Merge pull request #13 from chinmayapancholi13/add_unittests

[WIP] Adding unittests

.travis.yml

@@ -1,3 +1,6 @@
+sudo: false
+dist: trusty
+sudo: required
 language: python
 python:
   - "2.7"
@@ -7,10 +10,14 @@ before_install:
   - ./miniconda.sh -b
   - export PATH=/home/travis/miniconda2/bin:$PATH
   - conda update --yes conda
+  - sudo apt-get update
+  - sudo apt-get install libc6
 install:
   - conda create --yes -n shorttext-test python=$TRAVIS_PYTHON_VERSION pip numpy scipy
   - source activate shorttext-test
   - pip install unittest2
   - pip install pytest
+  - pip install google-compute-engine
+  - pip install tensorflow==1.1.0
   - pip install -U .
-script: python shorttext_tests.py
+script: python shorttext_tests.py

shorttext_tests.py

@@ -1,11 +1,92 @@
+import os
 import unittest
 
-class SampleTest(unittest.TestCase):
+import shorttext
+
+
+class TestVarNNEmbeddedVecClassifier(unittest.TestCase):
     def setUp(self):
-        self.sample_var = True
+		if not os.path.isfile("test_w2v_model"):
+			os.system("wget https://raw.githubusercontent.com/chinmayapancholi13/shorttext_test_data/master/test_w2v_model")  # download w2v model 
+
+		self.w2v_model = shorttext.utils.load_word2vec_model("test_w2v_model", binary=False)  # load word2vec model
+		self.trainclass_dict = shorttext.data.subjectkeywords()  # load training data
+
+    def tearDown(self):
+		if os.path.isfile("test_w2v_model"):
+			os.remove("test_w2v_model")  # delete downloaded w2v model
+
+    def testCNNWordEmbedWithoutGensim(self):
+  		# create keras model using `CNNWordEmbed` class
+  		keras_model = shorttext.classifiers.frameworks.CNNWordEmbed(wvmodel=self.w2v_model, nb_labels=len(self.trainclass_dict.keys()), vecsize=100, with_gensim=False)
+
+    	# create and train classifier using keras model constructed above
+		main_classifier = shorttext.classifiers.VarNNEmbeddedVecClassifier(self.w2v_model, with_gensim=False, vecsize=100)
+		main_classifier.train(self.trainclass_dict, keras_model, nb_epoch=2)
+
+		# compute classification score
+		score_vals = main_classifier.score('artificial intelligence')
+		self.assertTrue(score_vals['mathematics']>0.0 and score_vals['physics']>0.0 and score_vals['theology']>0.0)
+
+    def testDoubleCNNWordEmbedWithGensim(self):
+  		# create keras model using `DoubleCNNWordEmbed` class
+  		keras_model = shorttext.classifiers.frameworks.DoubleCNNWordEmbed(wvmodel=self.w2v_model, nb_labels=len(self.trainclass_dict.keys()), vecsize=100, with_gensim=True)
+
+    	# create and train classifier using keras model constructed above
+		main_classifier = shorttext.classifiers.VarNNEmbeddedVecClassifier(self.w2v_model, with_gensim=True, vecsize=100)
+		main_classifier.train(self.trainclass_dict, keras_model, nb_epoch=2)
+
+		# compute classification score
+		score_vals = main_classifier.score('artificial intelligence')
+		self.assertTrue(score_vals['mathematics']>0.0 and score_vals['physics']>0.0 and score_vals['theology']>0.0)
+
+    def testDoubleCNNWordEmbedWithoutGensim(self):
+  		# create keras model using `DoubleCNNWordEmbed` class
+  		keras_model = shorttext.classifiers.frameworks.DoubleCNNWordEmbed(wvmodel=self.w2v_model, nb_labels=len(self.trainclass_dict.keys()), vecsize=100, with_gensim=False)
+
+    	# create and train classifier using keras model constructed above
+		main_classifier = shorttext.classifiers.VarNNEmbeddedVecClassifier(self.w2v_model, with_gensim=False, vecsize=100)
+		main_classifier.train(self.trainclass_dict, keras_model, nb_epoch=2)
+
+		# compute classification score
+		score_vals = main_classifier.score('artificial intelligence')
+		self.assertTrue(score_vals['mathematics']>0.0 and score_vals['physics']>0.0 and score_vals['theology']>0.0)
+
+    def testCNNWordEmbedWithGensim(self):
+  		# create keras model using `CNNWordEmbed` class
+  		keras_model = shorttext.classifiers.frameworks.CNNWordEmbed(wvmodel=self.w2v_model, nb_labels=len(self.trainclass_dict.keys()), vecsize=100, with_gensim=True)
+
+    	# create and train classifier using keras model constructed above
+		main_classifier = shorttext.classifiers.VarNNEmbeddedVecClassifier(self.w2v_model, with_gensim=True, vecsize=100)
+		main_classifier.train(self.trainclass_dict, keras_model, nb_epoch=2)
+
+		# compute classification score
+		score_vals = main_classifier.score('artificial intelligence')
+		self.assertTrue(score_vals['mathematics']>0.0 and score_vals['physics']>0.0 and score_vals['theology']>0.0)
+
+    def testCLSTMWordEmbedWithoutGensim(self):
+  		# create keras model using `CLSTMWordEmbed` class
+  		keras_model = shorttext.classifiers.frameworks.CLSTMWordEmbed(wvmodel=self.w2v_model, nb_labels=len(self.trainclass_dict.keys()), vecsize=100, with_gensim=False)
+
+    	# create and train classifier using keras model constructed above
+		main_classifier = shorttext.classifiers.VarNNEmbeddedVecClassifier(self.w2v_model, with_gensim=False, vecsize=100)
+		main_classifier.train(self.trainclass_dict, keras_model, nb_epoch=2)
+
+		# compute classification score
+		score_vals = main_classifier.score('artificial intelligence')
+		self.assertTrue(score_vals['mathematics']>0.0 and score_vals['physics']>0.0 and score_vals['theology']>0.0)
+
+    def testCLSTMWordEmbedWithGensim(self):
+  		# create keras model using `CLSTMWordEmbed` class
+  		keras_model = shorttext.classifiers.frameworks.CLSTMWordEmbed(wvmodel=self.w2v_model, nb_labels=len(self.trainclass_dict.keys()), vecsize=100, with_gensim=True)
+
+    	# create and train classifier using keras model constructed above
+		main_classifier = shorttext.classifiers.VarNNEmbeddedVecClassifier(self.w2v_model, with_gensim=True, vecsize=100)
+		main_classifier.train(self.trainclass_dict, keras_model, nb_epoch=2)
 
-    def testSampleTestCase(self):
-        self.assertEqual(True, self.sample_var)
+		# compute classification score
+		score_vals = main_classifier.score('artificial intelligence')
+		self.assertTrue(score_vals['mathematics']>0.0 and score_vals['physics']>0.0 and score_vals['theology']>0.0)
 
 if __name__ == '__main__':
-    unittest.main()
+    unittest.main()