asahi417 commited on
Commit
4317bcb
·
1 Parent(s): 43f66af
Files changed (2) hide show
  1. download_audio.py +24 -12
  2. main.sh +2 -2
download_audio.py CHANGED
@@ -92,6 +92,18 @@ def to_json_serializable(val):
92
  return str(val)
93
 
94
 
 
 
 
 
 
 
 
 
 
 
 
 
95
  def get_audio(dataframe: pd.DataFrame):
96
  features = {"line_no": int(dataframe.pop('line_no').values[0])}
97
  feature_file = p_join(cache_dir_feature, f'{features["line_no"]}.json')
@@ -106,10 +118,7 @@ def get_audio(dataframe: pd.DataFrame):
106
  flag = wget(features[f"{side}.url"], output_file=features[f"{side}.path"])
107
  if not flag:
108
  print("\n#### ERROR: wget failure ####\n")
109
- if os.path.exists(feature_file):
110
- os.remove(feature_file)
111
- if os.path.exists(features[f"{side}.path"]):
112
- os.remove(features[f"{side}.path"])
113
  return None
114
  else:
115
  try:
@@ -133,16 +142,13 @@ def get_audio(dataframe: pd.DataFrame):
133
  sf.write(features[f"{side}.path"], wav["array"][start:end], wav["sampling_rate"])
134
  else:
135
  print(f"\n#### ERROR ####\n truncation error: {start}, {end}, {len(wav['array'])}")
136
- os.remove(features[f"{side}.path"])
137
- if os.path.exists(feature_file):
138
- os.remove(feature_file)
139
  return None
140
  except Exception as e:
141
  print(f"\n#### ERROR ####\n {e}")
142
- os.remove(features[f"{side}.path"])
143
- if os.path.exists(feature_file):
144
- os.remove(feature_file)
145
  return None
 
146
  with open(feature_file, "w") as f:
147
  json.dump(features, f)
148
  return features["line_no"]
@@ -177,8 +183,14 @@ if __name__ == '__main__':
177
  return json.load(f_reader)
178
 
179
 
180
- features = [loader(i) for i in glob(p_join(cache_dir_feature, '*.json'))]
181
- features = [i for i in features if line_no_start <= int(i["line_no"]) < line_no_end]
 
 
 
 
 
 
182
  print(f"push {len(features)} records to hub")
183
  data_dict = {}
184
  for side in sides:
 
92
  return str(val)
93
 
94
 
95
+ def cleanup(features, feature_file):
96
+ os.remove(features[f"{side}.path"])
97
+ if os.path.exists(feature_file):
98
+ os.remove(feature_file)
99
+ for _side in sides:
100
+ if os.path.exists(features[f"{_side}.path"]):
101
+ os.remove(features[f"{_side}.path"])
102
+ # create a dummy so that we can skip from next run
103
+ with open(feature_file, "w") as f:
104
+ json.dump({"dummy": "dummy"})
105
+
106
+
107
  def get_audio(dataframe: pd.DataFrame):
108
  features = {"line_no": int(dataframe.pop('line_no').values[0])}
109
  feature_file = p_join(cache_dir_feature, f'{features["line_no"]}.json')
 
118
  flag = wget(features[f"{side}.url"], output_file=features[f"{side}.path"])
119
  if not flag:
120
  print("\n#### ERROR: wget failure ####\n")
121
+ cleanup(features, feature_file)
 
 
 
122
  return None
123
  else:
124
  try:
 
142
  sf.write(features[f"{side}.path"], wav["array"][start:end], wav["sampling_rate"])
143
  else:
144
  print(f"\n#### ERROR ####\n truncation error: {start}, {end}, {len(wav['array'])}")
145
+ cleanup(features, feature_file)
 
 
146
  return None
147
  except Exception as e:
148
  print(f"\n#### ERROR ####\n {e}")
149
+ cleanup(features, feature_file)
 
 
150
  return None
151
+ print(f"\n### SUCCESS! ###\n:{features['line_no']}")
152
  with open(feature_file, "w") as f:
153
  json.dump(features, f)
154
  return features["line_no"]
 
183
  return json.load(f_reader)
184
 
185
 
186
+ print("UPLOADING TO HF!!!")
187
+ features = [p_join(cache_dir_feature, f'{i}.json') for i in range(line_no_start, line_no_end)]
188
+ print(f"- raw feature: {len(features)}")
189
+ features = [i for i in features if os.path.exists(i)]
190
+ print(f"- path exists: {len(features)}")
191
+ features = [loader(i) for i in features]
192
+ features = [i for i in features if "dummy" not in i]
193
+ print(f"- dummy removed: {len(features)}")
194
  print(f"push {len(features)} records to hub")
195
  data_dict = {}
196
  for side in sides:
main.sh CHANGED
@@ -1,9 +1,9 @@
 
1
  rm -rf download/audio
2
  rm -rf download/feature
3
  python -c 'n=41; import os; from glob import glob; tmp = [int(os.path.basename(i).split(".")[0]) for i in glob("download/feature/enA-jaA/*.json")]; print(len([x for x in tmp if (n-1) * 2500 <= x < n * 2500]))'
4
  python -c 'n=41; import os; from glob import glob; tmp = [int(os.path.basename(i).split(".")[0]) for i in glob("download/audio/enA-jaA/enA/*")]; print(len([x for x in tmp if (n-1) * 2500 <= x < n * 2500]))'
5
- export CUDA_VISIBLE_DEVICES=
6
- file_name='download/audio/enA-jaA/jaA/100020.mp3'; from datasets import Audio; a=Audio(); a.decode_example({"path": file_name, "bytes": None})
7
  ####################
8
  # enA-jaA: 718_606 #
9
  ####################
 
1
+ export CUDA_VISIBLE_DEVICES=
2
  rm -rf download/audio
3
  rm -rf download/feature
4
  python -c 'n=41; import os; from glob import glob; tmp = [int(os.path.basename(i).split(".")[0]) for i in glob("download/feature/enA-jaA/*.json")]; print(len([x for x in tmp if (n-1) * 2500 <= x < n * 2500]))'
5
  python -c 'n=41; import os; from glob import glob; tmp = [int(os.path.basename(i).split(".")[0]) for i in glob("download/audio/enA-jaA/enA/*")]; print(len([x for x in tmp if (n-1) * 2500 <= x < n * 2500]))'
6
+ python -c 'file_name="download/audio/enA-jaA/jaA/100020.mp3"; from datasets import Audio; a=Audio(); wav=a.decode_example({"path": file_name, "bytes": None}): print(wav)'
 
7
  ####################
8
  # enA-jaA: 718_606 #
9
  ####################