On Thu, Feb 18, 2016, Ross Walker wrote:
> I have
> many many examples for the machine learning space where one has to use
> Python virtual environments for each package or you trash everything
> else on the machine.
Seems to me that this is a good argument to push *most* users to run
the configure_python script, which creates a local python (inside of
$AMBERHOME) that has everything Amber needs, and doesn't mess up any other
python stuff, or require root access; (or indeed, require any knowledge
of python at all.)
A handful of very knowledgeable pythonistas (say, who are switching back and
forth between python2 and python3 all the time, and are heavily involved in
developing the python-based connections between Amber and Phenix, pyRosetta,
openMM, etc.) will want to use their favorite python environments. But the
rest of the world (including Ross and me...) are probably better off using the
configure_python script. For that, you don't have to worry about which
version of python your Linux distro comes with, or whether Amber's python
needs are compatible with those of Google, etc.
> Note, while the above is good it does have deficiencies if one doesn't
> have network access (or limited network access since it downloads a
> LOT!)
I think this concern is overblown. The Amber download itself is many times
larger than the miniconda download (and we have cut the size of the miniconda
download quite a bit from its earlier version.) It's true that one *could* get
caught on an airplane when running configure, but that is unlikely to be the
case for the vast majority of our users.
....dac
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Thu Feb 18 2016 - 15:30:03 PST
