RichardScottOZ

I will try a generic


```python grid = pv.UniformGrid() grid.dimensions=np.array( [25, 25, 24] ) + 1 grid.origin= (100000,7000000,-3400) grid.spacing = (200,200,200) p = pv.Plotter(notebook=True) p.add_mesh(mesh=grid) p.show_grid(color='black') p.set_background(color='brown') p.screenshot(r'BananaSplits.png') p.deep_clean() ```

so maybe have to check those exist first?

I will try and give that a shot this afternoon, have to go out shortly.

Still definitely leaking.

Thanks Wesley, will see if I can try it on the weekend.

Thanks for the tips - currently using 1.8.18 [I think], helped with a file not recognised error.

It took a few hours with the rio commandline to warp a 40GB vrt dataset to a set bounds and 400GB upscale - be interesting to see how the dask/rioxarray...

> Related: > > * [Port xr_reproject from odc.algo opendatacube/odc-geo#26](https://github.com/opendatacube/odc-geo/issues/26) > * [Dask resampler and gradient search overhaul pytroll/pyresample#341](https://github.com/pytroll/pyresample/pull/341) Yes, looking forward to trying the odc-geo version out sometime.