Matplotlib update to version 2.0

[khs26]

A placeholder issue for matplotlib changes that might break our existing calls. Not sure exactly which changes might need making and I don't currently have time to test it out.

[smcantab]

my opinion is that matplotlib should typically be wrapped in a try
statement, since one typically does not uses matplotlib unless he is
analysing the data. That way it won't break anything if the import is not
resolved at run time, and one can fix it at a later stage during data
analysis if need to

On Mon, Dec 14, 2015 at 5:08 PM, Kyle Sutherland-Cash <
[email protected]> wrote:

A placeholder issue for matplotlib changes that might break our existing
calls. Not sure exactly which changes might need making and I don't
currently have time to test it out.

—
Reply to this email directly or view it on GitHub
#139.

[khs26]

I agree with that. Especially given that it can take some fiddling with the various graphical backends to get it to work reliably. Does the GUI use matplotlib for things that we need to work reliably?

[smcantab]

yes the gui relies heavily on matplotlib, and as far as I can tell it uses
exclusively the QT4Agg backend. As far as I can tell the whole point of the
gui is that one wants to visualise results, so I don't see how you can
bypass matplotlib when using the gui. What I mean is that if there is a
matplotlib import in the NEB source because there is a test snippet at the
end of the script then that should be wrapped in a try statement because it
is not strictly necessary to run a NEB calculation.

On Mon, Dec 14, 2015 at 5:15 PM, Kyle Sutherland-Cash <
[email protected]> wrote:

I agree with that. Especially given that it can take some fiddling with
the various graphical backends to get it to work reliably. Does the GUI use
matplotlib for things that we need to work reliably?

—
Reply to this email directly or view it on GitHub
#139 (comment).

[smcantab]

re-reading my post I noticed that it might sound a bit passive-aggressive,
it wasn't my intention though! (I was just writing in a bit of a rush) let
me know if I can help on a particular issue related to this

On Mon, Dec 14, 2015 at 5:32 PM, Stefano Martiniani <
[email protected]> wrote:

yes the gui relies heavily on matplotlib, and as far as I can tell it uses
exclusively the QT4Agg backend. As far as I can tell the whole point of the
gui is that one wants to visualise results, so I don't see how you can
bypass matplotlib when using the gui. What I mean is that if there is a
matplotlib import in the NEB source because there is a test snippet at the
end of the script then that should be wrapped in a try statement because it
is not strictly necessary to run a NEB calculation.

On Mon, Dec 14, 2015 at 5:15 PM, Kyle Sutherland-Cash <
[email protected]> wrote:

I agree with that. Especially given that it can take some fiddling with
the various graphical backends to get it to work reliably. Does the GUI use
matplotlib for things that we need to work reliably?

—
Reply to this email directly or view it on GitHub
#139 (comment).