Add a test for plotting global grid without the redundant 360 longitude

[seisman]

Description of proposed changes

This PR adds a test for the bug reported in #3331. See #3331 (comment) for the comprehensive tests for when the bug can be triggered. This PR only test the central longitude=120 case. The baseline image is created by saving da3/da4 into a netCDF file and passing the netCDF files directly to Fig.grdimage.

The new test only passes with the upstream fix in GenericMappingTools/gmt#8554. So, the test is marked as xfail for GMT<=6.5.0.

The test passes in the "GMT Dev Tests" workflow: https://github.com/GenericMappingTools/pygmt/actions/runs/11138377051?pr=3358

Closes #3331.

[github-actions]

Summary of changed images

This is an auto-generated report of images that have changed on the DVC remote

Status	Path
added	pygmt/tests/baseline/test_grdimage_grid_no_redunant_360.png

Image diff(s)

Added images

• test_grdimage_grid_no_redunant_360.png

Modified images

Path	Old	New

Report last updated at commit 86b6d7d

[weiji14]

It's not a real global grid, which doesn't have the 360° longitude defined and also the north/south poles are missing. I'll open a separate PR to refine the xrgrid into a true global grid, so mark this PR "Draft" now.

Maybe best if we keep the original xrgrid that has data from 0-359° longitude (no redundant data at 360°), and a new grid from 0-360° longitude (redundant data at 360°). There isn't really a correct 'real' global grid here, we just need to know GMT/PyGMT can handle both.

[seisman]

It's not a real global grid, which doesn't have the 360° longitude defined and also the north/south poles are missing. I'll open a separate PR to refine the xrgrid into a true global grid, so mark this PR "Draft" now.

Maybe best if we keep the original xrgrid that has data from 0-359° longitude (no redundant data at 360°), and a new grid from 0-360° longitude (redundant data at 360°). There isn't really a correct 'real' global grid here, we just need to know GMT/PyGMT can handle both.

Actually the original issue is not due to the missing 360 longitude. For any grids that is not exactly 0-360, the same issue occurs for central longitude < 180 and for specific projections. If interested, you can play with the script in #3331 (comment).