base_spec.rb +41 -0
  169. data/spec/vandalism_detection/features/biased_frequency_spec.rb +30 -0
  170. data/spec/vandalism_detection/features/biased_impact_spec.rb +52 -0
  171. data/spec/vandalism_detection/features/blanking_spec.rb +35 -0
  172. data/spec/vandalism_detection/features/character_diversity_spec.rb +30 -0
  173. data/spec/vandalism_detection/features/character_sequence_spec.rb +31 -0
  174. data/spec/vandalism_detection/features/comment_bad_frequency_spec.rb +28 -0
  175. data/spec/vandalism_detection/features/comment_biased_frequency_spec.rb +28 -0
  176. data/spec/vandalism_detection/features/comment_length_spec.rb +21 -0
  177. data/spec/vandalism_detection/features/comment_markup_frequency_spec.rb +28 -0
  178. data/spec/vandalism_detection/features/comment_pronoun_frequency_spec.rb +28 -0
  179. data/spec/vandalism_detection/features/comment_sex_frequency_spec.rb +28 -0
  180. data/spec/vandalism_detection/features/comment_vulgarism_frequency_spec.rb +28 -0
  181. data/spec/vandalism_detection/features/compressibility_spec.rb +38 -0
  182. data/spec/vandalism_detection/features/contains_base_spec.rb +27 -0
  183. data/spec/vandalism_detection/features/copyedit_spec.rb +30 -0
  184. data/spec/vandalism_detection/features/digit_ratio_spec.rb +30 -0
  185. data/spec/vandalism_detection/features/edits_per_user_spec.rb +51 -0
  186. data/spec/vandalism_detection/features/emoticons_frequency_spec.rb +30 -0
  187. data/spec/vandalism_detection/features/emoticons_impact_spec.rb +45 -0
  188. data/spec/vandalism_detection/features/frequency_base_spec.rb +20 -0
  189. data/spec/vandalism_detection/features/impact_base_spec.rb +36 -0
  190. data/spec/vandalism_detection/features/inserted_character_distribution_spec.rb +40 -0
  191. data/spec/vandalism_detection/features/inserted_external_links_spec.rb +29 -0
  192. data/spec/vandalism_detection/features/inserted_internal_links_spec.rb +29 -0
  193. data/spec/vandalism_detection/features/inserted_size_spec.rb +29 -0
  194. data/spec/vandalism_detection/features/inserted_words_spec.rb +29 -0
  195. data/spec/vandalism_detection/features/longest_word_spec.rb +29 -0
  196. data/spec/vandalism_detection/features/markup_frequency_spec.rb +31 -0
  197. data/spec/vandalism_detection/features/markup_impact_spec.rb +53 -0
  198. data/spec/vandalism_detection/features/non_alphanumeric_ratio_spec.rb +30 -0
  199. data/spec/vandalism_detection/features/personal_life_spec.rb +22 -0
  200. data/spec/vandalism_detection/features/pronoun_frequency_spec.rb +30 -0
  201. data/spec/vandalism_detection/features/pronoun_impact_spec.rb +53 -0
  202. data/spec/vandalism_detection/features/removed_all_wordlists_frequency_spec.rb +30 -0
  203. data/spec/vandalism_detection/features/removed_bad_frequency_spec.rb +30 -0
  204. data/spec/vandalism_detection/features/removed_biased_frequency_spec.rb +30 -0
  205. data/spec/vandalism_detection/features/removed_character_distribution_spec.rb +40 -0
  206. data/spec/vandalism_detection/features/removed_emoticons_frequency_spec.rb +30 -0
  207. data/spec/vandalism_detection/features/removed_markup_frequency_spec.rb +30 -0
  208. data/spec/vandalism_detection/features/removed_pronoun_frequency_spec.rb +30 -0
  209. data/spec/vandalism_detection/features/removed_sex_frequency_spec.rb +30 -0
  210. data/spec/vandalism_detection/features/removed_size_spec.rb +29 -0
  211. data/spec/vandalism_detection/features/removed_vulgarism_frequency_spec.rb +30 -0
  212. data/spec/vandalism_detection/features/removed_words_spec.rb +29 -0
  213. data/spec/vandalism_detection/features/replacement_similarity_spec.rb +37 -0
  214. data/spec/vandalism_detection/features/reverted_spec.rb +24 -0
  215. data/spec/vandalism_detection/features/revisions_character_distribution_spec.rb +40 -0
  216. data/spec/vandalism_detection/features/same_editor_spec.rb +71 -0
  217. data/spec/vandalism_detection/features/sex_frequency_spec.rb +30 -0
  218. data/spec/vandalism_detection/features/sex_impact_spec.rb +53 -0
  219. data/spec/vandalism_detection/features/size_increment_spec.rb +29 -0
  220. data/spec/vandalism_detection/features/size_ratio_spec.rb +48 -0
  221. data/spec/vandalism_detection/features/term_frequency_spec.rb +32 -0
  222. data/spec/vandalism_detection/features/time_interval_spec.rb +56 -0
  223. data/spec/vandalism_detection/features/time_of_day_spec.rb +16 -0
  224. data/spec/vandalism_detection/features/upper_case_ratio_spec.rb +30 -0
  225. data/spec/vandalism_detection/features/upper_case_words_ratio_spec.rb +33 -0
  226. data/spec/vandalism_detection/features/upper_to_lower_case_ratio_spec.rb +30 -0
  227. data/spec/vandalism_detection/features/vulgarism_frequency_spec.rb +30 -0
  228. data/spec/vandalism_detection/features/vulgarism_impact_spec.rb +52 -0
  229. data/spec/vandalism_detection/features/weekday_spec.rb +16 -0
  230. data/spec/vandalism_detection/features/words_increment_spec.rb +29 -0
  231. data/spec/vandalism_detection/instances_spec.rb +146 -0
  232. data/spec/vandalism_detection/page_parser_spec.rb +190 -0
  233. data/spec/vandalism_detection/page_spec.rb +134 -0
  234. data/spec/vandalism_detection/revision_parser_spec.rb +53 -0
  235. data/spec/vandalism_detection/revision_spec.rb +148 -0
  236. data/spec/vandalism_detection/test_dataset_spec.rb +227 -0
  237. data/spec/vandalism_detection/text_spec.rb +29 -0
  238. data/spec/vandalism_detection/training_dataset_spec.rb +266 -0
  239. data/spec/vandalism_detection/wikitext_extractor_spec.rb +97 -0
  240. data/spec/weka/classifiers/meta/one_class_classifier_spec.rb +82 -0
  241. data/spec/weka/classifiers/meta/real_ada_boost_spec.rb +37 -0
  242. data/spec/weka/classifiers/trees/balanced_random_forest_spec.rb +37 -0
  243. data/spec/weka/filters/supervised/instance/smote_spec.rb +5 -0
  244. data/wikipedia-vandalism_detection.gemspec +37 -0
  245. metadata +550 -0
@@ -0,0 +1,601 @@
1
+ require 'spec_helper'
2
+ require 'yaml'
3
+
4
+ describe Wikipedia::VandalismDetection do
5
+ DEFAULTS = Wikipedia::VandalismDetection::DefaultConfiguration::DEFAULTS
6
+
7
+ describe 'Configuration class' do
8
+ before do
9
+ allow_any_instance_of(Wikipedia::VandalismDetection::DefaultConfiguration)
10
+ .to receive(:source)
11
+ .and_return(source_dir)
12
+
13
+ @config = Wikipedia::VandalismDetection::Configuration.instance
14
+
15
+ use_test_configuration
16
+ end
17
+
18
+ %i[
19
+ data
20
+ features
21
+ classifier_type
22
+ classifier_options
23
+ cross_validation_fold
24
+ training_data_options
25
+ balanced_training_data?
26
+ unbalanced_training_data?
27
+ oversampled_training_data?
28
+ test_output_classification_file
29
+ oversampling_options
30
+ training_output_arff_file
31
+ test_output_arff_file
32
+ replace_training_data_missing_values?
33
+ ].each do |attribute|
34
+ it "responds to ##{attribute}" do
35
+ expect(@config).to respond_to attribute
36
+ end
37
+ end
38
+
39
+ it 'returns a hash for #data (the full config hash)' do
40
+ expect(@config.data).to be_a Hash
41
+ end
42
+
43
+ it 'returns a feature array for #feature' do
44
+ expect(@config.features).to be_an Array
45
+ end
46
+
47
+ it 'returns a numeric for #cross-validation-fold' do
48
+ expect(@config.cross_validation_fold).to be_a Numeric
49
+ end
50
+
51
+ describe '#test_output_classification_file' do
52
+ it 'returns the classification file path extended by classifier name and training data options' do
53
+ file_path = @config.test_output_classification_file
54
+ classifier_name = @config.classifier_type.split('::').last.downcase
55
+ dataset_options = @config.training_data_options
56
+ file_name = DEFAULTS['output']['test']['classification_file']
57
+
58
+ path = File.join(
59
+ @config.output_base_directory,
60
+ classifier_name,
61
+ dataset_options,
62
+ file_name
63
+ )
64
+
65
+ expect(file_path).to eq path
66
+ end
67
+ end
68
+
69
+ describe 'output arff files' do
70
+ describe '#training_output_arff_file' do
71
+ it 'returns the arff file path extended by classifier name and training data options' do
72
+ file_path = @config.training_output_arff_file
73
+ classifier_name = @config.classifier_type.split('::').last.downcase
74
+ dataset_options = @config.training_data_options
75
+ file_name = DEFAULTS['output']['training']['arff_file']
76
+
77
+ path = File.join(
78
+ @config.output_base_directory,
79
+ classifier_name,
80
+ dataset_options,
81
+ file_name
82
+ )
83
+
84
+ expect(file_path).to eq path
85
+ end
86
+ end
87
+
88
+ describe '#test_output_arff_file' do
89
+ it 'returns the arff file path extended by classifier name and training data options' do
90
+ file_path = @config.test_output_arff_file
91
+ classifier_name = @config.classifier_type.split('::').last.downcase
92
+ dataset_options = @config.training_data_options
93
+ file_name = DEFAULTS['output']['test']['arff_file']
94
+
95
+ path = File.join(
96
+ @config.output_base_directory,
97
+ classifier_name,
98
+ dataset_options,
99
+ file_name
100
+ )
101
+
102
+ expect(file_path).to eq path
103
+ end
104
+ end
105
+ end
106
+
107
+ describe '#use_occ?' do
108
+ it 'returns true if used classifier is a one class classifier' do
109
+ config = test_config
110
+ classifier_type = Weka::Classifiers::Meta::OneClassClassifier.type
111
+ config.instance_variable_set(:@classifier_type, classifier_type)
112
+ use_configuration(config)
113
+ use_occ = Wikipedia::VandalismDetection.config.use_occ?
114
+
115
+ expect(use_occ).to be true
116
+ end
117
+
118
+ it 'returns false if used classifier isn’t a one class classifier' do
119
+ use_test_configuration
120
+ use_occ = Wikipedia::VandalismDetection.config.use_occ?
121
+
122
+ expect(use_occ).to be false
123
+ end
124
+ end
125
+
126
+ describe '#balanced_training_data?' do
127
+ context 'if it is set in config' do
128
+ it 'returns true' do
129
+ config = test_config
130
+ config.instance_variable_set(:@training_data_options, 'balanced')
131
+ use_configuration(config)
132
+ balanced = Wikipedia::VandalismDetection.config.balanced_training_data?
133
+
134
+ expect(balanced).to be true
135
+ end
136
+ end
137
+
138
+ context 'if it is not set in config' do
139
+ it 'returns false ' do
140
+ config = test_config
141
+ config.instance_variable_set(:@training_data_options, nil)
142
+ use_configuration(config)
143
+ balanced = Wikipedia::VandalismDetection.config.balanced_training_data?
144
+
145
+ expect(balanced).to be false
146
+ end
147
+ end
148
+
149
+ context 'if it is set to other value than "balanced"' do
150
+ it 'returns false' do
151
+ config = test_config
152
+ config.instance_variable_set(:@training_data_options, 'other')
153
+ use_configuration(config)
154
+ balanced = Wikipedia::VandalismDetection.config.balanced_training_data?
155
+
156
+ expect(balanced).to be false
157
+ end
158
+ end
159
+ end
160
+
161
+ describe '#unbalanced_training_data?' do
162
+ context 'if it is set in config' do
163
+ it 'returns true' do
164
+ config = test_config
165
+ config.instance_variable_set(:@training_data_options, 'unbalanced')
166
+ use_configuration(config)
167
+ unbalanced = Wikipedia::VandalismDetection.config.unbalanced_training_data?
168
+
169
+ expect(unbalanced).to be true
170
+ end
171
+ end
172
+
173
+ context 'if it is not set in config' do
174
+ it 'returns true' do
175
+ config = test_config
176
+ config.instance_variable_set(:@training_data_options, nil)
177
+ use_configuration(config)
178
+ unbalanced = Wikipedia::VandalismDetection.config.unbalanced_training_data?
179
+
180
+ expect(unbalanced).to be true
181
+ end
182
+ end
183
+
184
+ context 'if set to other value than "unbalanced" or "oversampled"' do
185
+ it 'returns true' do
186
+ config = test_config
187
+ config.instance_variable_set(:@training_data_options, 'other value')
188
+ use_configuration(config)
189
+ unbalanced = Wikipedia::VandalismDetection.config.unbalanced_training_data?
190
+
191
+ expect(unbalanced).to be true
192
+ end
193
+ end
194
+
195
+ context 'if it is set to other value than "unbalanced"' do
196
+ it 'returns false' do
197
+ config = test_config
198
+ config.instance_variable_set(:@training_data_options, 'balanced')
199
+ use_configuration(config)
200
+ unbalanced = Wikipedia::VandalismDetection.config.unbalanced_training_data?
201
+
202
+ expect(unbalanced).to be false
203
+ end
204
+ end
205
+ end
206
+
207
+ describe '#oversampled_training_data?' do
208
+ context 'if it is set in config' do
209
+ it 'returns true' do
210
+ config = test_config
211
+ config.instance_variable_set(:@training_data_options, 'oversampled')
212
+ use_configuration(config)
213
+ oversampled = Wikipedia::VandalismDetection.config.oversampled_training_data?
214
+
215
+ expect(oversampled).to be true
216
+ end
217
+ end
218
+
219
+ context 'if it is not set in config' do
220
+ it 'returns false' do
221
+ config = test_config
222
+ config.instance_variable_set(:@training_data_options, nil)
223
+ use_configuration(config)
224
+ oversampled = Wikipedia::VandalismDetection.config.oversampled_training_data?
225
+
226
+ expect(oversampled).to be false
227
+ end
228
+ end
229
+
230
+ context 'if it is set to other value than "balanced"' do
231
+ it 'returns false' do
232
+ config = test_config
233
+ config.instance_variable_set(:@training_data_options, 'other')
234
+ use_configuration(config)
235
+ oversampled = Wikipedia::VandalismDetection.config.oversampled_training_data?
236
+
237
+ expect(oversampled).to be false
238
+ end
239
+ end
240
+ end
241
+
242
+ describe '#oversampled_options' do
243
+ it 'returns a hash' do
244
+ expect(@config.oversampling_options).to be_a Hash
245
+ end
246
+
247
+ it 'returns a hash with the :percent and :undersampling keys' do
248
+ config = test_config
249
+ config.instance_variable_set(:@training_data_options, 'oversampled')
250
+ use_configuration(config)
251
+
252
+ options = Wikipedia::VandalismDetection.config.oversampling_options
253
+ hash = { percentage: 0, undersampling: true }
254
+
255
+ expect(options.keys).to eq hash.keys
256
+ end
257
+
258
+ context 'if training data is not oversampled' do
259
+ it 'returns an empty hash' do
260
+ config = test_config
261
+ config.instance_variable_set(:@training_data_options, 'other')
262
+ use_configuration(config)
263
+ options = Wikipedia::VandalismDetection.config.oversampling_options
264
+
265
+ expect(options).to eq({})
266
+ end
267
+ end
268
+
269
+ describe 'Returning of configured options' do
270
+ let(:percentage) { 300.0 }
271
+ let(:undersampling) { 200.0 }
272
+
273
+ let(:sampling_options) do
274
+ {
275
+ percentage: percentage,
276
+ undersampling: undersampling
277
+ }
278
+ end
279
+
280
+ it 'returns the configured options with downcase params' do
281
+ config = test_config
282
+ options = "oversampled -p #{percentage} -u true #{undersampling}"
283
+ config.instance_variable_set(:@training_data_options, options)
284
+ use_configuration(config)
285
+ hash = Wikipedia::VandalismDetection.config.oversampling_options
286
+
287
+ expect(hash).to eq sampling_options
288
+ end
289
+
290
+ it 'returns the configured options with upcase params' do
291
+ config = test_config
292
+ options = "oversampled -P #{percentage} -U true #{undersampling}"
293
+ config.instance_variable_set(:@training_data_options, options)
294
+ use_configuration(config)
295
+ hash = Wikipedia::VandalismDetection.config.oversampling_options
296
+
297
+ expect(hash).to eq sampling_options
298
+ end
299
+
300
+ it 'returns the configured options with full params' do
301
+ config = test_config
302
+ options = "oversampled -Percentage #{percentage} -Undersampling true #{undersampling}"
303
+ config.instance_variable_set(:@training_data_options, options)
304
+ use_configuration(config)
305
+ hash = Wikipedia::VandalismDetection.config.oversampling_options
306
+
307
+ expect(hash).to eq sampling_options
308
+ end
309
+ end
310
+
311
+ it 'returns a default value for percent if not set' do
312
+ percentage = 100 # default value
313
+ undersampling = 200
314
+ sampling_options = {
315
+ percentage: percentage,
316
+ undersampling: undersampling
317
+ }
318
+
319
+ config = test_config
320
+ options = "oversampled -u true #{undersampling}"
321
+ config.instance_variable_set(:@training_data_options, options)
322
+ use_configuration(config)
323
+ hash = Wikipedia::VandalismDetection.config.oversampling_options
324
+
325
+ expect(hash).to eq sampling_options
326
+ end
327
+
328
+ it 'returns a default true for undersampling if not set' do
329
+ percentage = 200
330
+ undersampling = 100 # default value
331
+ sampling_options = {
332
+ percentage: percentage,
333
+ undersampling: undersampling
334
+ }
335
+
336
+ config = test_config
337
+ options = "oversampled -p #{percentage}"
338
+ config.instance_variable_set(:@training_data_options, options)
339
+ use_configuration(config)
340
+ hash = Wikipedia::VandalismDetection.config.oversampling_options
341
+
342
+ expect(hash).to eq sampling_options
343
+ end
344
+
345
+ it 'returns a percentange value for undersampling if set in -u option' do
346
+ percentage = 200
347
+ undersampling = 0.001
348
+ sampling_options = {
349
+ percentage: percentage,
350
+ undersampling: undersampling
351
+ }
352
+
353
+ config = test_config
354
+ options = "oversampled -p #{percentage} -u true #{undersampling}"
355
+ config.instance_variable_set(:@training_data_options, options)
356
+ use_configuration(config)
357
+ hash = Wikipedia::VandalismDetection.config.oversampling_options
358
+
359
+ expect(hash).to eq sampling_options
360
+ end
361
+ end
362
+
363
+ describe '#replace_missing_values?' do
364
+ ['no', 'No', 'false', 'nope', '', nil].each do |option|
365
+ it 'returns false if not set' do
366
+ config = test_config
367
+ config.instance_variable_set(:@replace_missing_values, option)
368
+ use_configuration(config)
369
+ replace = Wikipedia::VandalismDetection.config.replace_training_data_missing_values?
370
+
371
+ expect(replace).to be false
372
+ end
373
+ end
374
+
375
+ %w[yes t T YES True true].each do |option|
376
+ it 'returns true if set' do
377
+ config = test_config
378
+ config.instance_variable_set(:@replace_missing_values, option)
379
+ use_configuration(config)
380
+ replace = Wikipedia::VandalismDetection.config.replace_training_data_missing_values?
381
+
382
+ expect(replace).to be true
383
+ end
384
+ end
385
+ end
386
+
387
+ %i[
388
+ training_corpus_edits_file
389
+ training_corpus_annotations_file
390
+ training_corpus_revisions_directory
391
+ test_corpus_edits_file
392
+ test_corpus_revisions_directory
393
+ test_corpus_ground_truth_file
394
+ training_output_arff_file
395
+ training_output_index_file
396
+ test_output_arff_file
397
+ test_output_index_file
398
+ classifier_type
399
+ classifier_options
400
+ output_base_directory
401
+ ].each do |attribute|
402
+ it "returns a string when calling ##{attribute}" do
403
+ expect(@config.send(attribute)).to be_a String
404
+ end
405
+ end
406
+ end
407
+
408
+ describe '#configuration' do
409
+ it 'can be overridden by a wikipedia-vandalism-detection.yml file' do
410
+ allow_any_instance_of(Wikipedia::VandalismDetection::DefaultConfiguration)
411
+ .to receive(:source)
412
+ .and_return(source_dir)
413
+
414
+ default_config = DEFAULTS
415
+ config_file = '../../resources/config/wikipedia-vandalism-detection.yml'
416
+ custom_config = YAML.load_file(File.expand_path(config_file, __FILE__))
417
+
418
+ expect(Wikipedia::VandalismDetection.config.data)
419
+ .to eq default_config.deep_merge(custom_config)
420
+ end
421
+
422
+ it 'returns a Wikipedia::VandalismDetection::Configuration' do
423
+ expect(Wikipedia::VandalismDetection.config)
424
+ .to be_a Wikipedia::VandalismDetection::Configuration
425
+ end
426
+
427
+ it 'has all features as default configuration' do
428
+ use_default_configuration
429
+
430
+ features = [
431
+ 'anonymity',
432
+ 'anonymity previous',
433
+ 'all wordlists frequency',
434
+ 'all wordlists impact',
435
+ 'article size',
436
+ 'bad frequency',
437
+ 'bad impact',
438
+ 'biased frequency',
439
+ 'biased impact',
440
+ 'blanking',
441
+ 'character sequence',
442
+ 'character diversity',
443
+ 'comment length',
444
+ 'comment biased frequency',
445
+ 'comment pronoun frequency',
446
+ 'comment vulgarism frequency',
447
+ 'compressibility',
448
+ 'copyedit',
449
+ 'digit ratio',
450
+ 'edits per user',
451
+ 'emoticons frequency',
452
+ 'emoticons impact',
453
+ 'inserted size',
454
+ 'inserted words',
455
+ 'inserted character distribution',
456
+ 'inserted external links',
457
+ 'inserted internal links',
458
+ 'longest word',
459
+ 'markup frequency',
460
+ 'markup impact',
461
+ 'non-alphanumeric ratio',
462
+ 'personal life',
463
+ 'pronoun frequency',
464
+ 'pronoun impact',
465
+ 'removed size',
466
+ 'removed words',
467
+ 'removed all wordlists frequency',
468
+ 'removed bad frequency',
469
+ 'removed biased frequency',
470
+ 'removed character distribution',
471
+ 'removed emoticons frequency',
472
+ 'removed markup frequency',
473
+ 'removed pronoun frequency',
474
+ 'removed sex frequency',
475
+ 'removed vulgarism frequency',
476
+ 'replacement similarity',
477
+ 'reverted',
478
+ 'revisions character distribution',
479
+ 'sex frequency',
480
+ 'sex impact',
481
+ 'same editor',
482
+ 'size increment',
483
+ 'size ratio',
484
+ 'term frequency',
485
+ 'time interval',
486
+ 'time of day',
487
+ 'upper case ratio',
488
+ 'upper case words ratio',
489
+ 'upper to lower case ratio',
490
+ 'vulgarism frequency',
491
+ 'vulgarism impact',
492
+ 'weekday',
493
+ 'words increment'
494
+ ]
495
+
496
+ configured_features = Wikipedia::VandalismDetection.config['features']
497
+
498
+ expect(configured_features).to eq features
499
+ end
500
+
501
+ describe '#configuration#corpora' do
502
+ before do
503
+ use_default_configuration
504
+ @corpora = Wikipedia::VandalismDetection.config['corpora']
505
+ end
506
+
507
+ it 'has a corpora config' do
508
+ expect(@corpora).to be_a Hash
509
+ end
510
+
511
+ %i[training test].each do |attribute|
512
+ it "has a #{attribute}-corpus config" do
513
+ expect(@corpora[attribute.to_s]).to be_a Hash
514
+ end
515
+ end
516
+
517
+ it 'has a default nil corpora-base_directory config' do
518
+ expect(@corpora['base_directory']).to be_nil
519
+ end
520
+
521
+ %i[
522
+ base_directory
523
+ revisions_directory
524
+ edits_file
525
+ annotations_file
526
+ ].each do |attribute|
527
+ it "has an default nil '#{attribute}' config for the training-corpus" do
528
+ expect(@corpora['training'][attribute.to_s]).to be_nil
529
+ end
530
+ end
531
+
532
+ %i[
533
+ base_directory
534
+ revisions_directory
535
+ edits_file
536
+ ].each do |attribute|
537
+ it "has an default nil '#{attribute}' config for the test-corpus" do
538
+ expect(@corpora['test'][attribute.to_s]).to be_nil
539
+ end
540
+ end
541
+ end
542
+
543
+ describe '#configuration#output' do
544
+ before do
545
+ use_default_configuration
546
+ @output_config = Wikipedia::VandalismDetection.config['output']
547
+ end
548
+
549
+ it 'has an output-config' do
550
+ expect(@output_config).to be_a Hash
551
+ end
552
+
553
+ describe 'output sub configs' do
554
+ it 'has a default "base_directory" output-config' do
555
+ expect(@output_config['base_directory']).to_not be_nil
556
+ end
557
+
558
+ %i[arff_file index_file].each do |attribute|
559
+ it "has a default '#{attribute}' config for the training-output" do
560
+ output_training_config = @output_config['training']
561
+ expect(output_training_config).to_not be_nil
562
+ expect(output_training_config[attribute.to_s]).to_not be_nil
563
+ end
564
+ end
565
+
566
+ %i[arff_file index_file].each do |attribute|
567
+ it "has a default '#{attribute}' config for the test-output" do
568
+ output_test_config = @output_config['training']
569
+ expect(output_test_config).to_not be nil
570
+ expect(output_test_config[attribute.to_s]).to_not be nil
571
+ end
572
+ end
573
+ end
574
+ end
575
+
576
+ describe '#configuration#classifier' do
577
+ before do
578
+ use_default_configuration
579
+ @classifier = Wikipedia::VandalismDetection.config['classifier']
580
+ end
581
+
582
+ it 'return a classifier Hash' do
583
+ expect(@classifier).to be_a Hash
584
+ end
585
+
586
+ %i[type options].each do |attribute|
587
+ it "has a default nil '#{attribute}' config for classification" do
588
+ expect(@classifier[attribute.to_s]).to be_nil
589
+ end
590
+ end
591
+
592
+ it 'has a default 10-fold CV config for evaluation' do
593
+ expect(@classifier['cross-validation-fold']).to eq 10
594
+ end
595
+
596
+ it 'has has unbalances training data by default for training' do
597
+ expect(@classifier['training-data-options']).to eq 'unbalanced'
598
+ end
599
+ end
600
+ end
601
+ end
@@ -0,0 +1,40 @@
1
+ require 'spec_helper'
2
+
3
+ describe Wikipedia::VandalismDetection::Diff do
4
+ let(:old_text) { Text.new("hello\nworld\nmy name is Luke\n") }
5
+ let(:new_text) { Text.new("world\nhello\nmy name is Mr. Skywalker\n") }
6
+ let(:diff) { Wikipedia::VandalismDetection::Diff.new(old_text, new_text) }
7
+
8
+ it 'can deal with invalid byte sequences' do
9
+ text = "text \255".force_encoding('UTF-8')
10
+ diff = Wikipedia::VandalismDetection::Diff.new("#{text} a", "#{text} b")
11
+
12
+ expect(diff).to be_a Wikipedia::VandalismDetection::Diff
13
+ end
14
+
15
+ describe '#inserted_words' do
16
+ let(:words) { diff.inserted_words }
17
+
18
+ it 'returns the inserted words as array' do
19
+ expect(words).to be_an Array
20
+ expect(words.count).to eq 3
21
+ end
22
+
23
+ it 'returns the right inserted words' do
24
+ expect(words).to eq %w[hello Mr. Skywalker]
25
+ end
26
+ end
27
+
28
+ describe '#removed_words' do
29
+ let(:words) { diff.removed_words }
30
+
31
+ it 'returns the removed words as array' do
32
+ expect(words).to be_an Array
33
+ expect(words.count).to eq 2
34
+ end
35
+
36
+ it 'returns the right removed words' do
37
+ expect(words).to eq %w[hello Luke]
38
+ end
39
+ end
40
+ end