grass.jupyter: fix resolution handling for InteractiveMap

[petrasovaa]

Should address #2369. Now when use_region=True or saved_region, it should reproject the map with the correct resolution. Note that in these cases the resolution is not estimated by r.proj but is estimated by reprojecting the extent and using the same ncols/nrows as in the source region, which might not be the most precise way.

[neteler]

I have tested this with my larger dataset:

Then, when displaying it. I get:

Checking in parallel what is run underneath, with ps aux | grep proj I see:

r.proj input=opennrw_lidar_bonn_p95_50cm_DOM@bonn_lidar output=opennrw_lidar_bonn_p95_50cm_DOM_bonn_lidar mapset=bonn_lidar location=openNRW_epsg25832 dbase=/home/mneteler/grassdata resolution=0.8094142783201477

What I notice is that I set the computational region to res=2m while r.proj uses resolution=0.809 which is resulting in way too many pixels. Possibly my mistake?

[petrasovaa]

Try use_region=True

[neteler]

Try use_region=True

Ok, this then changes the resolution (not sure how to inspect the bbox):

r.proj input=opennrw_lidar_bonn_p95_50cm_DOM@bonn_lidar output=opennrw_lidar_bonn_p95_50cm_DOM_bonn_lidar mapset=bonn_lidar location=openNRW_epsg25832 dbase=/home/mneteler/grassdata resolution=3.21940318

And it helps:

The question is now what should be the default: use_region=True or not, as currently.
Perhaps catch the error above and in case suggest to the user to add use_region=True?

[veroandreo]

I added grammar suggestions in the documentation paragraph :)

[wenzeslaus]

The code and behavior make sense to me.