-
Notifications
You must be signed in to change notification settings - Fork 0
/
Who wrote this poem.py
217 lines (187 loc) · 7.47 KB
/
Who wrote this poem.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
from typing import List
from collections import defaultdict
LANDAS = [0.9, 0.09, 0.01]
EPSILON = 0.00001
UNK = '<ناشناخته>'
start = '<شعر> '
end = ' <\شعر>'
ferdowsi_train = "ferdowsi_train.txt"
hafez_train = "hafez_train.txt"
molana_train = "molavi_train.txt"
def read_input(path):
with open(path, 'r', encoding="utf-8") as f:
sentences = f.readlines()
edit_sentences(sentences)
dictionary = create_dict(sentences)
# sentences = unkonwn_finder(sentences)
return {
"dict": dictionary,
"sentences": sentences
}
def edit_sentences(sentences: List[str]):
for i in range(len(sentences)):
sentences[i] = sentences[i].replace("?", " ")
sentences[i] = sentences[i].replace(":", " ")
sentences[i] = sentences[i].replace("؟", " ")
sentences[i] = sentences[i].replace("،", " ")
sentences[i] = sentences[i].replace("*", " ")
sentences[i] = sentences[i].replace("\"", " ")
sentences[i] = sentences[i].replace("!", " ")
sentences[i] = sentences[i].strip()
while ' ' in sentences[i]:
sentences[i] = sentences[i].replace(' ', ' ')
sentences[i] = start + sentences[i] + end
#This section has not been used
def unkonwn_finder(sentences: List[str]):
curpus = ' - '.join(sentences)
words = set(curpus.split())
for word in words:
if curpus.count(word) == 1:
curpus = curpus.replace(word, UNK)
new_sentences = curpus.split(' - ')
return new_sentences
#This section has not been used
def create_dict(sentences: List[str]):
l = []
curpus = ' - '.join(sentences)
words = curpus.split()
for word in words:
if word not in ['-', '<شعر>', '<\شعر>'] :
if curpus.count(word) > 1:
if word not in l:
l.append(word)
return l
def generate_unigram(sentences: List[str]):
grams = []
for sentence in sentences:
words = sentence.split()
for i in range(1, len(words) - 1):
gram = words[i:i + 1]
if gram not in grams:
grams.append(gram)
return grams
def generate_bigram(sentences: List[str]):
grams = []
for sentence in sentences:
words = sentence.split()
for i in range(0, len(words) - 2):
gram = words[i:i + 2]
if gram not in grams:
grams.append(gram)
return grams
def generate_bigram_for_test(sentence: str):
grams = []
words = sentence.split()
for i in range(0, len(words) - 2):
gram = words[i:i + 2]
grams.append(gram)
return grams
def learn(sentences: List[str]):
unigrams = defaultdict(lambda: 0)
bigrams = defaultdict(lambda: defaultdict(lambda: 0))
p_unigrams = defaultdict(lambda: 0)
p_bigrams = defaultdict(lambda: defaultdict(lambda: 0 ))
text = ' '.join(sentences)
size = len(text.split()) - len(sentences) * 2
for k in generate_unigram(sentences):
unigrams[k[0]] += text.count(k[0])
p_unigrams[k[0]] = unigrams[k[0]] / size
unigrams['<شعر>'] = len(sentences)
for k in generate_bigram(sentences):
tmp = ' '.join(k)
bigrams[k[0]][k[1]] = text.count(tmp)
p_bigrams[k[0]][k[1]] = bigrams[k[0]][k[1]] / unigrams[k[0]]
return{
"unigram" : p_unigrams,
"bigram": p_bigrams
}
def backoff_model(unigram, bigram, words):
return bigram[words[0]][words[1]] * LANDAS[0] + unigram[words[1]] * LANDAS[1] + EPSILON * LANDAS[2]
def read_test_set():
with open("test_file.txt", "r", encoding="utf-8") as f:
ferdowsi_test_sentences = []
hafez_test_sentences = []
molana_test_sentences = []
lines = f.readlines()
for line in lines:
l = line.split("\t")
if int(l[0]) == 1:
ferdowsi_test_sentences.append(l[1])
elif int(l[0]) == 2:
hafez_test_sentences.append(l[1])
elif int(l[0]) == 3:
molana_test_sentences.append(l[1])
edit_sentences(ferdowsi_test_sentences)
edit_sentences(hafez_test_sentences)
edit_sentences(molana_test_sentences)
return {
"ferdowsi_test_sentences": ferdowsi_test_sentences,
"hafez_test_sentences": hafez_test_sentences,
"molana_test_sentences": molana_test_sentences
}
def replace_with_UNK(fd: List[str], sentences:List[str]):
text = ' - '.join(sentences)
for element in text:
if element != '-' and element != '<شعر>' and element != '<\شعر>':
if element not in fd:
text = text.replace(element, UNK)
return text.split(' - ')
ferdowsi = read_input(ferdowsi_train)
ferdowsi_model = learn(ferdowsi["sentences"])
ferdowsi_unigram = ferdowsi_model["unigram"]
ferdowsi_bigram = ferdowsi_model["bigram"]
print("Complete Ferdowsi training ...")
hafez = read_input(hafez_train)
hafez_model = learn(hafez["sentences"])
hafez_unigram = hafez_model["unigram"]
hafez_bigram = hafez_model["bigram"]
print("Complete Hafez training ...")
molana = read_input(molana_train)
molana_model = learn(molana["sentences"])
molana_unigram = molana_model["unigram"]
molana_bigram = molana_model["bigram"]
print("Complete Molana training ...")
l = read_test_set()
ferdowsi_test_sentences = l["ferdowsi_test_sentences"]
hafez_test_sentences = l["hafez_test_sentences"]
molana_test_sentences = l["molana_test_sentences"]
# ferdowsi_test_sentences = replace_with_UNK(ferdowsi_test_sentences, ferdowsi["dict"])
# hafez_test_sentences = replace_with_UNK(hafez_test_sentences, hafez["dict"])
# molana_test_sentences = replace_with_UNK(molana_test_sentences, molana["dict"])
ferdowsi_correct = 0
for sentence in ferdowsi_test_sentences:
res_f = 1
res_h = 1
res_m = 1
for k in generate_bigram_for_test(sentence):
res_f *= backoff_model(ferdowsi_unigram, ferdowsi_bigram, k)
res_h *= backoff_model(hafez_unigram, hafez_bigram, k)
res_m *= backoff_model(molana_unigram, molana_bigram, k)
if max(res_m, res_h, res_f) == res_f:
ferdowsi_correct += 1
hafez_correct = 0
for sentence in hafez_test_sentences:
res_f = 1
res_h = 1
res_m = 1
for k in generate_bigram_for_test(sentence):
res_f *= backoff_model(ferdowsi_unigram, ferdowsi_bigram, k)
res_h *= backoff_model(hafez_unigram, hafez_bigram, k)
res_m *= backoff_model(molana_unigram, molana_bigram, k)
if max(res_m, res_h, res_f) == res_h:
hafez_correct += 1
molana_correct = 0
for sentence in molana_test_sentences:
res_f = 1
res_h = 1
res_m = 1
for k in generate_bigram_for_test(sentence):
res_f *= backoff_model(ferdowsi_unigram, ferdowsi_bigram, k)
res_h *= backoff_model(hafez_unigram, hafez_bigram, k)
res_m *= backoff_model(molana_unigram, molana_bigram, k)
if max(res_m, res_h, res_f) == res_m:
molana_correct += 1
print("Ferdowsi : ", ferdowsi_correct / len(ferdowsi_test_sentences)*100 , "%")
print("Hafez : ", hafez_correct / len(hafez_test_sentences)*100, '%')
print("Molana : ", molana_correct / len(molana_test_sentences)*100, '%')
print("Overall :", (ferdowsi_correct+hafez_correct+molana_correct) / (len(ferdowsi_test_sentences)+len(hafez_test_sentences)+len(molana_test_sentences)) *100,'%')