elif word.istitle():
word = "_CF_"
elif all(c in string.punctuation or c.isdigit() for c in word):
word = "_NP_"
else:
word = "_RARE_"
# Iterate over u and v
for u in K[k - 1]:
for v in K[k]:
# Find max over w in K[k-2]
w_candidates = defaultdict(float)
for w in K[k - 2]:
w_candidates[w] = pi[k - 1][(w, u)] * counter.calc_mle(
[w, u, v]) * counter.calc_emissions(word, v)
final_w = max(w_candidates.iteritems(),
key=operator.itemgetter(1))
# Assign pi value
pi[k][(u, v)] = final_w[1]
# Get the (tag, probability) of v in max(pi[k](u,v))
final_k_idx = max(pi[k].iteritems(), key=operator.itemgetter(1))
prob = final_k_idx[1]
# Log probability
log_prob = math.log(prob)
# Ouput format: word, tag, log probability
sys.stdout.write("%s %s %s\n" %
sys.exit(2)
try:
counts_file = file(sys.argv[1], "r")
trigram_file = sys.argv[2]
except IOError:
sys.stderr.write("ERROR: Cannot read inputfile %s.\n" % arg)
sys.exit(1)
# Initialize a trigram counter
counter = Hmm(3)
# Read in counts
counter.read_counts(count_file)
# Iterate through trigrams in trigram_file and calculate the log probability of each trigram.
for line in test_file:
trigram = line.strip().split(" ")
if trigram: # Nonempty line
prob = counter.calc_mle(trigram)
# Get the log of the probability
log_prob = math.log(prob)
# Write log probability to output file
sys.stdout.write(
"%s %s %s %s\n" %
(trigram[0], trigram[1], trigram[2], str(log_prob)))
else:
print ""
