[Biopython-dev] Should we recommend "pip install biopython"?

[Peter Cock]

On Tue, Oct 27, 2015 at 3:38 PM, Eric Talevich <eric.talevich at gmail.com> wrote:
> On Fri, Oct 23, 2015 at 8:17 AM, Peter Cock <p.j.a.cock at googlemail.com>
> wrote:
>>
>> Actually installing pip on older Pythons is a pain, so I'd like any
>> wording to make it clear pip may not be installed, in which case
>> the traditional "python setup.py ..." route is the best bet.
>
> Sure. The Python community has been fairly clear for years that pip is the
> way forward. As long as we're being good citizens, recommending Python 3.5
> and so on, we should suggest a sane setup as the default one for Biopython's
> users.

So suggesting Python 3.5, and "pip install biopython --user" or similar
as first port of call?

> I haven't used Windows in about a decade, but on Linux in recent years pip
> has always been available to me. I can vouch for conda, too; whatever the
> usual problems are on Windows, I think conda is meant to fix/avoid them.

Given conda/anaconda already package Biopython, that's a good option.
Although we may want to ping them for Biopython 1.66, currently they
only have up to Biopython 1.65: https://anaconda.org/anaconda/biopython

conda install -c https://conda.anaconda.org/anaconda biopython

Peter