wikipedia-vandalism_detection 0.0.1
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.gitignore +17 -0
- data/Gemfile +6 -0
- data/LICENSE.txt +4 -0
- data/README.md +265 -0
- data/Rakefile +12 -0
- data/lib/java/LibSVM.jar +0 -0
- data/lib/java/SMOTE.jar +0 -0
- data/lib/java/balancedRandomForest.jar +0 -0
- data/lib/java/diffutils-1.3.0.jar +0 -0
- data/lib/java/libsvm.jar +0 -0
- data/lib/java/oneClassClassifier.jar +0 -0
- data/lib/java/realAdaBoost.jar +0 -0
- data/lib/java/swc-engine-1.1.0-jar-with-dependencies.jar +0 -0
- data/lib/java/sweble-wikitext-extractor.jar +0 -0
- data/lib/weka/classifiers/functions/lib_svm.rb +15 -0
- data/lib/weka/classifiers/meta/one_class_classifier.rb +25 -0
- data/lib/weka/classifiers/meta/real_ada_boost.rb +17 -0
- data/lib/weka/classifiers/trees/balanced_random_forest.rb +18 -0
- data/lib/weka/filters/supervised/instance/smote.rb +22 -0
- data/lib/wikipedia.rb +51 -0
- data/lib/wikipedia/vandalism_detection.rb +30 -0
- data/lib/wikipedia/vandalism_detection/algorithms.rb +18 -0
- data/lib/wikipedia/vandalism_detection/algorithms/kullback_leibler_divergence.rb +69 -0
- data/lib/wikipedia/vandalism_detection/classifier.rb +186 -0
- data/lib/wikipedia/vandalism_detection/configuration.rb +321 -0
- data/lib/wikipedia/vandalism_detection/diff.rb +27 -0
- data/lib/wikipedia/vandalism_detection/edit.rb +75 -0
- data/lib/wikipedia/vandalism_detection/evaluator.rb +606 -0
- data/lib/wikipedia/vandalism_detection/exceptions.rb +40 -0
- data/lib/wikipedia/vandalism_detection/feature_calculator.rb +89 -0
- data/lib/wikipedia/vandalism_detection/features.rb +67 -0
- data/lib/wikipedia/vandalism_detection/features/all_wordlists_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/all_wordlists_impact.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/anonymity.rb +19 -0
- data/lib/wikipedia/vandalism_detection/features/anonymity_previous.rb +29 -0
- data/lib/wikipedia/vandalism_detection/features/article_size.rb +18 -0
- data/lib/wikipedia/vandalism_detection/features/bad_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/bad_impact.rb +21 -0
- data/lib/wikipedia/vandalism_detection/features/base.rb +54 -0
- data/lib/wikipedia/vandalism_detection/features/biased_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/biased_impact.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/blanking.rb +25 -0
- data/lib/wikipedia/vandalism_detection/features/character_diversity.rb +25 -0
- data/lib/wikipedia/vandalism_detection/features/character_sequence.rb +19 -0
- data/lib/wikipedia/vandalism_detection/features/comment_bad_frequency.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/comment_biased_frequency.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/comment_length.rb +17 -0
- data/lib/wikipedia/vandalism_detection/features/comment_markup_frequency.rb +27 -0
- data/lib/wikipedia/vandalism_detection/features/comment_pronoun_frequency.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/comment_sex_frequency.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/comment_vulgarism_frequency.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/compressibility.rb +27 -0
- data/lib/wikipedia/vandalism_detection/features/contains_base.rb +18 -0
- data/lib/wikipedia/vandalism_detection/features/copyedit.rb +18 -0
- data/lib/wikipedia/vandalism_detection/features/digit_ratio.rb +24 -0
- data/lib/wikipedia/vandalism_detection/features/edits_per_user.rb +65 -0
- data/lib/wikipedia/vandalism_detection/features/emoticons_frequency.rb +27 -0
- data/lib/wikipedia/vandalism_detection/features/emoticons_impact.rb +29 -0
- data/lib/wikipedia/vandalism_detection/features/frequency_base.rb +20 -0
- data/lib/wikipedia/vandalism_detection/features/impact_base.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/inserted_character_distribution.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/inserted_external_links.rb +18 -0
- data/lib/wikipedia/vandalism_detection/features/inserted_internal_links.rb +18 -0
- data/lib/wikipedia/vandalism_detection/features/inserted_size.rb +20 -0
- data/lib/wikipedia/vandalism_detection/features/inserted_words.rb +18 -0
- data/lib/wikipedia/vandalism_detection/features/longest_word.rb +20 -0
- data/lib/wikipedia/vandalism_detection/features/markup_frequency.rb +27 -0
- data/lib/wikipedia/vandalism_detection/features/markup_impact.rb +27 -0
- data/lib/wikipedia/vandalism_detection/features/non_alphanumeric_ratio.rb +24 -0
- data/lib/wikipedia/vandalism_detection/features/personal_life.rb +18 -0
- data/lib/wikipedia/vandalism_detection/features/pronoun_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/pronoun_impact.rb +21 -0
- data/lib/wikipedia/vandalism_detection/features/removed_all_wordlists_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/removed_bad_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/removed_biased_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/removed_character_distribution.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/removed_emoticons_frequency.rb +27 -0
- data/lib/wikipedia/vandalism_detection/features/removed_markup_frequency.rb +28 -0
- data/lib/wikipedia/vandalism_detection/features/removed_pronoun_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/removed_sex_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/removed_size.rb +20 -0
- data/lib/wikipedia/vandalism_detection/features/removed_vulgarism_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/removed_words.rb +18 -0
- data/lib/wikipedia/vandalism_detection/features/replacement_similarity.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/reverted.rb +18 -0
- data/lib/wikipedia/vandalism_detection/features/revisions_character_distribution.rb +21 -0
- data/lib/wikipedia/vandalism_detection/features/same_editor.rb +29 -0
- data/lib/wikipedia/vandalism_detection/features/sex_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/sex_impact.rb +21 -0
- data/lib/wikipedia/vandalism_detection/features/size_increment.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/size_ratio.rb +26 -0
- data/lib/wikipedia/vandalism_detection/features/term_frequency.rb +25 -0
- data/lib/wikipedia/vandalism_detection/features/time_interval.rb +31 -0
- data/lib/wikipedia/vandalism_detection/features/time_of_day.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/upper_case_ratio.rb +24 -0
- data/lib/wikipedia/vandalism_detection/features/upper_case_words_ratio.rb +31 -0
- data/lib/wikipedia/vandalism_detection/features/upper_to_lower_case_ratio.rb +24 -0
- data/lib/wikipedia/vandalism_detection/features/user_reputation.rb +38 -0
- data/lib/wikipedia/vandalism_detection/features/vulgarism_frequency.rb +23 -0
- data/lib/wikipedia/vandalism_detection/features/vulgarism_impact.rb +22 -0
- data/lib/wikipedia/vandalism_detection/features/weekday.rb +21 -0
- data/lib/wikipedia/vandalism_detection/features/words_increment.rb +22 -0
- data/lib/wikipedia/vandalism_detection/instances.rb +130 -0
- data/lib/wikipedia/vandalism_detection/page.rb +88 -0
- data/lib/wikipedia/vandalism_detection/page_parser.rb +52 -0
- data/lib/wikipedia/vandalism_detection/revision.rb +69 -0
- data/lib/wikipedia/vandalism_detection/revision_parser.rb +43 -0
- data/lib/wikipedia/vandalism_detection/test_dataset.rb +367 -0
- data/lib/wikipedia/vandalism_detection/text.rb +18 -0
- data/lib/wikipedia/vandalism_detection/training_dataset.rb +303 -0
- data/lib/wikipedia/vandalism_detection/version.rb +5 -0
- data/lib/wikipedia/vandalism_detection/wikitext_extractor.rb +80 -0
- data/lib/wikipedia/vandalism_detection/word_lists.rb +19 -0
- data/lib/wikipedia/vandalism_detection/word_lists/bad.rb +12 -0
- data/lib/wikipedia/vandalism_detection/word_lists/biased.rb +21 -0
- data/lib/wikipedia/vandalism_detection/word_lists/emoticons.rb +22 -0
- data/lib/wikipedia/vandalism_detection/word_lists/markup.rb +12 -0
- data/lib/wikipedia/vandalism_detection/word_lists/pronouns.rb +15 -0
- data/lib/wikipedia/vandalism_detection/word_lists/sex.rb +12 -0
- data/lib/wikipedia/vandalism_detection/word_lists/vulgarism.rb +97 -0
- data/spec/factories/edit.rb +20 -0
- data/spec/factories/page.rb +13 -0
- data/spec/factories/revision.rb +51 -0
- data/spec/resources/config/config.yml +35 -0
- data/spec/resources/corpora/test/edits.csv +8 -0
- data/spec/resources/corpora/test/ground-truth.txt +3 -0
- data/spec/resources/corpora/test/revisions/part-1/326471754.txt +199 -0
- data/spec/resources/corpora/test/revisions/part-1/326873205.txt +137 -0
- data/spec/resources/corpora/test/revisions/part-1/328774035.txt +162 -0
- data/spec/resources/corpora/test/revisions/part-2/307084144.txt +137 -0
- data/spec/resources/corpora/test/revisions/part-2/326978767.txt +199 -0
- data/spec/resources/corpora/test/revisions/part-2/328774110.txt +162 -0
- data/spec/resources/corpora/training/annotations.csv +7 -0
- data/spec/resources/corpora/training/edits.csv +7 -0
- data/spec/resources/corpora/training/revisions/part-1/326471754.txt +199 -0
- data/spec/resources/corpora/training/revisions/part-1/326873205.txt +137 -0
- data/spec/resources/corpora/training/revisions/part-1/328774035.txt +162 -0
- data/spec/resources/corpora/training/revisions/part-2/307084144.txt +137 -0
- data/spec/resources/corpora/training/revisions/part-2/326978767.txt +199 -0
- data/spec/resources/corpora/training/revisions/part-2/328774110.txt +162 -0
- data/spec/resources/page_with_redirects.xml +85 -0
- data/spec/resources/redirect_page.xml +59 -0
- data/spec/resources/revision_simplified.xml +13 -0
- data/spec/resources/sample_revision.txt +137 -0
- data/spec/resources/sample_revision_clean_text.txt +1 -0
- data/spec/resources/sample_revision_plain_text.txt +183 -0
- data/spec/resources/vandalism_on_wikipedia.xml +234 -0
- data/spec/resources/vandalism_on_wikipedia_simplified.xml +119 -0
- data/spec/resources/wikipedia_tokens.txt +30 -0
- data/spec/spec_helper.rb +22 -0
- data/spec/support/macros/file_reading.rb +7 -0
- data/spec/support/macros/test_configuration.rb +71 -0
- data/spec/vandalism_detection/algorithms/kullback_leibler_divergence_spec.rb +36 -0
- data/spec/vandalism_detection/classifier_spec.rb +317 -0
- data/spec/vandalism_detection/configuration_spec.rb +517 -0
- data/spec/vandalism_detection/diff_spec.rb +40 -0
- data/spec/vandalism_detection/edit_spec.rb +137 -0
- data/spec/vandalism_detection/evaluator_spec.rb +671 -0
- data/spec/vandalism_detection/feature_calculator_spec.rb +128 -0
- data/spec/vandalism_detection/features/all_wordlists_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/all_wordlists_impact_spec.rb +58 -0
- data/spec/vandalism_detection/features/anonymity_previous_spec.rb +61 -0
- data/spec/vandalism_detection/features/anonymity_spec.rb +23 -0
- data/spec/vandalism_detection/features/article_size_spec.rb +35 -0
- data/spec/vandalism_detection/features/bad_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/bad_impact_spec.rb +59 -0
- data/spec/vandalism_detection/features/base_spec.rb +49 -0
- data/spec/vandalism_detection/features/biased_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/biased_impact_spec.rb +58 -0
- data/spec/vandalism_detection/features/blanking_spec.rb +38 -0
- data/spec/vandalism_detection/features/character_diversity_spec.rb +35 -0
- data/spec/vandalism_detection/features/character_sequence_spec.rb +37 -0
- data/spec/vandalism_detection/features/comment_bad_frequency_spec.rb +34 -0
- data/spec/vandalism_detection/features/comment_biased_frequency_spec.rb +34 -0
- data/spec/vandalism_detection/features/comment_length_spec.rb +27 -0
- data/spec/vandalism_detection/features/comment_markup_frequency_spec.rb +34 -0
- data/spec/vandalism_detection/features/comment_pronoun_frequency_spec.rb +34 -0
- data/spec/vandalism_detection/features/comment_sex_frequency_spec.rb +34 -0
- data/spec/vandalism_detection/features/comment_vulgarism_frequency_spec.rb +34 -0
- data/spec/vandalism_detection/features/compressibility_spec.rb +42 -0
- data/spec/vandalism_detection/features/contains_base_spec.rb +33 -0
- data/spec/vandalism_detection/features/copyedit_spec.rb +33 -0
- data/spec/vandalism_detection/features/digit_ratio_spec.rb +35 -0
- data/spec/vandalism_detection/features/edits_per_user_spec.rb +49 -0
- data/spec/vandalism_detection/features/emoticons_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/emoticons_impact_spec.rb +51 -0
- data/spec/vandalism_detection/features/frequency_base_spec.rb +26 -0
- data/spec/vandalism_detection/features/impact_base_spec.rb +41 -0
- data/spec/vandalism_detection/features/inserted_character_distribution_spec.rb +46 -0
- data/spec/vandalism_detection/features/inserted_external_links_spec.rb +35 -0
- data/spec/vandalism_detection/features/inserted_internal_links_spec.rb +35 -0
- data/spec/vandalism_detection/features/inserted_size_spec.rb +35 -0
- data/spec/vandalism_detection/features/inserted_words_spec.rb +35 -0
- data/spec/vandalism_detection/features/longest_word_spec.rb +35 -0
- data/spec/vandalism_detection/features/markup_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/markup_impact_spec.rb +59 -0
- data/spec/vandalism_detection/features/non_alphanumeric_ratio_spec.rb +35 -0
- data/spec/vandalism_detection/features/personal_life_spec.rb +26 -0
- data/spec/vandalism_detection/features/pronoun_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/pronoun_impact_spec.rb +59 -0
- data/spec/vandalism_detection/features/removed_all_wordlists_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/removed_bad_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/removed_biased_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/removed_character_distribution_spec.rb +46 -0
- data/spec/vandalism_detection/features/removed_emoticons_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/removed_markup_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/removed_pronoun_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/removed_sex_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/removed_size_spec.rb +35 -0
- data/spec/vandalism_detection/features/removed_vulgarism_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/removed_words_spec.rb +35 -0
- data/spec/vandalism_detection/features/replacement_similarity_spec.rb +44 -0
- data/spec/vandalism_detection/features/reverted_spec.rb +28 -0
- data/spec/vandalism_detection/features/revisions_character_distribution_spec.rb +46 -0
- data/spec/vandalism_detection/features/same_editor_spec.rb +60 -0
- data/spec/vandalism_detection/features/sex_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/sex_impact_spec.rb +59 -0
- data/spec/vandalism_detection/features/size_increment_spec.rb +35 -0
- data/spec/vandalism_detection/features/size_ratio_spec.rb +57 -0
- data/spec/vandalism_detection/features/term_frequency_spec.rb +38 -0
- data/spec/vandalism_detection/features/time_interval_spec.rb +50 -0
- data/spec/vandalism_detection/features/time_of_day_spec.rb +22 -0
- data/spec/vandalism_detection/features/upper_case_ratio_spec.rb +35 -0
- data/spec/vandalism_detection/features/upper_case_words_ratio_spec.rb +37 -0
- data/spec/vandalism_detection/features/upper_to_lower_case_ratio_spec.rb +35 -0
- data/spec/vandalism_detection/features/user_reputation_spec.rb +52 -0
- data/spec/vandalism_detection/features/vulgarism_frequency_spec.rb +36 -0
- data/spec/vandalism_detection/features/vulgarism_impact_spec.rb +58 -0
- data/spec/vandalism_detection/features/weekday_spec.rb +22 -0
- data/spec/vandalism_detection/features/words_increment_spec.rb +35 -0
- data/spec/vandalism_detection/instances_spec.rb +156 -0
- data/spec/vandalism_detection/page_parser_spec.rb +184 -0
- data/spec/vandalism_detection/page_spec.rb +135 -0
- data/spec/vandalism_detection/revision_parser_spec.rb +53 -0
- data/spec/vandalism_detection/revision_spec.rb +115 -0
- data/spec/vandalism_detection/test_dataset_spec.rb +231 -0
- data/spec/vandalism_detection/text_spec.rb +29 -0
- data/spec/vandalism_detection/training_dataset_spec.rb +264 -0
- data/spec/vandalism_detection/wikitext_extractor_spec.rb +72 -0
- data/spec/weka/classifiers/functions/lib_svm_spec.rb +38 -0
- data/spec/weka/classifiers/meta/one_class_classifier_spec.rb +76 -0
- data/spec/weka/classifiers/meta/real_ada_boost_spec.rb +40 -0
- data/spec/weka/classifiers/trees/balanced_random_forest_spec.rb +40 -0
- data/spec/weka/filters/supervised/instance/smote_spec.rb +6 -0
- data/wikipedia-vandalism_detection.gemspec +30 -0
- metadata +512 -0
@@ -0,0 +1,517 @@
|
|
1
|
+
require 'spec_helper'
|
2
|
+
require 'yaml'
|
3
|
+
|
4
|
+
describe Wikipedia::VandalismDetection do
|
5
|
+
|
6
|
+
describe "Configuration class" do
|
7
|
+
|
8
|
+
before do
|
9
|
+
Wikipedia::VandalismDetection::DefaultConfiguration.any_instance.stub(source: source_dir)
|
10
|
+
@configuration = Wikipedia::VandalismDetection::Configuration.instance
|
11
|
+
|
12
|
+
use_test_configuration
|
13
|
+
end
|
14
|
+
|
15
|
+
[ :data,
|
16
|
+
:features,
|
17
|
+
:classifier_type,
|
18
|
+
:classifier_options,
|
19
|
+
:cross_validation_fold,
|
20
|
+
:training_data_options,
|
21
|
+
:balanced_training_data?,
|
22
|
+
:unbalanced_training_data?,
|
23
|
+
:oversampled_training_data?,
|
24
|
+
:test_output_classification_file,
|
25
|
+
:oversampling_options,
|
26
|
+
:training_output_arff_file,
|
27
|
+
:test_output_arff_file,
|
28
|
+
:replace_training_data_missing_values?
|
29
|
+
].each do |attribute|
|
30
|
+
it "responds to ##{attribute}" do
|
31
|
+
expect(@configuration).to respond_to attribute
|
32
|
+
end
|
33
|
+
end
|
34
|
+
|
35
|
+
it "returns a hash for #data (the full config hash) " do
|
36
|
+
expect(@configuration.data).to be_a Hash
|
37
|
+
end
|
38
|
+
|
39
|
+
it "returns a feature array for #feature" do
|
40
|
+
expect(@configuration.features).to be_an Array
|
41
|
+
end
|
42
|
+
|
43
|
+
it "returns a numeric for #cross-validation-fold" do
|
44
|
+
expect(@configuration.cross_validation_fold).to be_a Numeric
|
45
|
+
end
|
46
|
+
|
47
|
+
describe "#test_output_classification_file" do
|
48
|
+
it "returns the classification file path extended by classifier name and training data options" do
|
49
|
+
file_path = @configuration.test_output_classification_file
|
50
|
+
classifier_name = @configuration.classifier_type.split('::').last.downcase
|
51
|
+
dataset_options = @configuration.training_data_options
|
52
|
+
file_name = Wikipedia::VandalismDetection::DefaultConfiguration::DEFAULTS['output']['test']['classification_file']
|
53
|
+
|
54
|
+
path = File.join(@configuration.output_base_directory, classifier_name, dataset_options, file_name)
|
55
|
+
expect(file_path).to eq path
|
56
|
+
end
|
57
|
+
end
|
58
|
+
|
59
|
+
describe "output arff files" do
|
60
|
+
describe "#training_output_arff_file" do
|
61
|
+
it "returns the arff file path extended by classifier name and training data options" do
|
62
|
+
file_path = @configuration.training_output_arff_file
|
63
|
+
classifier_name = @configuration.classifier_type.split('::').last.downcase
|
64
|
+
dataset_options = @configuration.training_data_options
|
65
|
+
file_name = Wikipedia::VandalismDetection::DefaultConfiguration::DEFAULTS['output']['training']['arff_file']
|
66
|
+
|
67
|
+
path = File.join(@configuration.output_base_directory, classifier_name, dataset_options, file_name)
|
68
|
+
expect(file_path).to eq path
|
69
|
+
end
|
70
|
+
end
|
71
|
+
|
72
|
+
describe "#test_output_arff_file" do
|
73
|
+
it "returns the arff file path extended by classifier name and training data options" do
|
74
|
+
file_path = @configuration.test_output_arff_file
|
75
|
+
classifier_name = @configuration.classifier_type.split('::').last.downcase
|
76
|
+
dataset_options = @configuration.training_data_options
|
77
|
+
file_name = Wikipedia::VandalismDetection::DefaultConfiguration::DEFAULTS['output']['test']['arff_file']
|
78
|
+
|
79
|
+
path = File.join(@configuration.output_base_directory, classifier_name, dataset_options, file_name)
|
80
|
+
expect(file_path).to eq path
|
81
|
+
end
|
82
|
+
end
|
83
|
+
end
|
84
|
+
|
85
|
+
describe "#use_occ?" do
|
86
|
+
it "returns true if used classifier is a one class classifier (occ)" do
|
87
|
+
config = test_config
|
88
|
+
config.instance_variable_set(:@classifier_type, Weka::Classifiers::Meta::OneClassClassifier.type)
|
89
|
+
use_configuration(config)
|
90
|
+
|
91
|
+
expect(Wikipedia::VandalismDetection.configuration.use_occ?).to be true
|
92
|
+
end
|
93
|
+
|
94
|
+
it "returns false if used classifier is not a one class classifier (occ)" do
|
95
|
+
use_test_configuration
|
96
|
+
expect(Wikipedia::VandalismDetection.configuration.use_occ?).to be false
|
97
|
+
end
|
98
|
+
end
|
99
|
+
|
100
|
+
describe "#balanced_training_data?" do
|
101
|
+
|
102
|
+
it "returns true if it is set in config" do
|
103
|
+
config = test_config
|
104
|
+
config.instance_variable_set(:@training_data_options, 'balanced')
|
105
|
+
use_configuration(config)
|
106
|
+
|
107
|
+
expect(Wikipedia::VandalismDetection.configuration.balanced_training_data?).to be true
|
108
|
+
end
|
109
|
+
|
110
|
+
it "returns false if it is not set in config" do
|
111
|
+
config = test_config
|
112
|
+
config.instance_variable_set(:@training_data_options, nil)
|
113
|
+
use_configuration(config)
|
114
|
+
|
115
|
+
expect(Wikipedia::VandalismDetection.configuration.balanced_training_data?).to be false
|
116
|
+
end
|
117
|
+
|
118
|
+
it "returns false if it is set to other value than 'balanced'" do
|
119
|
+
config = test_config
|
120
|
+
config.instance_variable_set(:@training_data_options, 'other')
|
121
|
+
use_configuration(config)
|
122
|
+
|
123
|
+
expect(Wikipedia::VandalismDetection.configuration.balanced_training_data?).to be false
|
124
|
+
end
|
125
|
+
end
|
126
|
+
|
127
|
+
describe "#unbalanced_training_data?" do
|
128
|
+
|
129
|
+
it "returns true if it is set in config" do
|
130
|
+
config = test_config
|
131
|
+
config.instance_variable_set(:@training_data_options, 'unbalanced')
|
132
|
+
use_configuration(config)
|
133
|
+
|
134
|
+
expect(Wikipedia::VandalismDetection.configuration.unbalanced_training_data?).to be true
|
135
|
+
end
|
136
|
+
|
137
|
+
it "returns true if it is not set in config" do
|
138
|
+
config = test_config
|
139
|
+
config.instance_variable_set(:@training_data_options, nil)
|
140
|
+
use_configuration(config)
|
141
|
+
|
142
|
+
expect(Wikipedia::VandalismDetection.configuration.unbalanced_training_data?).to be true
|
143
|
+
end
|
144
|
+
|
145
|
+
it "returns true if it is set to other value than 'unbalanced' or 'oversampled'" do
|
146
|
+
config = test_config
|
147
|
+
config.instance_variable_set(:@training_data_options, 'other value')
|
148
|
+
use_configuration(config)
|
149
|
+
|
150
|
+
expect(Wikipedia::VandalismDetection.configuration.unbalanced_training_data?).to be true
|
151
|
+
end
|
152
|
+
|
153
|
+
it "returns false if it is set to other value than 'unbalanced'" do
|
154
|
+
config = test_config
|
155
|
+
config.instance_variable_set(:@training_data_options, 'balanced')
|
156
|
+
use_configuration(config)
|
157
|
+
|
158
|
+
expect(Wikipedia::VandalismDetection.configuration.unbalanced_training_data?).to be false
|
159
|
+
end
|
160
|
+
end
|
161
|
+
|
162
|
+
describe "#oversampled_training_data?" do
|
163
|
+
|
164
|
+
it "returns true if it is set in config" do
|
165
|
+
config = test_config
|
166
|
+
config.instance_variable_set(:@training_data_options, 'oversampled')
|
167
|
+
use_configuration(config)
|
168
|
+
|
169
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampled_training_data?).to be true
|
170
|
+
end
|
171
|
+
|
172
|
+
it "returns false if it is not set in config" do
|
173
|
+
config = test_config
|
174
|
+
config.instance_variable_set(:@training_data_options, nil)
|
175
|
+
use_configuration(config)
|
176
|
+
|
177
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampled_training_data?).to be false
|
178
|
+
end
|
179
|
+
|
180
|
+
it "returns false if it is set to other value than 'balanced'" do
|
181
|
+
config = test_config
|
182
|
+
config.instance_variable_set(:@training_data_options, 'other')
|
183
|
+
use_configuration(config)
|
184
|
+
|
185
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampled_training_data?).to be false
|
186
|
+
end
|
187
|
+
end
|
188
|
+
|
189
|
+
describe "#oversampled_options" do
|
190
|
+
it "returns a hash" do
|
191
|
+
expect(@configuration.oversampling_options).to be_a Hash
|
192
|
+
end
|
193
|
+
|
194
|
+
it "returns a hash with the :percent and :undersampling keys" do
|
195
|
+
config = test_config
|
196
|
+
config.instance_variable_set(:@training_data_options, 'oversampled')
|
197
|
+
use_configuration(config)
|
198
|
+
|
199
|
+
hash = { percentage: 0, undersampling: true }
|
200
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampling_options.keys).to eq hash.keys
|
201
|
+
end
|
202
|
+
|
203
|
+
it "returns an empty hash if training data is not oversampled" do
|
204
|
+
config = test_config
|
205
|
+
config.instance_variable_set(:@training_data_options, 'other')
|
206
|
+
use_configuration(config)
|
207
|
+
|
208
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampling_options).to be {}
|
209
|
+
end
|
210
|
+
|
211
|
+
describe "Returning of configured options" do
|
212
|
+
before do
|
213
|
+
@percentage = 300.0
|
214
|
+
@undersampling = 200.0
|
215
|
+
@hash = { percentage: @percentage, undersampling: @undersampling }
|
216
|
+
end
|
217
|
+
|
218
|
+
it "returns the configured options with downcase params" do
|
219
|
+
config = test_config
|
220
|
+
options = "oversampled -p #{@percentage} -u true #{@undersampling}"
|
221
|
+
config.instance_variable_set(:@training_data_options, options)
|
222
|
+
use_configuration(config)
|
223
|
+
|
224
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampling_options).to eq @hash
|
225
|
+
end
|
226
|
+
|
227
|
+
it "returns the configured options with upcase params" do
|
228
|
+
config = test_config
|
229
|
+
options = "oversampled -P #{@percentage} -U true #{@undersampling}"
|
230
|
+
config.instance_variable_set(:@training_data_options, options)
|
231
|
+
use_configuration(config)
|
232
|
+
|
233
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampling_options).to eq @hash
|
234
|
+
end
|
235
|
+
|
236
|
+
it "returns the configured options with full params" do
|
237
|
+
config = test_config
|
238
|
+
options = "oversampled -Percentage #{@percentage} -Undersampling true #{@undersampling}"
|
239
|
+
config.instance_variable_set(:@training_data_options, options)
|
240
|
+
use_configuration(config)
|
241
|
+
|
242
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampling_options).to eq @hash
|
243
|
+
end
|
244
|
+
end
|
245
|
+
|
246
|
+
it "returns a default value for percent if not set" do
|
247
|
+
percentage = 100 # default value
|
248
|
+
undersampling = 200
|
249
|
+
hash = { percentage: percentage, undersampling: undersampling }
|
250
|
+
|
251
|
+
config = test_config
|
252
|
+
config.instance_variable_set(:@training_data_options, "oversampled -u true #{undersampling}")
|
253
|
+
use_configuration(config)
|
254
|
+
|
255
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampling_options).to eq hash
|
256
|
+
end
|
257
|
+
|
258
|
+
it "returns a default true for undersampling if not set" do
|
259
|
+
percentage = 200
|
260
|
+
undersampling = 100 # default value
|
261
|
+
hash = { percentage: percentage, undersampling: undersampling }
|
262
|
+
|
263
|
+
config = test_config
|
264
|
+
config.instance_variable_set(:@training_data_options, "oversampled -p #{percentage}")
|
265
|
+
use_configuration(config)
|
266
|
+
|
267
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampling_options).to eq hash
|
268
|
+
end
|
269
|
+
|
270
|
+
it "returns a percentange value for undersampling if set in -u option" do
|
271
|
+
percentage = 200
|
272
|
+
undersampling = 0.001
|
273
|
+
hash = { percentage: percentage, undersampling: undersampling }
|
274
|
+
|
275
|
+
config = test_config
|
276
|
+
config.instance_variable_set(:@training_data_options, "oversampled -p #{percentage} -u true #{undersampling}")
|
277
|
+
use_configuration(config)
|
278
|
+
|
279
|
+
expect(Wikipedia::VandalismDetection.configuration.oversampling_options).to eq hash
|
280
|
+
end
|
281
|
+
end
|
282
|
+
|
283
|
+
describe "#replace_missing_values?" do
|
284
|
+
|
285
|
+
['no', 'No', 'false', 'nope', '', nil].each do |option|
|
286
|
+
it "returns false if not set" do
|
287
|
+
config = test_config
|
288
|
+
config.instance_variable_set(:@replace_missing_values, option)
|
289
|
+
use_configuration(config)
|
290
|
+
|
291
|
+
expect(Wikipedia::VandalismDetection.configuration.replace_training_data_missing_values?).to be false
|
292
|
+
end
|
293
|
+
end
|
294
|
+
|
295
|
+
['yes', 't', 'T', 'YES', 'True', 'true'].each do |option|
|
296
|
+
it "returns true if set" do
|
297
|
+
config = test_config
|
298
|
+
config.instance_variable_set(:@replace_missing_values, option)
|
299
|
+
use_configuration(config)
|
300
|
+
|
301
|
+
expect(Wikipedia::VandalismDetection.configuration.replace_training_data_missing_values?).to be true
|
302
|
+
end
|
303
|
+
end
|
304
|
+
end
|
305
|
+
|
306
|
+
[ :training_corpus_edits_file,
|
307
|
+
:training_corpus_annotations_file,
|
308
|
+
:training_corpus_revisions_directory,
|
309
|
+
:test_corpus_edits_file,
|
310
|
+
:test_corpus_revisions_directory,
|
311
|
+
:test_corpus_ground_truth_file,
|
312
|
+
:training_output_arff_file,
|
313
|
+
:training_output_index_file,
|
314
|
+
:test_output_arff_file,
|
315
|
+
:test_output_index_file,
|
316
|
+
:classifier_type,
|
317
|
+
:classifier_options,
|
318
|
+
:output_base_directory
|
319
|
+
].each do |attribute|
|
320
|
+
it "returns a string when calling ##{attribute}" do
|
321
|
+
expect(@configuration.send(attribute)).to be_a String
|
322
|
+
end
|
323
|
+
end
|
324
|
+
end
|
325
|
+
|
326
|
+
describe "#configuration" do
|
327
|
+
|
328
|
+
it "can be overridden by a config.yml file" do
|
329
|
+
Wikipedia::VandalismDetection::DefaultConfiguration.any_instance.stub(source: source_dir)
|
330
|
+
|
331
|
+
default_config = Wikipedia::VandalismDetection::DefaultConfiguration::DEFAULTS
|
332
|
+
custom_config = YAML.load_file(File.expand_path('../../resources/config/config.yml', __FILE__))
|
333
|
+
|
334
|
+
expect(Wikipedia::VandalismDetection.configuration.data).to eq default_config.deep_merge(custom_config)
|
335
|
+
end
|
336
|
+
|
337
|
+
it "returns a Wikipedia::VandalismDetection::Configuration" do
|
338
|
+
expect(Wikipedia::VandalismDetection.configuration).to be_a Wikipedia::VandalismDetection::Configuration
|
339
|
+
end
|
340
|
+
|
341
|
+
it "has all features as default configuration" do
|
342
|
+
use_default_configuration
|
343
|
+
|
344
|
+
features = [
|
345
|
+
"anonymity",
|
346
|
+
"anonymity previous",
|
347
|
+
"all wordlists frequency",
|
348
|
+
"all wordlists impact",
|
349
|
+
"article size",
|
350
|
+
"bad frequency",
|
351
|
+
"bad impact",
|
352
|
+
"biased frequency",
|
353
|
+
"biased impact",
|
354
|
+
"blanking",
|
355
|
+
"character sequence",
|
356
|
+
"character diversity",
|
357
|
+
"comment length",
|
358
|
+
"comment biased frequency",
|
359
|
+
"comment pronoun frequency",
|
360
|
+
"comment vulgarism frequency",
|
361
|
+
"compressibility",
|
362
|
+
"copyedit",
|
363
|
+
"digit ratio",
|
364
|
+
"edits per user",
|
365
|
+
"emoticons frequency",
|
366
|
+
"emoticons impact",
|
367
|
+
"inserted size",
|
368
|
+
"inserted words",
|
369
|
+
"inserted character distribution",
|
370
|
+
"inserted external links",
|
371
|
+
"inserted internal links",
|
372
|
+
"longest word",
|
373
|
+
"markup frequency",
|
374
|
+
"markup impact",
|
375
|
+
"non-alphanumeric ratio",
|
376
|
+
"personal life",
|
377
|
+
"pronoun frequency",
|
378
|
+
"pronoun impact",
|
379
|
+
"removed size",
|
380
|
+
"removed words",
|
381
|
+
"removed all wordlists frequency",
|
382
|
+
"removed bad frequency",
|
383
|
+
"removed biased frequency",
|
384
|
+
"removed character distribution",
|
385
|
+
"removed emoticons frequency",
|
386
|
+
"removed markup frequency",
|
387
|
+
"removed pronoun frequency",
|
388
|
+
"removed sex frequency",
|
389
|
+
"removed vulgarism frequency",
|
390
|
+
"replacement similarity",
|
391
|
+
"reverted",
|
392
|
+
"revisions character distribution",
|
393
|
+
"sex frequency",
|
394
|
+
"sex impact",
|
395
|
+
"same editor",
|
396
|
+
"size increment",
|
397
|
+
"size ratio",
|
398
|
+
"term frequency",
|
399
|
+
"time interval",
|
400
|
+
"time of day",
|
401
|
+
"upper case ratio",
|
402
|
+
"upper case words ratio",
|
403
|
+
"upper to lower case ratio",
|
404
|
+
"user reputation",
|
405
|
+
"vulgarism frequency",
|
406
|
+
"vulgarism impact",
|
407
|
+
"weekday",
|
408
|
+
"words increment"
|
409
|
+
]
|
410
|
+
|
411
|
+
expect(Wikipedia::VandalismDetection.configuration['features']).to eq features
|
412
|
+
end
|
413
|
+
|
414
|
+
describe "#configuration#corpora" do
|
415
|
+
|
416
|
+
before do
|
417
|
+
use_default_configuration
|
418
|
+
end
|
419
|
+
|
420
|
+
it "has a corpora config" do
|
421
|
+
expect(Wikipedia::VandalismDetection.configuration['corpora']).to be_a Hash
|
422
|
+
end
|
423
|
+
|
424
|
+
[:training, :test].each do |attribute|
|
425
|
+
it "has a #{attribute}-corpus config" do
|
426
|
+
expect(Wikipedia::VandalismDetection.configuration['corpora'][attribute.to_s]).to be_a Hash
|
427
|
+
end
|
428
|
+
end
|
429
|
+
|
430
|
+
it "has a default nil corpora-base_directory config" do
|
431
|
+
expect(Wikipedia::VandalismDetection.configuration['corpora']['base_directory']).to be nil
|
432
|
+
end
|
433
|
+
|
434
|
+
[:base_directory, :revisions_directory, :edits_file, :annotations_file].each do |attribute|
|
435
|
+
it "has an default nil '#{attribute}' config for the training-corpus" do
|
436
|
+
expect(Wikipedia::VandalismDetection.configuration['corpora']['training'][attribute.to_s]).to be nil
|
437
|
+
end
|
438
|
+
end
|
439
|
+
|
440
|
+
[:base_directory, :revisions_directory, :edits_file].each do |attribute|
|
441
|
+
it "has an default nil '#{attribute}' config for the test-corpus" do
|
442
|
+
expect(Wikipedia::VandalismDetection.configuration['corpora']['test'][attribute.to_s]).to be nil
|
443
|
+
end
|
444
|
+
end
|
445
|
+
end
|
446
|
+
|
447
|
+
describe '#configuration#output' do
|
448
|
+
|
449
|
+
before do
|
450
|
+
use_default_configuration
|
451
|
+
end
|
452
|
+
|
453
|
+
it "has an output-config" do
|
454
|
+
expect(Wikipedia::VandalismDetection.configuration['output']).to be_a Hash
|
455
|
+
end
|
456
|
+
|
457
|
+
describe "output sub configs" do
|
458
|
+
|
459
|
+
before do
|
460
|
+
@output_config = Wikipedia::VandalismDetection.configuration['output']
|
461
|
+
expect(@output_config).to be_a Hash
|
462
|
+
end
|
463
|
+
|
464
|
+
it "has a default 'base_directory' output-config" do
|
465
|
+
expect(@output_config['base_directory']).to_not be nil
|
466
|
+
end
|
467
|
+
|
468
|
+
[:arff_file, :index_file].each do |attribute|
|
469
|
+
it "has a default '#{attribute}' config for the training-output" do
|
470
|
+
output_training_config = @output_config['training']
|
471
|
+
expect(output_training_config).to_not be nil
|
472
|
+
expect(output_training_config[attribute.to_s]).to_not be nil
|
473
|
+
end
|
474
|
+
end
|
475
|
+
|
476
|
+
[:arff_file, :index_file].each do |attribute|
|
477
|
+
it "has a default '#{attribute}' config for the test-output" do
|
478
|
+
output_test_config = @output_config['training']
|
479
|
+
expect(output_test_config).to_not be nil
|
480
|
+
expect(output_test_config[attribute.to_s]).to_not be nil
|
481
|
+
end
|
482
|
+
end
|
483
|
+
|
484
|
+
end
|
485
|
+
end
|
486
|
+
|
487
|
+
describe "#configuration#classifier" do
|
488
|
+
|
489
|
+
before do
|
490
|
+
use_default_configuration
|
491
|
+
end
|
492
|
+
|
493
|
+
it "has a classifier config" do
|
494
|
+
expect { Wikipedia::VandalismDetection.configuration['classifier'] }.not_to raise_error
|
495
|
+
end
|
496
|
+
|
497
|
+
it "return a classifier Hash" do
|
498
|
+
expect(Wikipedia::VandalismDetection.configuration['classifier']).to be_a Hash
|
499
|
+
end
|
500
|
+
|
501
|
+
[:type, :options].each do |attribute|
|
502
|
+
it "has a default nil '#{attribute}' config for classification" do
|
503
|
+
expect(Wikipedia::VandalismDetection.configuration['classifier'][attribute.to_s]).to be_nil
|
504
|
+
end
|
505
|
+
end
|
506
|
+
|
507
|
+
it "has a default 10 'cross-validation-fold' config for classifier evaluation" do
|
508
|
+
expect(Wikipedia::VandalismDetection.configuration['classifier']['cross-validation-fold']).to eq 10
|
509
|
+
end
|
510
|
+
|
511
|
+
it "has a default 'training-data-options' config of unbalanced for classifier training" do
|
512
|
+
expect(Wikipedia::VandalismDetection.configuration['classifier']['training-data-options']).to eq 'unbalanced'
|
513
|
+
end
|
514
|
+
end
|
515
|
+
end
|
516
|
+
|
517
|
+
end
|