This commit is contained in:
guangyusong 2025-11-19 11:42:12 -05:00 committed by GitHub
commit 8328ffb087
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -438,10 +438,12 @@ def enwik8_path():
enwik8_local_path_zip = os.path.join(base_dir, "enwik8.zip") enwik8_local_path_zip = os.path.join(base_dir, "enwik8.zip")
if not os.path.exists(enwik8_local_path): if not os.path.exists(enwik8_local_path):
print(f"Downloading enwik8 to {enwik8_local_path_zip}") print(f"Downloading enwik8 to {enwik8_local_path_zip}")
import requests import urllib.request, urllib.error
response = requests.get(enwik8_url) try:
with open(enwik8_local_path_zip, "wb") as f: with urllib.request.urlopen(enwik8_url, timeout=30) as resp, open(enwik8_local_path_zip, "wb") as f:
f.write(response.content) f.write(resp.read())
except (urllib.error.URLError, urllib.error.HTTPError) as e:
pytest.skip(f"Network unavailable or download failed: {e}")
with zipfile.ZipFile(enwik8_local_path_zip, "r") as zip_ref: with zipfile.ZipFile(enwik8_local_path_zip, "r") as zip_ref:
zip_ref.extractall(base_dir) zip_ref.extractall(base_dir)
print(f"Unzipped enwik8 to {enwik8_local_path}") print(f"Unzipped enwik8 to {enwik8_local_path}")