Drew Leonard
Drew Leonard
Oh wait that build is passing and it shouldn't be. Never mind.
From a little basic testing so far, the helper function is not resilient against different kinds of WCS. For example, resampling the data before trying to reproject so that it...
This has been at least partially addressed by #370 but I guess we probably want to make further improvements if we can.
Not any more. Now I think the main issues are the last two things on the list above but I'm not sure what the status is of those now.
Could we use sparse arrays: https://docs.dask.org/en/latest/array-sparse.html ?
> Could we use sparse arrays: https://docs.dask.org/en/latest/array-sparse.html ? (Tentatively): yes. PR incoming.
Tagged as Upstream Issue but may be worth testing in user tools first and then pushing up when it's a bit stable.
A similar error occurs when updating a 2D plot if `clip_interval` is set, due to being unable to calculate `vmin` and `vmax` when data is missing.
I have a lot of previous work I can put towards this, which I'm porting onto my fork in bits when I have time, so I've assigned myself to this...
Figure failures are because of using a PR branch of ndcube and should go away once it's merged and we switch back to main.