def getFreqRepr(X, D):
freqX = OrderedDict()
reprX = OrderedDict()
for i, w in enumerate(X.words):
freqX[w] = X.freq[i]
reprX[w] = {j: X.features[i, j] for j in xrange(D)}
return freqX, reprX
# read input parameters
if __name__ == '__main__':
N = int(sys.argv[1])
D = int(sys.argv[2])
Nseed = int(sys.argv[3])
# make X,Y mock data
(X, Y, pi) = make(N, Nseed, D)
seed = [(i, i) for i in xrange(Nseed)]
# write to CSV files
freqX, reprX = getFreqRepr(X, D)
freqY, reprY = getFreqRepr(Y, D)
IO.writePickledWords('pockX.txt', freqX, reprX)
IO.writePickledWords('pockY.txt', freqY, reprY)
IO.writeWords('mockX.txt', X)
IO.writeWords('mockY.txt', Y)
IO.writeSeed('seedXY.txt', seed)
print X.asTuple()
# now need to save
