-
Notifications
You must be signed in to change notification settings - Fork 1
/
data.py
195 lines (144 loc) · 7.74 KB
/
data.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
from utils import (
update_iemocap_path,
update_file_path,
to_label,
is_common,
update_cafe_path,
update_iemocap_label,
update_fesc_label
)
import json
import pandas as pd
from datasets import DatasetDict, Audio, Dataset
import glob
import warnings
warnings.filterwarnings('ignore')
##########################################################################
def iemocap(session):
train_json_file = f"/m/triton/scratch/elec/t405-puhe/p/bijoym1/iemocap/session{session}/train.json"
test_json_file = f"/m/triton/scratch/elec/t405-puhe/p/bijoym1/iemocap/session{session}/test.json"
with open(train_json_file, "r") as f:
train_data = json.load(f)
with open(test_json_file, "r") as f:
test_data = json.load(f)
train_df = pd.DataFrame.from_dict(train_data, orient="index").reset_index()
train_df = train_df.rename(columns={'index': 'file_id', 'wav': 'audio'})
test_df = pd.DataFrame.from_dict(test_data, orient="index").reset_index()
test_df = test_df.rename(columns={'index': 'file_id', 'wav': 'audio'})
train_df["audio"] = train_df["audio"].apply(update_iemocap_path).values
test_df["audio"] = test_df["audio"].apply(update_iemocap_path).values
train_df["emo"] = train_df["emo"].apply(update_iemocap_label)
test_df["emo"] = test_df["emo"].apply(update_iemocap_label)
labels = ['anger', 'happiness', 'neutral', 'sadness'] #["ang", "hap", "neu", "sad"]
label2id, id2label = dict(), dict()
for i, label in enumerate(labels):
label2id[label] = str(i)
id2label[str(i)] = label
# convert categorical label to numerical labels
train_df["label"] = train_df["emo"].apply(to_label, args=(label2id,))
test_df["label"] = test_df["emo"].apply(to_label, args=(label2id,))
train_audio_data = Dataset.from_pandas(train_df[['audio', 'label']])
train_audio_data = train_audio_data.cast_column("audio", Audio(sampling_rate=16_000))
test_audio_data = Dataset.from_pandas(test_df[['audio', 'label']])
test_audio_data = test_audio_data.cast_column("audio", Audio(sampling_rate=16_000))
ds = DatasetDict({
'train' : train_audio_data,
'test' : test_audio_data,
'dev' : test_audio_data,
})
return label2id, id2label, ds
##########################################################################
# Finnish Spoken Emotion
def fesc(session):
SESSION_FOLDER_MAP = {
1: "TIRE", 2: "TIPE", 3: "JARA", 4: "MIKO", 5: "ANRO",
6: "RIGR", 7: "TUVA", 8: "PEKO", 9: "JAKA"
}
FOLDER_NAME = SESSION_FOLDER_MAP.get(session, "ERROR")
train_json_file = f"/m/triton/scratch/elec/t405-puhe/p/bijoym1/Finnish-emotion-spilits/{FOLDER_NAME}/train.json"
test_json_file = f"/m/triton/scratch/elec/t405-puhe/p/bijoym1/Finnish-emotion-spilits/{FOLDER_NAME}/test.json"
dev_json_file = f"/m/triton/scratch/elec/t405-puhe/p/bijoym1/Finnish-emotion-spilits/{FOLDER_NAME}/dev.json"
with open(train_json_file, "r") as f:
train_data = json.load(f)
with open(test_json_file, "r") as f:
test_data = json.load(f)
with open(dev_json_file, "r") as f:
dev_data = json.load(f)
train_df = pd.DataFrame.from_dict(train_data, orient="index").reset_index()
test_df = pd.DataFrame.from_dict(test_data, orient="index").reset_index()
dev_df = pd.DataFrame.from_dict(dev_data, orient="index").reset_index()
train_df["audio"] = train_df["file_path"].apply(update_file_path)
test_df["audio"] = test_df["file_path"].apply(update_file_path)
dev_df["audio"] = dev_df["file_path"].apply(update_file_path)
train_df = train_df.rename(columns={'index': 'file_id', 'label': 'emo'})
test_df = test_df.rename(columns={'index': 'file_id', 'label': 'emo'})
dev_df = dev_df.rename(columns={'index': 'file_id', 'label': 'emo'})
train_df = train_df.loc[train_df['emo'] != '5'].reset_index(drop=True)
test_df = test_df.loc[test_df['emo'] != '5'].reset_index(drop=True)
dev_df = dev_df.loc[dev_df['emo'] != '5'].reset_index(drop=True)
train_df["emo"] = train_df["emo"].apply(update_fesc_label)
test_df["emo"] = test_df["emo"].apply(update_fesc_label)
dev_df["emo"] = dev_df["emo"].apply(update_fesc_label)
labels = ['anger', 'happiness', 'neutral', 'sadness'] # ["1", "2", "3", "4"]
label2id, id2label = dict(), dict()
for i, label in enumerate(labels):
label2id[label] = str(i)
id2label[str(i)] = label
train_df["label"] = train_df["emo"].apply(to_label, args=(label2id,))
test_df["label"] = test_df["emo"].apply(to_label, args=(label2id,))
dev_df["label"] = dev_df["emo"].apply(to_label, args=(label2id,))
train_audio_data = Dataset.from_pandas(train_df[['audio', 'label']])
train_audio_data = train_audio_data.cast_column("audio", Audio(sampling_rate=16_000))
test_audio_data = Dataset.from_pandas(test_df[['audio', 'label']])
test_audio_data = test_audio_data.cast_column("audio", Audio(sampling_rate=16_000))
dev_audio_data = Dataset.from_pandas(dev_df[['audio', 'label']])
dev_audio_data = dev_audio_data.cast_column("audio", Audio(sampling_rate=16_000))
ds = DatasetDict({
'train' : train_audio_data,
'test' : test_audio_data,
'dev' : dev_audio_data
})
return label2id, id2label, ds
##########################################################################
def cafe(session=None):
cafe_df = pd.DataFrame()
# all_files = glob.glob("/m/triton/scratch/elec/t405-puhe/p/bijoym1/CaFE_json_splits/*")
all_files = glob.glob("/content/CaFE/CaFE_splits/*")
for file in all_files:
with open(file, "r") as f:
data = json.load(f)
df = pd.DataFrame.from_dict(data, orient="index").reset_index()
df["set"] = [file.split("/")[-1].split(".")[0] for _ in range(len(df))]
df = df.rename(columns={'wav': 'audio', 'emo': 'label'})
df["label_flag"] = df["label"].apply(is_common)
cafe_df = pd.concat([cafe_df, df], axis=0)
cafe_df = cafe_df.loc[cafe_df["label_flag"] == 1]
cafe_df = cafe_df.reset_index(drop=True)
cafe_df["dataset"] = ["CaFE" for _ in range(len(cafe_df))]
cafe_df_train = cafe_df.loc[cafe_df["set"] != "test"].reset_index(drop=True)
cafe_df_train = cafe_df_train[["audio", "label"]]
cafe_df_test = cafe_df.loc[cafe_df["set"] == "test"].reset_index(drop=True)
cafe_df_test = cafe_df_test[["audio", "label"]]
cafe_df_train["audio"] = cafe_df_train["audio"].apply(update_cafe_path).values
cafe_df_test["audio"] = cafe_df_test["audio"].apply(update_cafe_path).values
train_df = cafe_df_train
test_df = cafe_df_test
labels = ['anger', 'happiness', 'neutral', 'sadness']
label2id, id2label = dict(), dict()
for i, label in enumerate(labels):
label2id[label] = str(i)
id2label[str(i)] = label
# convert categorical label to numerical labels
train_df["label"] = train_df["label"].apply(to_label, args=(label2id,)).values
test_df["label"] = test_df["label"].apply(to_label, args=(label2id,)).values
train_audio_data = Dataset.from_pandas(train_df[['audio', 'label']])
train_audio_data = train_audio_data.cast_column("audio", Audio(sampling_rate=16_000))
test_audio_data = Dataset.from_pandas(test_df[['audio', 'label']])
test_audio_data = test_audio_data.cast_column("audio", Audio(sampling_rate=16_000))
ds = DatasetDict({
'train' : train_audio_data,
'test' : test_audio_data,
'dev' : test_audio_data
})
return label2id, id2label, ds
##########################################################################