WIP : draft of replacement for pyplot #3587
Conversation
pyplot_nng replaces most of pyplot plotting functions, a few things (like plt.subplots and the other axes/figure generation functions) need to be interactive gives a auto-redraw wrapper for explicitly using axes, but forces you through a functional interface rather than the OO interface.
|
I'm going to reply on the list, not here, because I think it is important that this be discussed among a wider audience than those who happen to subscribe to this Github PR. |
|
Fair enough, I just wanted to get a working version of what I was talking about up someplace readable. |
|
@tacaswell Because it is too big of a change in the way users interact with matplotlib, this could be targeted at V2.0? |
|
@fariza, I think a poll would be premature and not helpful; but the underlying point is correct. Yes, big changes must be designed for the long-term benefit of the user community, and introduced carefully. |
pyplot_nng replaces most of pyplot plotting functions, a few
things (like plt.subplots and the other axes/figure generation
functions) need to be
interactive gives a auto-redraw wrapper for explicitly using axes, but
forces you through a functional interface rather than the OO interface.
This a result of http://thread.gmane.org/gmane.comp.python.matplotlib.devel/12999
On one hand, this is a huge API change (but it can be done breaking back-compatibility as slowly as we want), but I think this will greatly simplify things going forward and solve the pyplot vs OO issues replacing both with a more functional approach (we still need to work out where the edges are, is it
set_yscale(ax, 'log'),ax.yscale('log'),ax.set_yscale('log')orax.yscale = 'log'). The pyplot/interactive repl interface is an automatic wrapping of the functions to handle the global state and automatic re-drawing and the functions take the place of the OO in scripts and user developed functions.This would also let us sort the plotting functions in to modules which might help with the documentation issues.
cc @efiring