PyPlot: How to change the aspect ratio of a plot?

[Daniel Carrera]

Hello,

I am using PyPlot and I cannot figure out how to change the aspect ratio of a plot. I googled for some answers and I found a thread from Python users about aspect ratios in Matplotlib:

http://stackoverflow.com/questions/7965743/how-can-i-set-the-aspect-ratio-in-matplotlib

I am trying to make sense of that thread, and imagine how to do something similar in Julia, but so far my attempts have failed. The closest I got is to make a "subplot", but I don't know how to plot inside a subplot. Here is what I tried:

julia> using PyPlot

INFO: Loading help data...

julia> aspect

ERROR: aspect not defined

julia> ax = axes()

PyObject <matplotlib.axes.AxesSubplot object at 0x7f5c378fc4d0>

julia> methods(ax)

ERROR: `methods` has no method matching methods(::PyObject)

julia> fig = figure(2)

Figure(PyObject <matplotlib.figure.Figure object at 0x7f5c376cda10>)

julia> methods(fig)

ERROR: `methods` has no method matching methods(::Figure)

julia> subplot(111, aspect=3)

PyObject <matplotlib.axes.AxesSubplot object at 0x7f5c376d3090>

AHA! So the last line correctly creates a plot with the desired aspect ratio. But I don't know where to go from there. The moment I try to plot anything, the aspect ratio resets. I do have the basic idea of how subplots work:

--------------------------

using PyPlot

x = linspace(0,2pi);

subplot(211)

plot(x,sin(x))

subplot(212)

plot(x,cos(x))

--------------------------

So I was expecting to be able to do this:

--------------------------

using PyPlot

x = linspace(0,2pi);

subplot(111, aspect=3)

plot(x,sin(x))

--------------------------

I was expecting that this would give me a plot that looks like a very tall sine wave. But it doesn't. The moment I run the plot() command, the aspect ratio resets.

Can anyone help me?

Cheers,

Daniel.

[Steven G. Johnson]

The answers in the stackoverflow thread you linked work for me, e.g.

w, h = plt[:figaspect](0.5)

figure(figsize=(w,h))

plot(rand(10))

[Daniel Carrera]

I often have trouble converting from Python+Matplotlib to Julia+PyPlot. Even with your code example, I don't really know what to do. Where do I get that "plt" variable? For future reference, how would I know that the figaspect() function would be located in "plt[:figaspect]". Is there a general rule I can follow?

I'm sure you are busy, but is there a chance you might write a tutorial for PyPlot, showing how to convert Python code one finds around the web into equivalent Julia code?

Thanks for the help.

Cheers,

Daniel.

[Nils Gudat]

https://github.com/stevengj/PyPlot.jl#exported-functions

[Nils Gudat]

FWIW, I always think the fig/ax syntax aligns best with Python, e.g.

fig, ax = PyPlot.subplots(2,2, ...)
ax[1,1][:plot](..., linestyle = ":", label = "...")
ax[1,1][:legend](loc="best")
ax[1,2][:hist](..., alpha = 0.5, label = "...")
fig[:title]("Some title")

with this syntax, you can more or less convert Python code to Julia 1-for-1 by just replacing all the dot method calls with [:method]-calls.

[Daniel Carrera]

Thanks. That is very helpful.

Cheers,

Daniel.