def testCreateFromString(self, ceptMock):
term = Term()
term = term.createFromString("fox")
# Check that our mock object was called
ceptMock.assert_called_with("fox")
# Check that we have a Term type
self.assertIsInstance(term, Term)
def testFeedTermReturnsTerm(self, mockBitmap, mockTPCompute):
model = self.Model()
term = Term().createFromString("test")
result = model.feedTerm(term)
self.assertIsInstance(result, Term, "Result is not a Term")
def testFeedTermReturnsTerm(self):
model = Model()
term = Term().createFromString("test")
result = model.feedTerm(term)
self.assertIsInstance(result, Term, "Result is not a Term")
def feedTerm(self, term):
""" Feed a Term to model, returning next predicted Term """
tp = self.tp
array = numpy.array(term.toArray(), dtype="uint32")
tp.compute(array, enableLearn=True, computeInfOutput=True)
predictedCells = tp.getPredictedState()
predictedColumns = predictedCells.max(axis=1)
predictedBitmap = predictedColumns.nonzero()[0].tolist()
return Term().createFromBitmap(predictedBitmap)
def readFile(filename, model, resetSequences=False, format=None):
if model.canCheckpoint() and model.hasCheckpoint():
model.load()
exclusions = ('!', '.', ':', ',', '"', '\'', '\n')
if format == "csv":
fmt = "%s,%s,%s,%s,%s,%s,%s"
else:
# No format specified, so pretty print it
fmt = "%10s | %10s | %20s | %20s | %20s | %20s | %20s"
print(fmt %
("Sequence #", "Term #", "Current Term",
"Predicted Term 1", "Predicted Term 2", "Predicted Term 3", "Predicted Term 3"))
print("-----------------------------------"
"-----------------------------------"
"-----------------------------------"
"-----------------------------------")
s = 1
t = 1
with open(filename) as f:
for line in f:
line = "".join([c for c in line if c not in exclusions])
strings = line.split(" ")
for string in strings:
if not len(string):
continue
term = Term().createFromString(string)
prediction = model.feedTerm(term)
closestStrings = prediction.closestStrings()
closestStringsIter = iter(closestStrings)
print(fmt %
(s, t, string,
next(closestStringsIter, ""),
next(closestStringsIter, ""),
next(closestStringsIter, ""),
next(closestStringsIter, "")))
t += 1
if model.canCheckpoint():
model.save()
if resetSequences:
model.resetSequence()
s += 1
t = 1
def test_createFromString(self):
# Test enablePlaceholder
term = Term().createFromString("thisisaninvalidterm",
enablePlaceholder=False)
self.assertEqual(sum(term.toArray()), 0)
term = Term().createFromString("thisisaninvalidterm",
enablePlaceholder=True)
self.assertGreater(sum(term.toArray()), 0)
self.assertGreater(term.sparsity, 0)
placeholder = term.bitmap
# Make sure we get the same placeholder back for the same term
term = Term().createFromString("thisisaninvalidterm",
enablePlaceholder=True)
self.assertEqual(term.bitmap, placeholder)
# Make sure we get a different placeholder back for a different term
term = Term().createFromString("differentinvalidterm",
enablePlaceholder=True)
self.assertNotEqual(term.bitmap, placeholder)
def test_training(self):
term0 = Term().createFromString("the")
term1 = Term().createFromString("fox")
term2 = Term().createFromString("eats")
term3 = Term().createFromString("rodent")
model = Model()
prediction = model.feedTerm(term0)
self.assertFalse(len(prediction.bitmap))
for _ in range(5):
model.feedTerm(term1)
model.feedTerm(term2)
model.feedTerm(term3)
model.resetSequence()
model.feedTerm(term1)
prediction = model.feedTerm(term2)
self.assertEqual(prediction.closestString(), "rodent")
def POST(self, uid, string):
model = getModel(uid)
term = Term().createFromString(string)
prediction = model.feedTerm(term)
model.save()
web.header('Content-Type', 'application/json')
return json.dumps([{
"type": "term",
"term": {
"string": prediction.closestString()
}
}])
def test_checkpoint(self):
model1 = Model(checkpointDir=MODEL_CHECKPOINT_DIR)
term = Term().createFromString("fox")
for _ in range(5):
prediction = model1.feedTerm(term)
self.assertTrue(len(prediction.bitmap))
model1.save()
model2 = Model(checkpointDir=MODEL_CHECKPOINT_DIR)
model2.load()
prediction = model2.feedTerm(term)
self.assertTrue(len(prediction.bitmap))
def POST(self, uid, string):
model = getModel(uid)
term = Term().createFromString(string)
learning = False if web.input().learning == "false" else True
prediction = model.feedTerm(term, learning)
model.save()
closestStrings = prediction.closestStrings()
web.header('Content-Type', 'application/json')
return json.dumps([{
"type": "term",
"term": {
"string": string
}
} for string in closestStrings])
def feedTerm(self, term, learn=True):
""" Feed a Term to model, returning next predicted Term """
tp = self.tp
array = numpy.array(term.toArray(), dtype="uint32")
tp.resetStats()
tp.compute(array, enableLearn = learn, computeInfOutput = True)
#print "ret: " + repr(ret)
#if ret.all() == array.all():
# print "EQUAL to input"
ret = tp.getStats()
#ret = tp.printStates()
print "ret: " + repr(ret)
print
print
print "*****************************************"
predictedCells = tp.getPredictedState()
predictedColumns = predictedCells.max(axis=1)
predictedBitmap = predictedColumns.nonzero()[0].tolist()
return Term().createFromBitmap(predictedBitmap)
def readFile(filename, model, resetSequences=False):
if model.canCheckpoint() and model.hasCheckpoint():
model.load()
exclusions = ('!', '.', ':', ',', '"', '\'', '\n')
print("%10s | %10s | %20s | %20s" %
("Sequence #", "Term #", "Current Term", "Predicted Term"))
print("-----------------------------------"
"-----------------------------------")
s = 1
t = 1
with open(filename) as f:
for line in f:
line = "".join([c for c in line if c not in exclusions])
strings = line.split(" ")
for string in strings:
if not len(string):
continue
term = Term().createFromString(string)
prediction = model.feedTerm(term)
print("%10i | %10i | %20s | %20s" %
(s, t, string, prediction.closestString()))
t += 1
if model.canCheckpoint():
model.save()
if resetSequences:
model.resetSequence()
s += 1
t = 1
def test_createFromString(self):
# Test enablePlaceholder
term = Term().createFromString("thisisaninvalidterm", enablePlaceholder=False)
self.assertEqual(sum(term.toArray()), 0)
term = Term().createFromString("thisisaninvalidterm", enablePlaceholder=True)
self.assertGreater(sum(term.toArray()), 0)
self.assertGreater(term.sparsity, 0)
placeholder = term.bitmap
# Make sure we get the same placeholder back for the same term
term = Term().createFromString("thisisaninvalidterm", enablePlaceholder=True)
self.assertEqual(term.bitmap, placeholder)
# Make sure we get a different placeholder back for a different term
term = Term().createFromString("differentinvalidterm", enablePlaceholder=True)
self.assertNotEqual(term.bitmap, placeholder)