datasets-report / app.py
albertvillanova's picture
Fix update datasets
c18321a verified
raw
history blame
799 Bytes
import datetime
import os
import gradio as gri
import huggingface_hub as hfh
from apscheduler.schedulers.background import BackgroundScheduler
def greet(name):
return "Hello " + name + "!!"
iface = gr.Interface(fn=greet, inputs="text", outputs="text")
iface.launch()
def update_datasets():
datasets = hfh.list_datasets()
yesterday = (datetime.datetime.now(datetime.timezone.utc).date() - datetime.timedelta(days=1)).isoformat()
os.makedirs("data", exist_ok=True)
with open(os.path.join("data", f"{yesterday}.json"), "w") as f:
json.dump([ds.id for ds in sorted(datasets, key=lambda item: item.id)], f)
scheduler = BackgroundScheduler()
scheduler.add_job(func=update_datasets, trigger="cron", hour=10, minute=35, timezone=datetime.timezone.utc)
scheduler.start()