-
Notifications
You must be signed in to change notification settings - Fork 0
/
utils_focus.py
156 lines (143 loc) · 7.93 KB
/
utils_focus.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
#(c) 2021 NCSOFT Corporation & Korea University. All rights reserved.
import json
import logging
import os
import torch
from transformers import cached_path
logger = logging.getLogger(__file__)
def get_dataset_only_train_dev(tokenizer, train_dataset_path, train_dataset_cache, dev_dataset_path, dev_dataset_cache):
def tokenize(obj):
if isinstance(obj, str):
return tokenizer.convert_tokens_to_ids(tokenizer.tokenize(obj))
if isinstance(obj, dict):
return dict((n, tokenize(o)) for n, o in obj.items())
return list(tokenize(o) for o in obj)
train_dataset_cache = train_dataset_cache + '_train_focus_' + type(tokenizer).__name__
dev_dataset_cache = dev_dataset_cache + '_dev_focus_' + type(tokenizer).__name__
if train_dataset_cache and os.path.isfile(train_dataset_cache):
logger.info("Load tokenized dataset from cache at %s", train_dataset_cache)
train_dataset = torch.load(train_dataset_cache)
dev_dataset = torch.load(dev_dataset_cache)
all_dataset = dict()
all_dataset["train"] = train_dataset["train"]
all_dataset["valid"] = dev_dataset["valid"]
else:
logger.info("Process dataset from %s", train_dataset_path)
plan_file_train = cached_path(train_dataset_path)
plan_file_dev = cached_path(dev_dataset_path)
file_dict = {"train": plan_file_train, "valid": plan_file_dev}
all_dataset = dict()
for name, file in file_dict.items():
with open(file, "r", encoding="utf-8") as f:
dataset = json.loads(f.read())
dataset_enc = dict()
dataset_enc[name] = list()
for dialogue in dataset["data"]:
ID = dialogue["dialogID"]
persona = dialogue["persona"]
knowledge = dialogue["knowledge"]
utterance = dialogue["utterance"]
new_dialogue = dict()
new_dialogue["utterance"] = list()
for i, utt in enumerate(utterance):
key = "dialogue" + str(i+1)
dial = utt[key]
dial_new = dict()
persona_can = utt["persona_candidate"]
if len(persona_can) > 5:
persona_can = persona_can[:5]
persona_ground = utt["persona_grounding"]
if len(persona_ground) > 5:
persona_ground = persona_ground[:5]
knowledge_can = utt["knowledge_candidates"]
knowledge_answer = utt["knowledge_answer_index"]
dial_enc = [tokenizer.convert_tokens_to_ids(tokenizer.tokenize(sentence.strip())) for sentence in dial]
persona_ground_enc = [1 if item==True else 0 for item in persona_ground]
knowledge_can_enc = [tokenizer.convert_tokens_to_ids(tokenizer.tokenize(sentence.strip())) for sentence in knowledge_can]
persona_can_enc = [tokenizer.convert_tokens_to_ids(tokenizer.tokenize(sentence.strip())) for sentence in persona_can]
dial_new["dialog"] = dial_enc
dial_new["persona_grounding"] = persona_ground_enc
dial_new["persona_candidates"] = persona_can_enc
dial_new["knowledge_candidates"] = knowledge_can_enc
dial_new["knowledge_answer_index"] = knowledge_answer
new_dialogue["utterance"].append(dial_new)
persona_enc = [tokenizer.convert_tokens_to_ids(tokenizer.tokenize(sentence.strip())) for sentence in persona]
knowledge_enc = [tokenizer.convert_tokens_to_ids(tokenizer.tokenize(sentence.strip())) for sentence in knowledge]
new_dialogue["persona"] = persona_enc
new_dialogue["knowledge"] = knowledge_enc
new_dialogue["dialogID"] = ID
dataset_enc[name].append(new_dialogue)
logger.info("Tokenize and encode the dataset")
dataset = dataset_enc
all_dataset[name] = dataset_enc[name]
if name == 'train':
torch.save(dataset, train_dataset_cache)
else:
torch.save(dataset, dev_dataset_cache)
return all_dataset
def get_dataset_only_test(tokenizer, dataset_path, dataset_cache):
def tokenize(obj):
if isinstance(obj, str):
return tokenizer.convert_tokens_to_ids(tokenizer.tokenize(obj))
if isinstance(obj, dict):
return dict((n, tokenize(o)) for n, o in obj.items())
return list(tokenize(o) for o in obj)
dataset_cache = dataset_cache + '_test_focus_' + type(tokenizer).__name__
if dataset_cache and os.path.isfile(dataset_cache):
logger.info("Load tokenized dataset from cache at %s", dataset_cache)
dataset = torch.load(dataset_cache)
else:
logger.info("Process dataset from %s", dataset_path)
plan_file = cached_path(dataset_path)
with open(plan_file, "r", encoding="utf-8") as f:
dataset = json.loads(f.read())
dataset_enc = dict()
dataset_enc["test"] = list()
for dialogue in dataset["data"]:
ID = dialogue["dialogID"]
persona = dialogue["persona"]
knowledge = dialogue["knowledge"]
utterance = dialogue["utterance"]
new_dialogue = dict()
new_dialogue["utterance"] = list()
for i, utt in enumerate(utterance):
key = "dialogue" + str(i+1)
dial = utt[key]
dial_new = dict()
persona_can = utt["persona_candidate"]
if len(persona_can) > 5:
persona_can = persona_can[:5]
persona_ground = utt["persona_grounding"]
if len(persona_ground) > 5:
persona_ground = persona_ground[:5]
knowledge_can = utt["knowledge_candidates"]
knowledge_answer = utt["knowledge_answer_index"]
dial_enc = [tokenizer.convert_tokens_to_ids(tokenizer.tokenize(sentence.strip())) for sentence in dial]
persona_ground_enc = [1 if item==True else 0 for item in persona_ground]
knowledge_can_enc = [tokenizer.convert_tokens_to_ids(tokenizer.tokenize(sentence.strip())) for sentence in knowledge_can]
persona_can_enc = [tokenizer.convert_tokens_to_ids(tokenizer.tokenize(sentence.strip())) for sentence in persona_can]
dial_new["dialog"] = dial_enc
dial_new["persona_candidates"] = persona_can_enc
dial_new["persona_grounding"] = persona_ground_enc
dial_new["knowledge_candidates"] = knowledge_can_enc
dial_new["knowledge_answer_index"] = knowledge_answer
new_dialogue["utterance"].append(dial_new)
persona_enc = [tokenizer.convert_tokens_to_ids(tokenizer.tokenize(sentence.strip())) for sentence in persona]
knowledge_enc = [tokenizer.convert_tokens_to_ids(tokenizer.tokenize(sentence.strip())) for sentence in knowledge]
new_dialogue["persona"] = persona_enc
new_dialogue["knowledge"] = knowledge_enc
new_dialogue["dialogID"] = ID
dataset_enc["test"].append(new_dialogue)
logger.info("Tokenize and encode the dataset")
dataset = dataset_enc
torch.save(dataset, dataset_cache)
return dataset
class AttrDict(dict):
def __init__(self, *args, **kwargs):
super(AttrDict, self).__init__(*args, **kwargs)
self.__dict__ = self
def make_focus_logdir(dir_name: str):
logdir = os.path.join(
'./models', dir_name
)
return logdir