Closes # (if applicable).

Change proposed in this Pull Request

Update, correct and synchronise dependencies.

Description

• minimum pandas version seems to be pandas=0.25 which is the minimum of xarray>=0.16.2 (which is now also set consistently).
• dask memory regression not yet resolved Regression: poor memory management in threads and sync schedulers dask/dask#7583 but newer versions released, resulting in dask<2021.04.0
• tqdm is used since recently, but was not added to list of dependencies
• environment.yaml missed progressbar and requests
• setup.py requested Python 3.6 but newer versions also work.
• Why was the conda-forge channel wedged between two defaults?

Motivation and Context

Some dependencies were not correctly specified and increasingly out of sync across setup.py, environment.yaml and the conda-feedstock (see conda-forge/atlite-feedstock#14)

How Has This Been Tested?

For environment.yaml:

conda env create -f environment.yaml
pip install --no-deps -e .
pytest

(another time where I fixed pandas=0.25 and python=3.6)

For setup.py:

conda env create -n atlite-pip python=3.8 pip pytest
pip install -e .
pytest

For all, I had problems with

• test_dx_dy_preservation_era5_weird
• test_all_non_na_era5_weird

and I could not figure out why. They just get stuck (not windows!).

I also had problems with update_feature_test:

>           ds = (cutout.data.merge(ds[missing_vars.values])
                  .assign_attrs(**non_bool_dict(cutout.data.attrs), **ds.attrs))
E           TypeError: xarray.core.common.DataWithCoords.assign_attrs() got multiple values for keyword argument 'Conventions'

atlite/data.py:159: TypeError

Any suspicions where the extra keyword argument comes from?

I did not run the SARAH tests.

I could open separate issues for these if they substantiate.

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist

• I tested my contribution locally and it seems to work fine.
• I locally ran pytest inside the repository and no unexpected problems came up.
• I have adjusted the docstrings in the code appropriately.
• I have documented the effects of my code changes in the documentation doc/.
• I have added newly introduced dependencies to environment.yaml file.
• I have added a note to release notes doc/release_notes.rst.