implementing forecast.plot for quadtree grid (Handling 'dh')

[khawajasim]

I am working on making forecast.plot() compatible for single-resolution quadtree grids. In quadtree grid, we do not have any fixed 'dh' value. So I request a little modification in the code for replacement of dh.

In file: utils.plot, line: 863-865
bbox = region.get_bbox()
if extent is None and not set_global:
extent = [bbox[0], bbox[1], bbox[2] + region.dh, bbox[3] + region.dh]

and lines: 933-934:
lons, lats = numpy.meshgrid(numpy.append(region.xs, region.xs[-1] + region.dh), numpy.append(region.ys, region.ys[-1] + region.dh))

In above cases, dh is only used to get the last corner point of the testing region.
My suggestion is that we can change this implementation by using region.get_bbox(). Currently, this function provides bounding box based on origin points only. I suggest to change its implementation from just origins to whole testing region (origin coords + top_right coords), like I did for Quadtree for gird. Then with this modification, we can simply change above lines to:

bbox = region.get_bbox()
if extent is None and not set_global:
extent = [bbox[0], bbox[1], bbox[2], bbox[3]]

and

lons, lats = numpy.meshgrid(numpy.append(region.xs, bbox[1]), numpy.append(region.ys, bbox[3]))

[wsavran]

@khawajasim can you make a pull request for this change if it's still needed?