Update install docs

[loforest]

I created a pysteps environment from the yml file and installed pysteps in development mode (pip -e). When importing pysteps I got this error: ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject As suggested by @dnerini, the error was solved by starting from an empty environment instead of the one from yml. I took the opportunity to update the installation/contribution docs and add some information here and there.

[codecov[bot]]

Codecov Report

Merging #206 (c0368a7) into master (6d9f0d8) will not change coverage. The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #206   +/-   ##
=======================================
  Coverage   78.01%   78.01%           
=======================================
  Files         131      131           
  Lines        9640     9640           
=======================================
  Hits         7521     7521           
  Misses       2119     2119           

Flag	Coverage Δ
unit_tests	78.01% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

---

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 6d9f0d8...910df34. Read the comment docs.

[aperezhortal]

Hi Loris, I like the idea of including in the documentation a workaround just in case that using the enviroment.yml file does not work. I left a small comment above about possible issues mixing conda and pip.

I couldn't reproduce the error on my computer, but the windows-py3.8 tests are failing with the error that you mentioned. Maybe this is related to the issues in the numpy-feedstock in conda-forge with the failed tests. Perhaps this error goes away with a newer NumPy minor or build release.

[aperezhortal]

Thanks Loris for updating the docs. Now all the tests are passing (maybe numpy was updated).