86 lines
2.7 KiB
Python
86 lines
2.7 KiB
Python
import pickle
|
|
import sys
|
|
from math import log
|
|
import regex as re
|
|
|
|
def smoothing(count, total, num_class):
|
|
probability = (count + 1.0) / (total + num_class)
|
|
if probability > 1.0:
|
|
return 1.0
|
|
else:
|
|
return probability
|
|
|
|
def prediction(file_in, file_out):
|
|
|
|
ngrams = pickle.load(open('ngrams.pkl', 'rb'))
|
|
dictionary_size = len(ngrams[1])
|
|
|
|
in_file = open(file_in, encoding = 'utf-8')
|
|
out_file = open(file_out, 'w', encoding='utf-8')
|
|
|
|
for line in in_file:
|
|
words = re.findall(r'.*\t.*\t.* (.*?) (.*?)\t(.*?) (.*?) ', line.lower())[0]
|
|
#left_words = [str(words[0]), str(words[1])]
|
|
#right_words = [str(words[2]), str(words[3])]
|
|
|
|
left_words = [str(words[0])]
|
|
right_words = [str(words[1])]
|
|
|
|
probabilities = []
|
|
|
|
for word in ngrams[1].keys():
|
|
word = str(word[0])
|
|
pre_ngram = tuple(left_words + [word])
|
|
post_ngram = tuple([word] + right_words)
|
|
|
|
pre_prob = smoothing(ngrams[2].get(pre_ngram, 0), ngrams[2].get(tuple(left_words), 0), dictionary_size)
|
|
post_prob = smoothing(ngrams[2].get(post_ngram, 0), ngrams[2].get(post_ngram[0:1], 0), dictionary_size)
|
|
probabilities.append((word, pre_prob * post_prob))
|
|
|
|
log_prob_0 = False
|
|
probabilities = sorted(probabilities, key=lambda t: t[1], reverse=True)[:50]
|
|
probability = 1.0
|
|
text = ''
|
|
counter = 0
|
|
|
|
|
|
for probab in probabilities:
|
|
word = probab[0]
|
|
prob = probab[1]
|
|
|
|
if counter == 0 and (probability - prob <= 0.0):
|
|
text = word + ':' + str(log(0.95)) + ' :' + str(log(0.05))
|
|
log_prob_0 = True
|
|
break
|
|
|
|
if counter > 0 and (probability - prob <= 0.0):
|
|
text += ':' + str(log(probability))
|
|
log_prob_0 = True
|
|
break
|
|
|
|
text += word + ':' + str(log(prob)) + ' '
|
|
probability -= prob
|
|
counter += 1
|
|
|
|
if not log_prob_0:
|
|
text += ':' + str(log(0.0001))
|
|
|
|
out_file.write(text)
|
|
out_file.write('\n')
|
|
|
|
|
|
if __name__ == '__main__':
|
|
|
|
in_dev_0 = 'C:/Users/eryk6/PycharmProjects/retro-gap/dev-0/in.tsv'
|
|
in_dev_1 = 'C:/Users/eryk6/PycharmProjects/retro-gap/dev-1/in.tsv'
|
|
in_test_a = 'C:/Users/eryk6/PycharmProjects/retro-gap/test-A/in.tsv'
|
|
|
|
out_dev_0 = 'C:/Users/eryk6/PycharmProjects/retro-gap/dev-0/out.tsv'
|
|
out_dev_1 = 'C:/Users/eryk6/PycharmProjects/retro-gap/dev-1/out.tsv'
|
|
out_test_a = 'C:/Users/eryk6/PycharmProjects/retro-gap/test-A/out.tsv'
|
|
|
|
#prediction(in_dev_0, out_dev_0)
|
|
#prediction(in_dev_1, out_dev_1)
|
|
#prediction(in_test_a, out_test_a)
|
|
|