CSS-LM

Форк
0
70 строк · 2.5 Кб
1
import json
2

3
#{"sentence": "i would like to use a different operating system altogether.", "aspect": "operating system", "sentiment": "neutral"}
4

5

6
with open("../GCAE/atsa-laptop/atsa_test.json") as f:
7
    laptop_data = json.load(f)
8
laptop_num_label_dict = dict()
9
for l in laptop_data:
10
    if l["sentiment"] == "conflict":
11
        continue
12
    try:
13
        data = {"sentence": l["sentence"],"aspect":l["aspect"], "sentiment":l["sentiment"], "domain":"laptop"}
14
        laptop_num_label_dict[l["sentiment"]].append(data)
15
    except:
16
        laptop_num_label_dict[l["sentiment"]]=[]
17
        data = {"sentence": l["sentence"],"aspect":l["aspect"], "sentiment":l["sentiment"], "domain":"laptop"}
18
        laptop_num_label_dict[l["sentiment"]].append(data)
19

20

21
min_num = 99999999
22
for label in set(laptop_num_label_dict.keys()):
23
    print(label,len(laptop_num_label_dict[label]))
24
    if len(laptop_num_label_dict[label]) < min_num:
25
        min_num = len(laptop_num_label_dict[label])
26
print("======")
27
for label in laptop_num_label_dict.keys():
28
    laptop_num_label_dict[label] = laptop_num_label_dict[label][:min_num]
29
for label in laptop_num_label_dict.keys():
30
    print(label,len(laptop_num_label_dict[label]))
31
print("======")
32

33

34
with open("../GCAE/acsa-restaurant-large/acsa_test.json") as f:
35
    restaurant_data = json.load(f)
36
restaurant_num_label_dict = dict()
37
for l in restaurant_data:
38
    try:
39
        data = {"sentence": l["sentence"],"aspect":l["aspect"], "sentiment":l["sentiment"], "domain":"restaurant"}
40
        restaurant_num_label_dict[l["sentiment"]].append(data)
41
    except:
42
        restaurant_num_label_dict[l["sentiment"]]=[]
43
        data = {"sentence": l["sentence"],"aspect":l["aspect"], "sentiment":l["sentiment"], "domain":"restaurant"}
44
        restaurant_num_label_dict[l["sentiment"]].append(data)
45

46
for label in set(restaurant_num_label_dict.keys()):
47
    print(label,len(restaurant_num_label_dict[label]))
48
print("======")
49
for label in restaurant_num_label_dict.keys():
50
    restaurant_num_label_dict[label] = restaurant_num_label_dict[label][:min_num]
51
for label in restaurant_num_label_dict.keys():
52
    print(label,len(restaurant_num_label_dict[label]))
53
print("======")
54

55
total_dict = dict()
56
for label in laptop_num_label_dict.keys():
57
    total_dict[label] = restaurant_num_label_dict[label] + laptop_num_label_dict[label]
58
    print(len(total_dict[label]))
59
    print(label)
60
    print("---")
61

62
total_list = list()
63
for label in total_dict.keys():
64
    print(label)
65
    total_list += total_dict[label]
66

67
print(128*6, len(total_list))
68

69
with open("test.json","w") as f:
70
    json.dump(total_list,f)
71

72

73

74

75

Использование cookies

Мы используем файлы cookie в соответствии с Политикой конфиденциальности и Политикой использования cookies.

Нажимая кнопку «Принимаю», Вы даете АО «СберТех» согласие на обработку Ваших персональных данных в целях совершенствования нашего веб-сайта и Сервиса GitVerse, а также повышения удобства их использования.

Запретить использование cookies Вы можете самостоятельно в настройках Вашего браузера.