Axis sharing doesn't work for geo-axis

[zxdawn]

Description

span and share don't work for geographic axis.

Steps to reproduce

import proplot as plot
f, axs = plot.subplots(nrows=2, ncols=2,
                       spany=3, sharey=3,
                       proj='lcc', proj_kw={'lon_0': 116}
                       )
axs.format(coast=True,
           latlines=1, lonlines=1,
           labels=True,
           lonlim=(113, 119),
           latlim=(37, 41.5),
           )

Expected behavior:

Lat labels are shared.

Actual behavior:

Equivalent steps in matplotlib

import matplotlib.pyplot as plt
import cartopy.crs as ccrs

proj = ccrs.LambertConformal(central_longitude=116)
fig, axs = plt.subplots(nrows=2, ncols=2,
                        subplot_kw={'projection': proj})

for idx,ax in enumerate(axs.flatten()):
    ax.set_extent([113, 119, 37, 41.5], ccrs.PlateCarree())
    gl = ax.gridlines(crs=ccrs.PlateCarree(), draw_labels=True,
                      color='gray', alpha=0.5, linestyle='--',
                      x_inline=False, y_inline=False)
    gl.xlabels_top = False
    gl.ylabels_right = False
    if idx%2 != 0:
        gl.ylabels_left = False
    ax.coastlines()

Proplot version

0.6.1

[lukelbd]

There's actually an open issue for this: #79

[lukelbd]

Might not get to #79 for a while because it's kind of a tricky problem (dealing with different projections, axis labels on different sides of the subplots, etc). But for the time being you can just toggle labels for the edge subplots like so:

import proplot as plot
f, axs = plot.subplots(
    nrows=2, ncols=2,
    spany=3, sharey=3,
    proj='lcc', proj_kw={'lon_0': 116}
)
axs.format(
    coast=True,
    latlines=1, lonlines=1,
    lonlim=(113, 119),
    latlim=(37, 41.5),
)
axs[:, 0].format(labels=True)
axs[-1, :].format(labels=True)

[zxdawn]

Nice trick. But, it would change the xtick label and ytick label at the same time.

I'm trying to get the gridline from the GeoAxes and set some to False:

import proplot as plot
f, axs = plot.subplots(
    nrows=2, ncols=2,
    spany=3, sharey=3,
    proj='lcc', proj_kw={'lon_0': 116}
)
axs.format(
    coast=True,
    latlines=1, lonlines=1,
    lonlim=(113, 119),
    latlim=(37, 41.5),
    labels=True,
)

axs[1].gridlines().left_labels = False

However, it doesn't work and the lines are thicker.

[lukelbd]

My bad, you’re right. What you want is: axs[:, 0].format(latlabels=True) axs[-1, :].format(lonlabels=True) See the full documentation for GeoAxes.format() <https://proplot.readthedocs.io/en/latest/api/proplot.axes.GeoAxes.format.html#proplot.axes.GeoAxes.format>.

…

On May 24, 2020, at 7:56 PM, Xin Zhang ***@***.***> wrote: Nice trick. But, it would change the xtick label and ytick label at the same time. I'm trying to get the gridline from the GeoAxes and set some to False: import proplot as plot f, axs = plot.subplots( nrows=2, ncols=2, spany=3, sharey=3, proj='lcc', proj_kw={'lon_0': 116} ) axs.format( coast=True, latlines=1, lonlines=1, lonlim=(113, 119), latlim=(37, 41.5), labels=True, ) axs[1].gridlines().left_labels = False However, it doesn't work and the lines are thicker. <https://user-images.githubusercontent.com/30388627/82771204-cf9cb900-9e6d-11ea-867c-4aad0a2f3ee9.png> — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#178 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEV7HNBXJGYOEBOK54R3GRDRTHF3HANCNFSM4NIXNT2A>.

[zxdawn]

Thanks, Luke!

It works well:

However the small space between tick and tick labels can't be changed using tick.pad and grid.pad.

[lukelbd]

I believe that issue is related to SciTools/cartopy#1556. The padding is dependent on the DPI of your saved PNG, because someone coded the padding as "dots" (pixels on your screen) instead of "points" (1/72 inches) in the cartopy source code. And proplot makes the default dpi for saved figures rc['savefig.dpi'] = 1200, much larger than the backend dpi rc['figure.dpi'] = 100, because 1200 dots per inch is recommended by most academic journals.

If you set the padding to a very large number, reduce the DPI with e.g. f.save('file.png', dpi=100), or save the file as PDF, this will change the padding.

Honesty if grid labels are important to you I recommend installing basemap and using basemap=True in your call to subplots(). Cartopy still has a lot of kinks to work out. For example:

proj = plot.Proj('lcc', lon_0=118, lonlim=(113, 119), latlim=(37, 42), basemap=True)
fig, axs = plot.subplots(ncols=2, nrows=2, proj=proj)
axs[:, 0].format(latlabels=True)
axs[-1, :].format(lonlabels=True)
...

[zxdawn]

Thanks! Hope that issue will be fixed soon.

The reason why I don't use Basemap is the KeyError: KeyError: 'PROJ_LIB'.

This problem doesn't exist in Linux Terminal, but shows up in the editor like sublime text and pycharm (stackoverflow link):

import os
os.environ["PROJ_LIB"] = '/home/xin/Software/miniconda3/envs/s5p/share/proj'

[lukelbd]

@zxdawn FYI SciTools/cartopy#1556 was just merged into cartopy's master branch