xtreme / pan-x cannot be downloaded

[severo]

Describe the bug

Dataset xtreme / pan-x cannot be loaded

Seems related to #2326

Steps to reproduce the bug

dataset = load_dataset("xtreme", "PAN-X.fr")

Expected results

Load the dataset

Actual results

FileNotFoundError: Couldn't find file at https://www.dropbox.com/s/12h3qqog6q4bjve/panx_dataset.tar?dl=1

Environment info

• datasets version: 1.9.0
• Platform: macOS-11.4-x86_64-i386-64bit
• Python version: 3.8.11
• PyArrow version: 4.0.1

[albertvillanova]

Hi @severo, thanks for reporting.

However I have not been able to reproduce this issue. Could you please confirm if the problem persists for you?

Maybe Dropbox (where the data source is hosted) was temporarily unavailable when you tried.

[severo]

Hmmm, the file (https://www.dropbox.com/s/dl/12h3qqog6q4bjve/panx_dataset.tar) really seems to be unavailable... I tried from various connexions and machines and got the same 404 error. Maybe the dataset has been loaded from the cache in your case?

[albertvillanova]

Yes @severo, weird... I could access the file when I answered to you, but now I cannot longer access it either... Maybe it was from the cache as you point out.

Anyway, I have opened an issue in the GitHub repository responsible for the original dataset: afshinrahimi/mmner#4
I have also contacted the maintainer by email.

I'll keep you informed with their answer.

[albertvillanova]

Reply from the author/maintainer:

Will fix the issue and let you know during the weekend.

[albertvillanova]

The author told that apparently Dropbox has changed their policy and no longer allow downloading the file without having signed in first. The author asked Hugging Face to host their dataset.