Non-interactive mode about pyplot.jl HOT 8 CLOSED

If you just want to block, why not do:

print("Hit <enter> to continue")
readline()

But it would be nice to provide a way to load PyPlot in non-interactive mode. (As I said on StackOverflow, if you just want non-interactive mode you could always @pyimport matplotlib.pyplot as plt directly...most of reason for PyPlot is to provide interactive graphics and IJulia integration. But you lose a few niceties of PyPlot that way.)

from pyplot.jl.

On a similar note.

I have a long running simulation and I would like to periodically update a Figure to show the progress of certain variables. Currently only the last plot will be updated and I also don't want to block.

My code looks like this

while t < largeNumber
   do calculation
   if t % interval
      plot([1:largeNumber], variable)
   end
   t += 1
end

Adding draw add the end of the call doesn't force an update.

from pyplot.jl.

Have you tried calling yield() occasionally so that the GUI event loop can run?

from pyplot.jl.

Thank you that was exactly what I needed

from pyplot.jl.

Any ideas how one could do that without yield? It seems to have pretty big performance penalty in my case attached to it. According to IProfile a quarter of my program runtime is spent waiting on yield.

It would be great if one could run the pyplot updates async to the main task.

from pyplot.jl.

Don't yield on every iteration of your inner loop. Just yield every 50ms or so.

from pyplot.jl.

@pyimport matplotlib.pyplot as plt to get non-interactive mode is quite useful for batch processing of data. would be great to have this documented. what "niceties" of PyPlot are lost by doing this?

from pyplot.jl.

@bjarthur, you lose the more Julian style of just typing plot (or PyPlot.plot) rather than having to qualify everything with plt., the LaTeXStrings import, the colormap conversions, some improved consistency of array-dimensionality handling, etc.

It should be straightforward to modify PyPlot to check the isinteractive() function to see if Julia is running in non-interactive mode, and if not then to initialize the pyplot module in non-interactive mode (skipping all of the GUI event-loop stuff), if anyone wants to take a crack at it.

from pyplot.jl.