PACYCx / restore_dataset.py
toecm's picture
Deploy hybrid backend with SSR disabled
881ea9e
Raw
History Blame Contribute Delete
1.25 kB
from huggingface_hub import hf_hub_download, HfApi
import pandas as pd
import os
api = HfApi()
commits = api.list_repo_commits(repo_id='toecm/IEDID', repo_type='dataset')
valid_commit = next((c for c in commits if '9bd61a137a86' in c.commit_id), None)
if valid_commit:
for filename in ['Nigerian Pidgin English.csv', 'Asụsụ Igbo.csv']:
try:
print(f"Restoring {filename}...")
path = hf_hub_download(repo_id='toecm/IEDID', filename=filename, repo_type='dataset', revision=valid_commit.commit_id)
df = pd.read_csv(path)
print(f"Loaded {filename} with {len(df)} rows.")
# Save it
local_path = f"restored_{filename.replace(' ', '_')}"
df.to_csv(local_path, index=False)
# Upload back to HF
api.upload_file(
path_or_fileobj=local_path,
path_in_repo=filename,
repo_id='toecm/IEDID',
repo_type='dataset',
commit_message=f"Restore {filename} before bug truncation"
)
print(f"Successfully restored {filename} to HF Hub")
except Exception as e:
print(f"Failed to restore {filename}: {e}")