aws credentials on the GTFS jupyterhub

[keewis]

In the s3_CMEMS_3D_kerchunk notebook, the first section says:

Credentials are stored in the gfts profile in your ~/.aws/credentials. This file is generated automatically on GFTS Jupyterhub.

You can view them with ~/.aws/credentials.

• access keys are in profile named gfts
• endpoint_url is https://s3.gra.perf.cloud.ovh.net
• region_name is gra

You should have read and write permissions to the bucket, but not delete

However, for me at least, that was not the case. Instead, the keys are set as environment variables, so I had to create the file myself.

For reference, this appears to be the description of the .aws/credentials file: https://boto3.amazonaws.com/v1/documentation/api/latest/guide/credentials.html#shared-credentials-file

[keewis]

I should also mention that the rclone command only works on jupyterhub because the aws credentials are set as environment variables, to get that to work without the environment variables you'd have to call

rclone --s3-profile gfts copy ...