Fix field plotting #1247
Fix field plotting #1247
Conversation
…r replication and fix #741
…issue coming from different cartopy versions. Solved it via try-except clauses.
….6.0 integration test failure
….0 integration test failure
….0 integration test failure
| # ax = fig.gca(projection='3d') | ||
| ax = fig.gca() |
There was a problem hiding this comment.
I don't understand this change. I think this is what breaks the unit-tests. The ax is now not set as 3D anymore, so there is no set_zlevel method. Revert back to original?
There was a problem hiding this comment.
The change was necessary cause that line broke bccb214. In short: in Windows, the workflow obtains a newer version of matplotlib, and in matplotlib 3.6.0, the function figure.gca() has no projection parameters. Unexpected change, but therefore I needed to make this projection option part of the figure creation procedure. I guess help would be welcome from bccb214 onwards.
Since this version still worked, and 2022.9.2 gives an error with https://github.com/OceanParcels/parcels/actions/runs/3180136958/jobs/5199374392
| @@ -28,4 +28,5 @@ dependencies: | |||
| - nbval | |||
| - scikit-learn | |||
| - pykdtree | |||
| - cartopy | |||
There was a problem hiding this comment.
Why is cartopy now need? That was not the case before, and I suggest not to keep it
| - cartopy |
There was a problem hiding this comment.
that will make the new test fail that actually tests platecarree plotting
| - cftime>=1.3.1 | ||
| - ipykernel | ||
| - pytest | ||
| - nbval | ||
| - pykdtree | ||
| - cartopy |
There was a problem hiding this comment.
See above, I suggest not to make cartopy an explicit dependency (as we had not done before)
| - cartopy |
| @@ -101,7 +102,7 @@ def plotparticles(particles, with_particles=True, show_time=None, field=None, do | |||
|
|
|||
|
|
|||
| def plotfield(field, show_time=None, domain=None, depth_level=0, projection='PlateCarree', land=True, | |||
| vmin=None, vmax=None, savefile=None, **kwargs): | |||
| vmin=None, vmax=None, savefile=None, use3D=False, **kwargs): | |||
There was a problem hiding this comment.
Is this variable still needed now? I am a bit confused what this does
| from parcels import FieldSet, ParticleSet, JITParticle, AdvectionRK4, ErrorCode | ||
| from datetime import timedelta as delta | ||
| import numpy as np | ||
| import dask.array as da # NOQA | ||
| from os import path | ||
| from matplotlib import pyplot as plt # NOQA | ||
| import pytest | ||
|
|
||
|
|
||
| def periodicBC(particle, fieldset, time): | ||
| if particle.lon > 180: | ||
| particle.lon -= 360 | ||
|
|
||
|
|
||
| def test_field_from_netcdf(): | ||
| data_path = path.join(path.dirname(__file__), 'test_data/') | ||
|
|
||
| filenames = {'U': {'lon': data_path + 'mask_nemo_cross_180lon.nc', | ||
| 'lat': data_path + 'mask_nemo_cross_180lon.nc', | ||
| 'data': data_path + 'Uu_eastward_nemo_cross_180lon.nc'}, | ||
| 'V': {'lon': data_path + 'mask_nemo_cross_180lon.nc', | ||
| 'lat': data_path + 'mask_nemo_cross_180lon.nc', | ||
| 'data': data_path + 'Vv_eastward_nemo_cross_180lon.nc'} | ||
| } | ||
| variables = {'U': 'U', | ||
| 'V': 'V'} | ||
| dimensions = {'lon': 'glamf', 'lat': 'gphif'} | ||
| return FieldSet.from_netcdf(filenames, variables, dimensions, interp_method='cgrid_velocity', chunksize='auto', allow_time_extrapolation=True) | ||
|
|
||
|
|
||
| @pytest.fixture(name="fieldset") | ||
| def fieldset_fixture(): | ||
| return test_field_from_netcdf() | ||
|
|
||
|
|
||
| def test_pset_create_field(fieldset, npart=100): | ||
| lonp = -180 * np.ones(npart) | ||
| latp = [i for i in np.linspace(-70, 88, npart)] | ||
| pset = ParticleSet.from_list(fieldset, JITParticle, lon=lonp, lat=latp) | ||
| return pset | ||
|
|
||
|
|
||
| def DeleteParticle(particle, fieldset, time): | ||
| particle.delete() | ||
|
|
||
|
|
||
| if __name__ == '__main__': | ||
| fset = test_field_from_netcdf() | ||
| print(fset) | ||
| pset = test_pset_create_field(fset) | ||
| kernels = pset.Kernel(AdvectionRK4) + periodicBC | ||
| pset.execute(kernels, dt=delta(hours=1), output_file=None, | ||
| recovery={ErrorCode.ErrorOutOfBounds: DeleteParticle}) | ||
| pset.show(field=fset.U) |
There was a problem hiding this comment.
What does this new set of functions test that was not tested before? Why is it needed?
Co-authored-by: Erik van Sebille <e.vansebille@uu.nl>
|
I think that this was now fixed in 0326f77 |
Added mini-test for loading and plotting NEMO data with dask for error replication and fix #741
Comment: little clean-up of open branches. This one is a small addition adding more stability to the plotting by checking that plotting works correctly. Valuable addition with little extra hassle - hence: merge-request.