Address issues of missing bucket_name in s3fs paths

[RobinHolzingerQC]

Motivation

The previous way of specifying paths for s3fs (not including the bucket_name) led to issues when using endpoint_url.

Changes

As paths in the format of <root_dir>/<subdir>/.../<file> are not officially supported by s3fs (see fsspec/s3fs#824), we are now using <bucket_name>/<root_dir>/<subdir>/.../<file>.
For that, we need a bucket_name config field.
def _bucket_map(...) is adjusted according so that every channel specification also includes the bucket_name

[codecov-commenter]

Codecov Report

Attention: 1 lines in your changes are missing coverage. Please review.

Comparison is base (76f3fda) 83.59% compared to head (f5c0022) 83.60%.
Report is 3 commits behind head on main.

Files	Patch %	Lines
quetz/pkgstores.py	75.00%	1 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #673   +/-   ##
=======================================
  Coverage   83.59%   83.60%           
=======================================
  Files          79       79           
  Lines        6230     6233    +3     
=======================================
+ Hits         5208     5211    +3     
  Misses       1022     1022           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[janjagusch]

Can you brielfy elaborate what's the purpose of this env var?

[RobinHolzingerQC]

As I found out with fsspec/s3fs#824, we need to specify the bucket_name in the paths for s3fs, we need to know how the bucket is called that we write to. Previously (at least for aws s3) this information could be encoded in S3_ENDPOINT, however, f3fs doesn't work with that (bucket cannot be encoded in URL). Therefore we need another variable explicitly stating the bucket_name.

[janjagusch]

But why do we need S3_BUCKET_NAME and QUETZ_S3_BUCKET_NAME?

[RobinHolzingerQC]

I already removed QUETZ_S3_BUCKET_NAME again, I just tested the QUETZ_ prefix to make sure that it wouldn't be mapped to S3_BUCKET_NAME in the tests (as the env variable was missing)

[janjagusch]

The tests are failing because it's using the CI definition from the main branch.