Datasets:
don't waste bandwidth
Browse files- vox_celeb.py +12 -9
vox_celeb.py
CHANGED
@@ -291,15 +291,18 @@ class VoxCeleb(datasets.GeneratorBasedBuilder):
|
|
291 |
with catch_warnings():
|
292 |
filterwarnings("ignore", category=InsecureRequestWarning)
|
293 |
head = requests.get(url.format(key=cred_key), verify=False, timeout=5, stream=True)
|
294 |
-
|
295 |
-
|
296 |
-
|
297 |
-
|
298 |
-
|
299 |
-
|
300 |
-
|
301 |
-
|
302 |
-
|
|
|
|
|
|
|
303 |
tmp_path = Path(path + "." + sha256(url.encode("utf-8")).hexdigest())
|
304 |
tmp_paths.append(tmp_path)
|
305 |
lengths.append(content_length)
|
|
|
291 |
with catch_warnings():
|
292 |
filterwarnings("ignore", category=InsecureRequestWarning)
|
293 |
head = requests.get(url.format(key=cred_key), verify=False, timeout=5, stream=True)
|
294 |
+
try:
|
295 |
+
if head.status_code == 401:
|
296 |
+
raise ValueError("failed to authenticate with VoxCeleb host")
|
297 |
+
if head.status_code < 200 or head.status_code >= 300:
|
298 |
+
raise ValueError("failed to fetch dataset")
|
299 |
+
save_credentials()
|
300 |
+
content_length = head.headers.get("Content-Length")
|
301 |
+
if content_length is None:
|
302 |
+
raise ValueError("expected non-empty Content-Length")
|
303 |
+
content_length = int(content_length)
|
304 |
+
finally:
|
305 |
+
head.close()
|
306 |
tmp_path = Path(path + "." + sha256(url.encode("utf-8")).hexdigest())
|
307 |
tmp_paths.append(tmp_path)
|
308 |
lengths.append(content_length)
|