continue if "val_recordings" missing; support data_dir
Browse files
nst.py
CHANGED
@@ -16,6 +16,7 @@
|
|
16 |
|
17 |
from pathlib import Path
|
18 |
import json
|
|
|
19 |
|
20 |
import datasets
|
21 |
from datasets.tasks import AutomaticSpeechRecognition
|
@@ -96,8 +97,17 @@ class NSTDataset(datasets.GeneratorBasedBuilder):
|
|
96 |
# split is hardcoded to 'train' for now; there is a test set, but
|
97 |
# it has not been modernised
|
98 |
def _split_generators(self, dl_manager):
|
99 |
-
|
100 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
101 |
return [
|
102 |
datasets.SplitGenerator(
|
103 |
name=datasets.Split.TRAIN,
|
@@ -119,7 +129,8 @@ class NSTDataset(datasets.GeneratorBasedBuilder):
|
|
119 |
data = json.load(json_file)
|
120 |
speaker_data = _get_speaker_data(data["info"])
|
121 |
pid = data["pid"]
|
122 |
-
|
|
|
123 |
for recording in data["val_recordings"]:
|
124 |
bare_path = recording['file'].replace(".wav", "")
|
125 |
text = recording["text"]
|
|
|
16 |
|
17 |
from pathlib import Path
|
18 |
import json
|
19 |
+
import os
|
20 |
|
21 |
import datasets
|
22 |
from datasets.tasks import AutomaticSpeechRecognition
|
|
|
97 |
# split is hardcoded to 'train' for now; there is a test set, but
|
98 |
# it has not been modernised
|
99 |
def _split_generators(self, dl_manager):
|
100 |
+
if hasattr(dl_manager, 'manual_dir'):
|
101 |
+
data_dir = os.path.abspath(os.path.expanduser(dl_manager.manual_dir))
|
102 |
+
JSON_FILE = _JSON_URL.split("/")[-1]
|
103 |
+
AUDIO_FILES = [
|
104 |
+
os.path.join(data_dir, a.split("/")[-1]) for a in _AUDIO_URLS
|
105 |
+
]
|
106 |
+
json_dir = dl_manager.extract(os.path.join(data_dir, JSON_FILE))
|
107 |
+
audio_dirs = dl_manager.extract(AUDIO_FILES)
|
108 |
+
else:
|
109 |
+
json_dir = dl_manager.download_and_extract(_JSON_URL)
|
110 |
+
audio_dirs = dl_manager.download_and_extract(_AUDIO_URLS)
|
111 |
return [
|
112 |
datasets.SplitGenerator(
|
113 |
name=datasets.Split.TRAIN,
|
|
|
129 |
data = json.load(json_file)
|
130 |
speaker_data = _get_speaker_data(data["info"])
|
131 |
pid = data["pid"]
|
132 |
+
if "val_recordings" not in data:
|
133 |
+
continue
|
134 |
for recording in data["val_recordings"]:
|
135 |
bare_path = recording['file'].replace(".wav", "")
|
136 |
text = recording["text"]
|