Thanks a lot Nathan … I think I can  make due by generating separate plots and then pasting them together manually.

Good to know that it wasn’t something dumb I was doing :)


Rick Sarmento
SESE Astronomy/Astrophysics Grad Student

On Nov 4, 2014, at 11:34 PM, Nathan Goldbaum <> wrote:

Hey Rick,

On Tue, Nov 4, 2014 at 9:21 PM, Rick Sarmento <> wrote:

Hello all,

I’m having trouble getting yt to plot 2 gas-phase diagrams side-by-side in a grid. I can do this fine for a projection plot, phase plots seems to behave differently. Here’s the code — any idea how to get these two plots side by side sharing the same scale/color-bar?

# fractional=True keyword -- divide the profile data by the sum. 
fns = [ "/Users/earnric/RAMSES-Data/coolingFixedTau/output_00010/info_00010.txt",

fig = plt.figure()
grid = AxesGrid(fig, (0.075,0.075,0.85,0.85),
                nrows_ncols = (1, 2),
                axes_pad = 0.05,
                label_mode = "L",
                share_all = True,
fig.suptitle("Gas Phase")

for i, fn in enumerate(fns):
    ds = yt.load(fn)
    ad = ds.all_data()
    p = yt.PhasePlot(ad, "density", "temperature", "cell_mass", x_bins=256, y_bins=256, weight_field=None, fractional=True )
    plot = p.plots['cell_mass']
    if i % 2 == 0:
        ptype = "Gas phase ${H_2}$ cooling "
        ptype = "Gas phase Standard cooling "
    zStr = ptype + "z = %.1f" % ds.current_redshift 
    plot.figure = fig
    plot.axes = grid[i].axes
    plot.cax = grid.cbar_axes[i]
    #plot.annotate_title(zStr) # YOU HAVE TO DO THIS AFTER THE AXES calls

    p['cell_mass'].axes.set_ylim (1e-1, 1e6)
    p['cell_mass'].axes.set_xlim (1e-29, 1e-24)
    p['cell_mass'].axes.set_title(zStr, fontsize=20)

FWIW, this is the first I've ever heard someone trying to do this. I can confirm that it doesn't work.

I tried experimenting a bit and unfortunately I think you're running into a limitation of matplotlib: AxesGrid doesn't support logarithmic x and y axes.

I can get something sort of sensible if I specify linear x and y axes and manually adjust the aspect ratio, but I don't think this is a particularly sustainable solution for supporting these sorts of plots:

Unfortunately when I try to make the same plot using logarithmic axes, I get junk:

I think the most straightforward way to do this would be to do it 100% manually (outside of PhasePlot) using a Profile2D object and matplotlib.  If you figure out something that works for you I think it would be a good addition for the cookbook. 

It might be possible to use PhasePlot with a hand-constructured set of axes, but I've never tried to set that up so I'm not sure if you'll need to modify yt to get it to work.

Sorry I don't have better news for you - good luck with getting this plot working.



Rick Sarmento

yt-users mailing list

yt-users mailing list