Support inset_axes zooms for all GeoAxes projections #133

Activity

1. lukelbd

   added

   feature

   on Apr 22, 2020
2. lukelbd commented on Apr 22, 2020

   

   lukelbd

   on Apr 22, 2020

   Last edited by lukelbd

   Collaborator

   More actions

   This is actually not an advertised feature but it's on my to-do list under the category "make XYAxes features work with rectangular projections like Mercator and Plate Carrée". See #131 for details.

   Will leave this issue open as a feature request for now. As a workaround you can probably make your own inset axes using the default cartopy API with e.g.

   import cartopy.crs as ccrs
   ax = fig.add_axes([x0, x1, y0, y1], projection=ccrs.PlateCarree())
   ax.set_extent([lon0, lon1, lat0, lat1])
   <plotting_code_here>

   👍React with 👍1Reacted by Xin Zhang
3. lukelbd

   closed this as completedin #172on May 19, 2020
4. zxdawn commented on Jul 30, 2021

   

   zxdawn

   on Jul 30, 2021

   Author

   More actions

   @lukelbd

   I tried this feature again:

   import proplot as pplt
   import cartopy.crs as ccrs
   
   pplt.rc.reso = 'med'
   fig, axs = pplt.subplots(proj='pcarree')
   axs.scatter(120, 30)
   
   bounds = (1, 1, 1, 1)
   # bounds = (118, 28, 4, 4)
   
   axs.inset(bounds, proj='pcarree',
             zoom=True,
            zoom_kw={'color': 'red3', 'lw': 2, 'ls': '--'}
            )
   
   axs.format(coast=True, lonlim=(100, 130), latlim=(15, 40))
   

   

   The bounds are relative to axis, is it possible to use lon and lat?
   If I set transform='data', I got this error:

   ~/new/miniconda3/envs/pyresample_min/lib/python3.8/site-packages/matplotlib/cbook/__init__.py in _check_isinstance(_types, **kwargs)
      2244     for k, v in kwargs.items():
      2245         if not isinstance(v, types):
   -> 2246             raise TypeError(
      2247                 "{!r} must be an instance of {}, not a {}".format(
      2248                     k,
   
   TypeError: 'transform' must be an instance of matplotlib.transforms.Transform, not a cartopy.crs.PlateCarree
   
5. lukelbd commented on Jul 30, 2021

   

   lukelbd

   on Jul 30, 2021

   Last edited by lukelbd

   Collaborator

   More actions

   In v0.7 I try to translate transform='data' into transform=ccrs.PlateCarree() so that you can position with longitude-latitude bounds rather than map projection bounds... but I guess matplotlib's axes positioning system can't handle cartopy transforms like its plotting commands can.

   You can use transform=ax.transData instead for now -- since you're using the Plate Carrée projection, "map projection coordinates" are equivalent to longitude-latitude coordinates
6. lukelbd

   added

   bug

   and removed

   feature

   on Jul 30, 2021
7. lukelbd

   reopened this on Jul 30, 2021
8. zxdawn commented on Jul 30, 2021

   

   zxdawn

   on Jul 30, 2021

   Last edited by zxdawn

   Author

   More actions

   @lukelbd Thanks, here's the example:

   import numpy as np
   import proplot as pplt
   import cartopy.crs as ccrs
   
   pplt.rc.reso = 'med'
   fig, axs = pplt.subplots(proj='pcarree')
   axs.scatter(120, 30)
   
   bounds = (120, 18, 10, 10)
   
   iax = axs.inset(bounds, proj='pcarree',
                   transform=axs.transData,
                   zoom=True,
                   zoom_kw={'color': 'red3', 'lw': 2, 'ls': '--'}
                  )
   iax.scatter(120, 30)
   
   iax.format(coast=True, lonlim=(118, 122), latlim=(28, 32), labels=True, dms=False)
   axs.format(coast=True, lonlim=(100, 130), latlim=(15, 40))
   

   However, grid labels are wrong and no "zoom in" lines are plotted:
   
9. lukelbd

   added this to the Version 0.8 milestone on Jul 30, 2021
10. lukelbd

    modified the milestones: Version 0.8, Version 0.9 on Aug 19, 2021
11. lukelbd commented on Jan 24, 2022

    

    lukelbd

    on Jan 24, 2022

    Last edited by lukelbd

    Collaborator

    More actions

    This seems to work pretty well for Plate Carree projections in the latest version. Have to set manually set the left/bottom spaces or else the result is wrong... but not too much effort.

    import proplot as pplt
    import cartopy.crs as ccrs
    pplt.rc.reso = 'med'
    fig, axs = pplt.subplots(
        proj='pcarree', coast=True, tight=True, bottom=7, left=1
    )
    fig.format(lonlim=(0, 180), latlim=(0, 90))
    axs.scatter(120, 30)
    bounds = (0.8, -0.6, 0.8, 0.8)
    iax = axs.inset(bounds, proj='pcarree', zoom=True, zoom_kw={'ec': 'red', 'lw': 2})
    iax.format(coast=True, rivers=True, lonlim=(100, 130), latlim=(15, 40))
    fig.show()

    

    Non-Plate Carree are a different story thought. Again matplotlib isn't supplying the correct bounds:

    import proplot as pplt
    import cartopy.crs as ccrs
    
    pplt.rc.reso = 'med'
    fig, axs = pplt.subplots(
        proj='robin', coast=True, tight=True, bottom=7, left=1
    )
    fig.format(lonlim=(0, 180), latlim=(0, 90))
    axs.scatter(120, 30)
    bounds = (0.8, -0.6, 0.8, 0.8)
    iax = axs.inset(bounds, proj='pcarree', zoom=True, zoom_kw={'ec': 'red', 'lw': 2})
    iax.format(coast=True, rivers=True, lonlim=(100, 130), latlim=(15, 40))

    

    Since the issue with using the wrong "transform" is fixed I'll remove the "bug" tag.
12. lukelbd

    added

    enhancement

    and removed

    bug

    on Jan 24, 2022
13. lukelbd

    changed the title [-]inset_axes for geographic plots[/-] [+]Support inset_axes zooms for all GeoAxes projections[/+] on Jan 24, 2022
14. lukelbd

    added

    feature

    on Jan 24, 2022