def sampleFromModel(prefix='', length=10):
"""Sample symbols from the model."""
startPos = seq.size()
for c in prefix:
seq.push_back(ord(c))
endPos = seq.size()
out = []
for i in xrange(length):
predictive = model.predictiveDistribution(startPos, endPos)
sample = libplump.sample_unnormalized_pdf(predictive)
out.append(chr(sample))
seq.push_back(sample)
endPos = seq.size()
return out
def sampleFromModel(prefix='', length=10):
"""Sample symbols from the model."""
startPos = seq.size()
for c in prefix:
seq.push_back(ord(c))
endPos = seq.size()
out = []
for i in xrange(length):
predictive = model.predictiveDistribution(startPos, endPos)
sample = libplump.sample_unnormalized_pdf(predictive)
out.append(chr(sample))
seq.push_back(sample)
endPos = seq.size()
return out
def completion():
"""Simple command line word completion example."""
key = None
startPos = seq.size()
endPos = seq.size()
while key != 27:
key = ord(getch())
if key == 127:
# backspace
seq.pop_back()
endPos -= 1
sys.stdout.write('\b')
sys.stdout.write(' ')
sys.stdout.write('\b')
elif key == 13:
#return
sys.stdout.write('\n')
startPos = seq.size()
endPos = seq.size()
elif key == 9:
# tab pressed -> predict (map)
sample = 0
while chr(int(sample)) != ' ':
predictive = model.predictiveDistribution(startPos, endPos)
sample = int(np.argmax(predictive))
sys.stdout.write(chr(sample))
seq.push_back(sample)
endPos += 1
elif key == 49:
# 1 pressed -> predict (sample)
sample = 0
while chr(int(sample)) != ' ':
predictive = model.predictiveDistribution(startPos, endPos)
sample = libplump.sample_unnormalized_pdf(predictive)
sys.stdout.write(chr(sample))
seq.push_back(sample)
endPos += 1
else:
sys.stdout.write(chr(key))
seq.push_back(key)
endPos += 1
def completion():
"""Simple command line word completion example."""
key = None
startPos = seq.size()
endPos = seq.size()
while key != 27:
key = ord(getch())
if key == 127:
# backspace
seq.pop_back()
endPos -= 1
sys.stdout.write('\b')
sys.stdout.write(' ')
sys.stdout.write('\b')
elif key == 13:
#return
sys.stdout.write('\n')
startPos = seq.size()
endPos = seq.size()
elif key == 9:
# tab pressed -> predict (map)
sample = 0
while chr(int(sample)) != ' ':
predictive = model.predictiveDistribution(startPos, endPos)
sample = int(np.argmax(predictive))
sys.stdout.write(chr(sample))
seq.push_back(sample)
endPos += 1
elif key == 49:
# 1 pressed -> predict (sample)
sample = 0
while chr(int(sample)) != ' ':
predictive = model.predictiveDistribution(startPos, endPos)
sample = libplump.sample_unnormalized_pdf(predictive)
sys.stdout.write(chr(sample))
seq.push_back(sample)
endPos += 1
else:
sys.stdout.write(chr(key))
seq.push_back(key)
endPos += 1
