[Newcomers] Guide on adding a new dataset (WP1) about flamby HOT 4 CLOSED

Hi Jean, I have been working on Federated Tumor Segmentation (FeTS) 2021 data. I have added a Readme file in the Erum/fets2021_dataset branch detailing the download process. My question is about the last part of 5th step 'The download experience should be as smooth as possible and be done only in Python/bash if possible'.

Once we get access to data from the challenge team, we are given a BOX app link to download the data which we can directly download by opening the link. However, I am unable to create a Python/bash script for the download. With wget, I am getting 'You must have JavaScript enabled to take full advantage of Box.' Error as wget is not compatible with javascript enabled webpages. I have tried requests and urlib packages as well but no success. Do you have any suggestions on how to fix this issue or is it okay if I move to further step without the python/bash download script.

from flamby.

@ErumMushtaq I've put it to FedML-s3 service. Please refer to this script:
https://github.com/FedML-AI/FedML/blob/master/data/FeTS2021/download.sh

from flamby.

Hi @ErumMushtaq and @chaoyanghe thank you for your contributions ! I have 3 questions:

• Where you able to recover the natural splits ?
• I will talk about it the afternoon but NeurIPS track requires as smooth an experience as possible for the user so BOX app link might be okay but do we need emailing PI or waiting days for registration confirmation ?
• Regarding S3 hosting I have not found any licence but I am would imagine the terms might forbid to mirror all data. Have you checked with the organizers that it is indeed allowed ?

from flamby.

Ok I found your branch ! Looks like you did find the splits ! Thank you very much !
Can you open a PR so I can better check the situation regarding your questions ?

from flamby.