def setUp(self):
self.b = NaiveBayes(backend=RedisBackend(host=HOST, port=PORT))
self.clsfr = FMClassifier.load(os.path.join(settings.DISQUS_PATH, 'analytics', 'hadoop', 'thread_views', 'classifiers', 'fm_cmap.b'))
class NaiveBayesianTest(unittest.TestCase):
def setUp(self):
self.b = NaiveBayes(backend=RedisBackend(host=HOST, port=PORT))
self.clsfr = FMClassifier.load(os.path.join(settings.DISQUS_PATH, 'analytics', 'hadoop', 'thread_views', 'classifiers', 'fm_cmap.b'))
def tearDown(self):
pass
def _clear_redis(self):
# clear out any existing redis data
conn = Redis(host=HOST, port=PORT)
conn.flushdb()
def _train_bayesian(self):
v0 = ('fm', 'politics', 'cnn')
v1 = ('fm', 'tech news', 'avc')
v2 = ('fm', 'music', 'likefm')
self.b.train([v0] * 10)
self.b.train([v1] * 11)
self.b.train([v2] * 12)
def test_initial_training(self):
self._clear_redis()
self._train_bayesian()
self.assertEquals('politics', self.b.classify(self.clsfr, 'politics', 'cnn')[0])
self.assertEquals('politics', self.b.classify(self.clsfr, 'summer', 'cnn')[0])
self.assertEquals('tech news', self.b.classify(self.clsfr, 'tech news', 'avc')[0])
self.assertEquals('tech news', self.b.classify(self.clsfr, 'summer', 'avc')[0])
self.assertEquals('tech news', self.b.classify(self.clsfr, 'techsummerpolitics', 'avc')[0])
self.assertEquals('music', self.b.classify(self.clsfr, 'music', 'likefm')[0])
self.assertEquals('music', self.b.classify(self.clsfr, 'relationships & dating', 'likefm')[0])