Port over to conda

[tnigon]

Hi, this package is not on the Anaconda Cloud platform, so I've begun to port it over (on the conda-forge channel). Please comment on this pull request your willingness to be listed as a maintainer of the PyErf package on conda-forge. Thanks!

[dougthor42]

Done!

I don't use conda myself and have never added a package to it, but I do have a rough idea of what it is and why it's used.

Questions:

1. Are there any special things I need to do in the future if I release new versions?
2. Should I update the README in some way?
3. Do you have any other tips/suggestions for proper conda integration?

[tnigon]

Great, thanks! This package is a dependency of a dependency of a package I'd like to be available on the Anaconda Cloud, so thanks.

1. Whenever a new version is released, we just have to use conda-build to be sure the new version can be built on the Anaconda environments (this is what the meta.yaml "recipe" is for). I think the protocol is a bit different for updating a package compared to uploading for the first time. I believe a separate repository will be created for this package on the conda-forge Github page, and I think we can update by pushing a new meta.yaml to that repo.
2. Once accepted to the Anaconda Cloud, you might update the readme to state that the package can also be installed via conda (once we get to that point), but it isn't absolutely necessary.
3. I am new to conda integration myself, so I'm definitely still learning. No specific suggestions at this point...

[tnigon]

PyErf was just merged to the conda-forge channel. There is now a pyerf conda-forge organization, a conda-forge feedstock, and the built package on Anaconda Cloud.

To install with conda:

conda install -c conda-forge pyerf

[dougthor42]

Great!

FYI: I'm reopening this Issue to remind myself to update documentation.