class TestCleanTextUtil(unittest.TestCase): """ Tests the CleanTextUtil class. """ def setUp(self): self.ctu = CleanTextUtil("french") self.words = [u"Nous", u"allions", u"à", u"la", u"plage"] def tearDown(self): rm_data_dir() def test_stem_words(self): """ Tests stem_words. 1- Verify is the result is correct. from [u"Nous", u"allions", u"à", u"la", u"plage"] to [u"Nous", u"allion", u"à", u"la", u"plag"] """ wanted = [u"Nous", u"allion", u"à", u"la", u"plag"] get = self.ctu.stem_words(self.words) self.assertEquals(get, wanted) # 1 def test_rm_stop_words(self): """ Tests rm_stop_words. 1- Verify is the result is correct. from [u"Nous", u"allions", u"à", u"la", u"plage"] to [u"allions", u"plage"] """ wanted = [u"allions", u"plage"] get = self.ctu.rm_stop_words(self.words) self.assertEquals(get, wanted) # 1 def test_clean_text(self): """ Tests clean_text. 1- Verify is the result is correct. from "Nous allions à la plage" to ["allion", "plag"] """ wanted = ["allion", "plag"] get = self.ctu.clean_text(" ".join(self.words)) self.assertEquals(get, wanted) # 1
def setUp(self): self.ctu = CleanTextUtil("french") self.words = [u"Nous", u"allions", u"à", u"la", u"plage"]