def test_nhamPersisted(self):
"""
L{SQLite3Classifier} tracks, in a database, the number of ham messages
it has been trained with.
"""
self.classifier.train(StringIO("very nice words"), False)
bayes = spam._SQLite3Classifier(self.path)
self.assertEqual(bayes.nham, 1)
def test_nhamPersisted(self):
"""
L{SQLite3Classifier} tracks, in a database, the number of ham messages
it has been trained with.
"""
self.classifier.train(StringIO("very nice words"), False)
bayes = spam._SQLite3Classifier(self.path)
self.assertEqual(bayes.nham, 1)
def test_nspamPersisted(self):
"""
L{SQLite3Classifier} tracks, in a database, the number of spam messages
it has been trained with.
"""
self.classifier.train(StringIO("spam words of spamfulness"), True)
bayes = spam._SQLite3Classifier(self.path)
self.assertEqual(bayes.nspam, 1)
def test_nspamPersisted(self):
"""
L{SQLite3Classifier} tracks, in a database, the number of spam messages
it has been trained with.
"""
self.classifier.train(StringIO("spam words of spamfulness"), True)
bayes = spam._SQLite3Classifier(self.path)
self.assertEqual(bayes.nspam, 1)
def test_hamClassificationWithoutCache(self):
"""
Like L{test_spamClassification}, but ensure no instance cache is used to
satisfied word info lookups.
"""
self.classifier.train(StringIO("very nice words"), False)
classifier = Hammie(spam._SQLite3Classifier(self.path), mode='r')
self.assertTrue(classifier.score(StringIO("words, very nice")) < 0.01)
def test_spamClassificationWithoutCache(self):
"""
Like L{test_spamClassification}, but ensure no instance cache is used to
satisfied word info lookups.
"""
self.classifier.train(StringIO("spam words of spamfulness"), True)
classifier = Hammie(spam._SQLite3Classifier(self.path), mode='r')
self.assertTrue(
classifier.score(StringIO("spamfulness words of spam")) > 0.99)
def test_spamClassificationWithoutCache(self):
"""
Like L{test_spamClassification}, but ensure no instance cache is used to
satisfied word info lookups.
"""
self.classifier.train(StringIO("spam words of spamfulness"), True)
classifier = Hammie(spam._SQLite3Classifier(self.path), mode='r')
self.assertTrue(
classifier.score(StringIO("spamfulness words of spam")) > 0.99)
def test_hamClassificationWithoutCache(self):
"""
Like L{test_spamClassification}, but ensure no instance cache is used to
satisfied word info lookups.
"""
self.classifier.train(StringIO("very nice words"), False)
classifier = Hammie(spam._SQLite3Classifier(self.path), mode='r')
self.assertTrue(
classifier.score(StringIO("words, very nice")) < 0.01)
def test_hamTokenIncremented(self):
"""
Encountered on a subsequent ham training operation, an existing word
info row has its spam count left alone and its ham count incremented.
"""
self.classifier.train(StringIO("spam bad puppies"), True)
self.classifier.train(StringIO("justice sunshine puppies"), False)
bayes = spam._SQLite3Classifier(self.path)
wordInfo = bayes._get(u"puppies")
self.assertEqual((u"puppies", 1, 1), wordInfo)
def test_hamTokenRecorded(self):
"""
The first time a token is encountered during ham training, a row is
inserted into the database counting it as never a spam token, once a ham
token.
"""
self.classifier.train(StringIO("justice sunshine puppies"), False)
bayes = spam._SQLite3Classifier(self.path)
wordInfo = bayes._get(u"sunshine")
self.assertEqual((u"sunshine", 0, 1), wordInfo)
def test_spamTokenRecorded(self):
"""
The first time a token is encountered during spam training, a row is
inserted into the database counting it as once a spam token, never a ham
token.
"""
self.classifier.train(StringIO("spam bad gross"), True)
bayes = spam._SQLite3Classifier(self.path)
wordInfo = bayes._get(u"spam")
self.assertEqual((u"spam", 1, 0), wordInfo)
def test_hamTokenIncremented(self):
"""
Encountered on a subsequent ham training operation, an existing word
info row has its spam count left alone and its ham count incremented.
"""
self.classifier.train(StringIO("spam bad puppies"), True)
self.classifier.train(StringIO("justice sunshine puppies"), False)
bayes = spam._SQLite3Classifier(self.path)
wordInfo = bayes._get(u"puppies")
self.assertEqual((u"puppies", 1, 1), wordInfo)
def test_hamTokenRecorded(self):
"""
The first time a token is encountered during ham training, a row is
inserted into the database counting it as never a spam token, once a ham
token.
"""
self.classifier.train(StringIO("justice sunshine puppies"), False)
bayes = spam._SQLite3Classifier(self.path)
wordInfo = bayes._get(u"sunshine")
self.assertEqual((u"sunshine", 0, 1), wordInfo)
def test_spamTokenRecorded(self):
"""
The first time a token is encountered during spam training, a row is
inserted into the database counting it as once a spam token, never a ham
token.
"""
self.classifier.train(StringIO("spam bad gross"), True)
bayes = spam._SQLite3Classifier(self.path)
wordInfo = bayes._get(u"spam")
self.assertEqual((u"spam", 1, 0), wordInfo)
def test_largeDocumentClassification(self):
"""
A document with more than 999 tokens can be successfully classified.
"""
words = []
for i in range(1000):
word = "word%d" % (i, )
words.append(word)
document = " ".join(words)
self.classifier.train(StringIO(document), False)
classifier = Hammie(spam._SQLite3Classifier(self.path), mode='r')
self.assertTrue(classifier.score(StringIO(document)) < 0.01)
def test_largeDocumentClassification(self):
"""
A document with more than 999 tokens can be successfully classified.
"""
words = []
for i in range(1000):
word = "word%d" % (i,)
words.append(word)
document = " ".join(words)
self.classifier.train(StringIO(document), False)
classifier = Hammie(spam._SQLite3Classifier(self.path), mode='r')
self.assertTrue(
classifier.score(StringIO(document)) < 0.01)
def setUp(self):
self.path = self.mktemp()
self.bayes = spam._SQLite3Classifier(self.path)
self.classifier = Hammie(self.bayes, mode='r')
def setUp(self):
self.path = self.mktemp()
self.bayes = spam._SQLite3Classifier(self.path)
self.classifier = Hammie(self.bayes, mode='r')
