Publish new versions or discontinue pyfolio package on anaconda

[suitablyquantified]

I spent a few hours tearing my hair out trying to figure out why I had anomalous results in Pyfolio.

As it turned out I installed Pyfolio fron anaconda, rather than doing a pip install.

The Quantopian channel on anaconda is horribly out-of-date (shows v0.5.1 released 3.5 years ago): https://anaconda.org/Quantopian/pyfolio/files

Could devs either maintain the packages on anaconda or, alternatively, just remove the package from anaconda, and have it only installed by pip instead.

[twiecki]

Sorry for the frustration, that is really out-of-date. CC @richafrank, I think removing this from our conda channel is a fine solution.

[willianpaixao]

Related #342 #560 #562

Any update on this issue, @twiecki @richafrank?

If the package are not being updated to conda-forge, I assume most people fetch it from PyPI, right?

[lionelyoung]

Most people I talk to who are on pandas >= 1.0 install the master branch via github, since PyPi release is at April 2019

(Pandas 1.0 release in Jan 2020)

pip install git+https://github.com/quantopian/pyfolio.git

[MattTriano]

If a package isn't on conda-forge, it makes me weigh how much I need the package, as using pip to install a package in a conda env can break the conda env. Per the conda documentation, conda is unaware of the dependencies of packages installed via pip so the conda documentation advises that users should basically make a new env if they need to install any packages after installing something via pip.