-
Notifications
You must be signed in to change notification settings - Fork 0
/
config.py
78 lines (68 loc) · 2.71 KB
/
config.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
import json
from io_utils import *
import numpy as np
class Config:
def add_config(self, value_dict):
for k, v in value_dict.items():
setattr(self, k, v)
def print_all_parameters(self):
attrs = vars(self)
print("---------------------------------")
for k, v in attrs.items():
if type(v) is dict or type(v).__module__ == np.__name__:
continue
print("%s : %s" % (k, str(v)))
print("---------------------------------")
def load_all_data_file(self):
self.tag_dict = read_item_file(self.tag_file, add_pad=False)
self.char_dict = read_item_file(self.char_set_file, add_pad=False)
self.pos_dict = read_item_file(self.pos_set_file, add_pad=False)
self.span_label_dict = read_item_file(self.span_label_file, add_pad=False) if self.use_span_label else constraint_index_mapping
self.relative_layer_dict = read_item_file(self.relative_layer_file, add_pad=False)
self.ancestors_label_dict = read_item_file(self.ancestors_label_file, add_pad=False)
self.dp_label_dict = read_item_file(self.dp_label_file, add_pad=False)
self.word_embedding, self.word_dict = create_word_embedding(self.embedding_file, self.word_embeds_dim)
self.nchars = len(self.char_dict)
self.ntags = len(self.tag_dict)
self.nwords = len(self.word_dict)
self.npos = len(self.pos_dict)
self.nspan = len(self.span_label_dict)
self.n_rel_label = len(self.relative_layer_dict)
self.n_ancestors_label = len(self.ancestors_label_dict)
self.n_dp_label = len(self.dp_label_dict)
def read_dataset(self, srl, parse, dp, pred_parse, pred_dp, data_type):
data_set = DataReader(self, self.word_dict, self.tag_dict, self.char_dict, self.pos_dict, self.span_label_dict, self.relative_layer_dict, self.ancestors_label_dict, self.dp_label_dict, data_type)
data_set.load_data(srl, parse, dp, pred_parse, pred_dp)
return data_set
def load_json(path):
with open(path) as input_json:
_config = json.loads(input_json.read())
return _config
def make_config(config_path):
overall_config = load_json(config_path)
config = Config()
model_config = load_json(overall_config['model'])
config.add_config(model_config)
data_dict = {}
data_config = load_json(overall_config['data'])
for k, v in data_config.items():
if type(v) is str:
data_dict[k] = v
elif type(v) is list:
for data_info in v:
for sub_k, sub_v in data_info.items():
key_info = k + "_" + sub_k
if key_info not in data_dict:
data_dict[key_info] = []
data_dict[key_info].append(sub_v)
elif type(v) is dict:
for sub_k, sub_v in v.items():
data_dict[k + "_" + sub_k] = sub_v
else:
raise ValueError('Invalid ' + str(v))
for k, v in overall_config.items():
if k in ['model', 'data']:
continue
data_dict[k] = v
config.add_config(data_dict)
return config