Revert temporary Dask pinning in CI

[jrbourbeau]

With dask=2021.3.1 out we can remove the temporary version pinning in CI to workaround #27. Additionally, this PR also switches to installing packages from conda-forge instead of defaults

[codecov]

Codecov Report

Merging #39 (835e20e) into master (a3b3660) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master      #39   +/-   ##
=======================================
  Coverage   94.41%   94.41%           
=======================================
  Files           3        3           
  Lines         215      215           
  Branches       58       58           
=======================================
  Hits          203      203           
  Misses          8        8           
  Partials        4        4           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a3b3660...835e20e. Read the comment docs.

[jrbourbeau]

The latest Dask release (2021.3.1) dropped support for Python 3.6. This means that all of xhistograms dependencies have now no longer support Python 3.6 so we should probably support as well. I've opened up a PR over in #40 to do this.

[dougiesquire]

Looks good to me. Thanks for doing this @jrbourbeau

[dougiesquire]

Actually, it looks as though some tests are still failing with the old error:

        if reduce(mul, shape, 1) != x.size:
>           raise ValueError("total size of new array must be unchanged")
E           ValueError: total size of new array must be unchanged

/usr/share/miniconda3/envs/xhistogram_test_env/lib/python3.6/site-packages/dask/array/reshape.py:206: ValueError

[jrbourbeau]

Those are due to Dask dropping support for Python 3.6 in 2021.03.1, which the Python 3.6 builds can't install the release which contains the fix for ValueError: total size of new array must be unchanged. Let's hold off on merging this PR until we've dropped Python 3.6 (xref #40), which should resole the issue altogether

[dougiesquire]

Thanks for this @jrbourbeau, merging now

[jrbourbeau]

Thanks @dougiesquire!