in split_data.py [0:0]
def filter_monosemous_data(data, senses):
#build sense inventory
sense_inv = {}
for label in senses:
key = get_key(label, use_pos=USE_POS)
if key in sense_inv: sense_inv[key] += 1
else: sense_inv[key] = 1
#filter out monosemous examples (examples with one sense)
polysemous = []
monosemous = []
for d in data:
key = get_key(d[1], use_pos=USE_POS)
if sense_inv[key] > 1:
polysemous.append(d)
else:
monosemous.append(d)
return polysemous, monosemous