1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
|
require 'rubygems'
require 'opentox-ruby'
require 'test/unit'
require "./validate-owl.rb"
class Float
def round_to(x)
(self * 10**x).round.to_f / 10**x
end
end
class LazarTest < Test::Unit::TestCase
def setup
@predictions = []
@compounds = []
@files = []
@dump_dir = FileUtils.mkdir_p File.join(File.dirname(__FILE__),"dump",File.basename(__FILE__,".rb"))
FileUtils.mkdir_p File.join(File.dirname(__FILE__),"reference",File.basename(__FILE__,".rb"))
end
def dump(object,file)
@files << file
FileUtils.mkdir_p File.dirname(file)
File.open(file,"w+"){|f| f.puts object.to_yaml}
end
def create_model(params)
params[:subjectid] = @@subjectid
model_uri = OpenTox::Algorithm::Lazar.new.run(params).to_s
@model = OpenTox::Model::Lazar.find model_uri, @@subjectid
dump @model, File.join(@dump_dir,caller[0][/`.*'/][1..-2],"model")+".yaml"
end
def predict_compound(compound)
@compounds << compound
prediction_uri = @model.run(:compound_uri => compound.uri, :subjectid => @@subjectid)
prediction = OpenTox::LazarPrediction.find(prediction_uri, @@subjectid)
@predictions << prediction
dump prediction, File.join(@dump_dir,caller[0][/`.*'/][1..-2],"compound_prediction")+@compounds.size.to_s+".yaml"
end
def predict_dataset(dataset)
prediction_uri = @model.run(:dataset_uri => dataset.uri, :subjectid => @@subjectid)
prediction = OpenTox::LazarPrediction.find(prediction_uri, @@subjectid)
@predictions << prediction
dump prediction, File.join(@dump_dir,caller[0][/`.*'/][1..-2],"dataset_prediction")+".yaml"
end
def cleanup # executed only when assertions succeed (teardown is called even when assertions fail)
@files.each do |f|
reference = f.sub(/dump/,"reference")
FileUtils.mkdir_p File.dirname(reference)
FileUtils.cp f, reference
FileUtils.rm f
end
#@predictions.each do |dataset|
# dataset.delete(@@subjectid)
#end
#@model.delete(@@subjectid)
end
def test_create_regression_pc_model
create_model :dataset_uri => @@regression_training_dataset.uri, :feature_dataset_uri => @@regression_feature_dataset.uri, :pc_type => "constitutional", :propositionalized => "false", :min_train_performance => -1000
predict_compound OpenTox::Compound.from_smiles("c1ccccc1NN")
assert_in_delta @predictions.first.value(@compounds.first), 3.5, 1.0
assert_equal 0.603, @predictions.first.confidence(@compounds.first).round_to(3)
assert_equal 74, @predictions.first.neighbors(@compounds.first).size
cleanup
end
def test_create_regression_pc_prop_model
create_model :dataset_uri => @@regression_training_dataset.uri, :feature_dataset_uri => @@regression_feature_dataset.uri, :pc_type => "constitutional", :propositionalized => "true", :min_train_performance => -1000
predict_compound OpenTox::Compound.from_smiles("c1ccccc1NN")
assert_in_delta @predictions.first.value(@compounds.first), 3.5, 1.0
assert_equal 0.603, @predictions.first.confidence(@compounds.first).round_to(3)
assert_equal 74, @predictions.first.neighbors(@compounds.first).size
cleanup
end
def test_create_regression_model
create_model :dataset_uri => @@regression_training_dataset.uri, :propositionalized => "false", :min_train_performance => -1000
predict_compound OpenTox::Compound.from_smiles("c1ccccc1NN")
assert_in_delta @predictions.first.value(@compounds.first), 0.7, 0.5
assert_equal 0.61, @predictions.first.confidence(@compounds.first).round_to(2)
assert_equal 253, @predictions.first.neighbors(@compounds.first).size
cleanup
end
def test_create_regression_prop_model
create_model :dataset_uri => @@regression_training_dataset.uri, :propositionalized => "true", :min_train_performance => -1000
predict_compound OpenTox::Compound.from_smiles("c1ccccc1NN")
assert_in_delta @predictions.first.value(@compounds.first), 0.6, 0.5
assert_equal 0.61, @predictions.first.confidence(@compounds.first).round_to(2)
assert_equal 253, @predictions.first.neighbors(@compounds.first).size
assert_equal 131, @model.features.size
cleanup
end
# def test_create_regression_prop_nr_hits_model
# create_model :dataset_uri => @@regression_training_dataset.uri, :propositionalized => "true", :nr_hits => "false"
# predict_compound OpenTox::Compound.from_smiles("c1ccccc1NN")
# assert_equal 0.61, @predictions.first.confidence(@compounds.first).round_to(2)
# assert_equal 253, @predictions.first.neighbors(@compounds.first).size
# assert_equal 131, @model.features.size
# cleanup
# end
def test_classification_model
create_model :dataset_uri => @@classification_training_dataset.uri
# single prediction
predict_compound OpenTox::Compound.from_smiles("c1ccccc1NN")
# dataset activity
predict_compound OpenTox::Compound.from_smiles("CNN")
# dataset prediction
predict_dataset OpenTox::Dataset.create_from_csv_file("data/multicolumn.csv", @@subjectid)
# assertions
# single prediction
assert_equal "false", @predictions[0].value(@compounds[0])
assert_equal 0.3383.round_to(4), @predictions[0].confidence(@compounds[0]).round_to(4)
assert_equal 16, @predictions[0].neighbors(@compounds[0]).size
# dataset activity
assert !@predictions[1].measured_activities(@compounds[1]).empty?
assert_equal "true", @predictions[1].measured_activities(@compounds[1]).first.to_s
assert @predictions[1].value(@compounds[1]).nil?
# dataset prediction
c = OpenTox::Compound.from_smiles("CC(=Nc1ccc2c(c1)Cc1ccccc21)O")
assert_equal nil, @predictions[2].value(c)
assert_equal "true", @predictions[2].measured_activities(c).first.to_s
c = OpenTox::Compound.from_smiles("c1ccccc1NN")
assert_equal "false", @predictions[2].value(c)
assert_equal 0.3383 , @predictions[2].confidence(c).round_to(4)
# model
assert_equal 41, @model.features.size
cleanup
end
# def test_classification_svm_model
# create_model :dataset_uri => @@classification_training_dataset.uri, :prediction_algorithm => "local_svm_classification"
# predict_compound OpenTox::Compound.from_smiles("c1ccccc1NN")
# predict_dataset OpenTox::Dataset.create_from_csv_file("data/multicolumn.csv", @@subjectid)
#
# assert_equal "true", @predictions[0].value(@compounds[0])
# assert_equal 0.5587, @predictions[0].confidence(@compounds[0]).round_to(4)
# assert_equal 16, @predictions[0].neighbors(@compounds[0]).size
#
# c = OpenTox::Compound.from_smiles("c1ccccc1NN")
# assert_equal 4, @predictions[1].compounds.size
# assert_equal "false", @predictions[1].value(c)
#
# assert_equal 41, @model.features.size
# cleanup
# end
# def test_classification_svm_prop_model
# create_model :dataset_uri => @@classification_training_dataset.uri, :prediction_algorithm => "local_svm_classification", :propositionalized => "true"
# predict_compound OpenTox::Compound.from_smiles("c1ccccc1NN")
# predict_dataset OpenTox::Dataset.create_from_csv_file("data/multicolumn.csv", @@subjectid)
#
# assert_equal "false", @predictions[0].value(@compounds[0])
# assert_equal 0.5587, @predictions[0].confidence(@compounds[0]).round_to(4)
# assert_equal 16, @predictions[0].neighbors(@compounds[0]).size
#
# c = OpenTox::Compound.from_smiles("c1ccccc1NN")
# assert_equal 4, @predictions[1].compounds.size
# assert_equal "false", @predictions[1].value(c)
#
# assert_equal 41, @model.features.size
# cleanup
# end
# DISABLED TEMPORARILY
# def test_create_regression_pc_mlr_prop_model
# create_model :dataset_uri => @@regression_training_dataset.uri, :feature_dataset_uri => @@regression_feature_dataset.uri, :pc_type => "constitutional", :prediction_algorithm => "local_mlr_prop"
# predict_compound OpenTox::Compound.from_smiles("c1ccccc1NN")
# assert_in_delta @predictions.first.value(@compounds.first), 1.02, 0.2
# assert_equal 0.728, @predictions.first.confidence(@compounds.first).round_to(3)
# #assert_equal 34, @predictions.first.neighbors(@compounds.first).size
# cleanup
# end
# DISABLED TEMPORARILY
# def test_ambit_classification_model
#
# # create model
# dataset_uri = "http://apps.ideaconsult.net:8080/ambit2/dataset/9?max=400"
# feature_uri ="http://apps.ideaconsult.net:8080/ambit2/feature/2153"
# #model_uri = OpenTox::Algorithm::Lazar.new.run({:dataset_uri => dataset_uri, :prediction_feature => feature_uri}).to_s
# #lazar = OpenTox::Model::Lazar.find model_uri
# model_uri = OpenTox::Algorithm::Lazar.new.run({:dataset_uri => dataset_uri, :prediction_feature => feature_uri, :subjectid => @@subjectid}).to_s
# validate_owl model_uri,@@subjectid
# lazar = OpenTox::Model::Lazar.find model_uri, @@subjectid
# puts lazar.features.size
# assert_equal lazar.features.size, 1874
# #puts "Model: #{lazar.uri}"
# #puts lazar.features.size
#
# # single prediction
# compound = OpenTox::Compound.from_smiles("c1ccccc1NN")
# #prediction_uri = lazar.run(:compound_uri => compound.uri)
# #prediction = OpenTox::LazarPrediction.find(prediction_uri)
# prediction_uri = lazar.run(:compound_uri => compound.uri, :subjectid => @@subjectid)
# prediction = OpenTox::LazarPrediction.find(prediction_uri, @@subjectid)
# #puts "Prediction: #{prediction.uri}"
# #puts prediction.value(compound)
# assert_equal prediction.value(compound), "3.0"
# #puts @prediction.confidence(compound).round_to(4)
# #assert_equal @prediction.confidence(compound).round_to(4), 0.3005.round_to(4)
# #assert_equal @prediction.neighbors(compound).size, 15
# #@prediction.delete(@@subjectid)
#
# # dataset activity
# #compound = OpenTox::Compound.from_smiles("CNN")
# #prediction_uri = @lazar.run(:compound_uri => compound.uri, :subjectid => @@subjectid)
# #@prediction = OpenTox::LazarPrediction.find prediction_uri, @@subjectid
# #assert !@prediction.measured_activities(compound).empty?
# #assert_equal @prediction.measured_activities(compound).first, true
# #assert @prediction.value(compound).nil?
# #@prediction.delete(@@subjectid)
#
# # dataset prediction
# #@lazar.delete(@@subjectid)
# end
end
|
