-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcheck_the_data.py
217 lines (183 loc) · 9.09 KB
/
check_the_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
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
from tqdm import tqdm
import pathlib
import pandas as pd
import json
IDs = ['20', '21', '23', '25', '29', '30']
def create_csv_from_json_files(): # one time execution
for id in IDs:
with open('/mnt/FS1/copd-data/speech/audio_files/copdpatient' + id + '/noNoise/segments.json') as f:
data = json.load(f)
df = pd.DataFrame.from_records(data)
df.drop(columns=['file'], inplace=True)
df['source'] = df['source'].str[35:]
df = df[['start_merged', 'end_merged', 'start_src', 'end_src', 'source']]
df.to_csv(LABELS_PATH + 'segments' + id + '.csv')
def detect_silence_labels(df, filename): # this function will be executed for each audio file
starts = df['start'].tolist() # all the starting times
ends = df['end'].tolist() # all the ending times
num = len(starts) # how many rows
missing_starts, missing_ends, duration = [], [], []
if starts[0] != 0:
missing_starts.append(0)
missing_ends.append(starts[0])
duration.append(starts[0])
for i in range(0, num - 1):
missing_starts.append(ends[i])
missing_ends.append(starts[i + 1])
duration.append(starts[i + 1] - ends[i])
if ends[-1] < 120:
missing_starts.append(ends[-1])
missing_ends.append(120)
duration.append(120 - ends[-1])
newdf = pd.DataFrame()
newdf['start'] = missing_starts
newdf['end'] = missing_ends
newdf['duration'] = duration
newdf['label'] = 'silence'
newdf['src_file'] = filename
return newdf
def append_non_speech_labels(df, filename): # this function will be executed for each audio file
df = df.sort_values(by='start')
starts = df['start'].tolist() # all the starting times
ends = df['end'].tolist() # all the ending times
num = len(starts) # how many rows
missing_starts, missing_ends, duration = [], [], []
if starts[0] > 0:
missing_starts.append(0)
missing_ends.append(starts[0])
duration.append(starts[0])
for i in range(0, num - 1):
missing_starts.append(ends[i])
missing_ends.append(starts[i + 1])
duration.append(starts[i + 1] - ends[i])
if ends[-1] < 120:
missing_starts.append(ends[-1])
missing_ends.append(120)
duration.append(120 - ends[-1])
newdf = pd.DataFrame()
newdf['start'] = missing_starts
newdf['end'] = missing_ends
newdf['duration'] = duration
newdf['label'] = 'nonspeech'
newdf['src_file'] = filename
# newdf = newdf[newdf['duration'] > 0] # two rows have wrong labels!
result = pd.concat([newdf, df])
return result.sort_values(by='start')
def find_silence_between_2_minutes(partial_df, filename):
df_silenece = detect_silence_labels(partial_df, filename)
df_silenece.reset_index(inplace=True, drop=True) # TODO this line added newly to reset index
return df_silenece
def find_start_end_merged_from_csv(grp):
"""
df ['start_merged', 'end_merged', 'src_file']
"""
df_ranges = pd.DataFrame(columns=['start', 'end', 'src_file'])
for filename, group in tqdm(grp):
group_start = group.iloc[0]['start_merged'] # this is for determining where to look for data in txt file
group_end = group.iloc[-1]['end_merged']
df2 = {'start': group_start, 'end': group_end, 'src_file': filename}
df_ranges = df_ranges.append(df2, ignore_index=True)
return df_ranges
def find_filenames_for_txt_files(df_ranges, txtlines):
df_txt_with_filename = pd.DataFrame(columns=['start_merged', 'end_merged', 'label', 'src_file'])
s_index = 0
counter = 0
for row in df_ranges.itertuples():
if counter > len(lines):
break
for line in txtlines[s_index:]:
# print(counter)
ln = line.strip().split()
if len(ln) != 3:
continue
# print(ln)
start, end, label = float(ln[0]), float(ln[1]), ln[2]
if start >= row.start and end <= row.end: # exactly between the ranges
df2 = {'start_merged': start, 'end_merged': end, 'label': label, 'src_file': row.src_file}
df_txt_with_filename = df_txt_with_filename.append(df2, ignore_index=True)
counter += 1
elif start <= row.end <= end: # half of it is in this range
df2 = {'start_merged': start, 'end_merged': row.end, 'label': label, 'src_file': row.src_file}
df3 = {'start_merged': row.end, 'end_merged': end, 'label': label,
'src_file': df_ranges['src_file'].iloc[[row.Index + 1]].values.item()}
df_txt_with_filename = df_txt_with_filename.append(df2, ignore_index=True)
df_txt_with_filename = df_txt_with_filename.append(df3, ignore_index=True)
counter += 1
elif start > row.end: # not in this range anymore
# counter += 1
s_index = counter
break
# df_txt_with_filename.to_csv('copdpatient20.csv')
return df_txt_with_filename
IDS = ['21', '23', '25', '29', '30'] # '20',
LABELS_PATH = '/home/tina/research/labels_and_maps/'
# txtfiles = list(pathlib.Path(LABELS_PATH).glob('*.txt'))
# txtfiles.sort()
# mapfiles = list(pathlib.Path(LABELS_PATH).glob('*.csv'))
# mapfiles.sort()
for id in IDS:
df_maps = pd.read_csv(LABELS_PATH + 'segments' + id + '.csv')
# print(df_maps['source'].unique())
df_maps['start_merged'] /= 16000
df_maps['end_merged'] /= 16000
df_maps['start_src'] /= 16000
df_maps['end_src'] /= 16000
grp_maps = df_maps.groupby(['source'])
df_ranges = find_start_end_merged_from_csv(grp_maps)
f = open(LABELS_PATH + 'copdpatient' + id + '.txt', 'r')
lines = f.readlines()
df_txt_with_filename = find_filenames_for_txt_files(df_ranges, lines)
# mapping them
result_df = pd.DataFrame(columns=['start', 'end', 'duration', 'label', 'src_file'])
for row in df_txt_with_filename.itertuples():
s = df_maps[
(df_maps['start_merged'] <= row.start_merged) & (row.start_merged <= df_maps['end_merged'])].index.values
e = df_maps[(df_maps['start_merged'] <= row.end_merged) & (row.end_merged <= df_maps['end_merged'])].index.values
related_df = df_maps.iloc[s[0]:e[-1] + 1, :]
related_df.reset_index(inplace=True)
# print(related_df)
if related_df.shape[0] == 1: # if it is in a single range thats easy just remap it with considering the difference
diff = (related_df['start_src'] - related_df['start_merged']).values.item()
df2 = {'start': row.start_merged + diff, 'end': row.end_merged + diff,
'duration': row.end_merged - row.start_merged, 'label': row.label,
'src_file': row.src_file}
result_df = result_df.append(df2, ignore_index=True)
else:
for i in range(0, len(related_df)):
if i == 0: # first
diff = (related_df.iloc[i]['start_src'] - related_df.iloc[i]['start_merged'])
if related_df.iloc[i]['end_merged'] - row.start_merged != 0 :
df2 = {'start': row.start_merged + diff, 'end': related_df.iloc[i]['end_merged'] + diff,
'duration': related_df.iloc[i]['end_merged'] - row.start_merged, 'label': row.label,
'src_file': row.src_file}
result_df = result_df.append(df2, ignore_index=True)
elif i == len(related_df) - 1: # last
# print("last")
diff = (related_df.iloc[i]['start_src'] - related_df.iloc[i]['start_merged'])
if row.end_merged - related_df.iloc[i]['start_merged'] != 0:
df2 = {'start': related_df.iloc[i]['start_merged'] + diff, 'end': row.end_merged + diff,
'duration': row.end_merged - related_df.iloc[i]['start_merged'], 'label': row.label,
'src_file': row.src_file}
# print(df2)
result_df = result_df.append(df2, ignore_index=True)
else: # middle
# print("middle")
if related_df.iloc[i]['end_src'] - related_df.iloc[i]['start_src']:
df2 = {'start': related_df.iloc[i]['start_src'], 'end': related_df.iloc[i]['end_src'],
'duration': related_df.iloc[i]['end_src'] - related_df.iloc[i]['start_src'], 'label': row.label,
'src_file': row.src_file}
result_df = result_df.append(df2, ignore_index=True)
#
print(result_df)
print(result_df['duration'].sum())
final_result = pd.DataFrame(columns=['start', 'end', 'duration', 'label', 'src_file'])
howmany = 0
for filename, group in tqdm(result_df.groupby(['src_file'])):
# print(group)
result = append_non_speech_labels(group, filename)
# print(result)
final_result = final_result.append(result, ignore_index=True)
howmany += 1
print(howmany)
print(final_result)
final_result.to_csv(LABELS_PATH + "mapped" + id + '.csv')