Add ying yang grid #20
Conversation
|
|
|
|
||
| @property | ||
| def lat_lon(self): | ||
| mesh_x, mesh_y = np.meshgrid(self.x, self.y, indexing='ij') |
There was a problem hiding this comment.
i needed to change this to use indexing=ij here versus the original which was indexing="xy", since meshgrid has unusual behavior. @simonbyrne please review. Specifically, do we want the data to have the shaped [nx, ny] or [ny,nx]?
There was a problem hiding this comment.
I don't really have a preference, I think I just copied LatLongGrid which uses Lat (Y) first.
There was a problem hiding this comment.
Do you need to update the LCC tests?
This PR adds a new projection based grid, the ying yang grid. It restructures some of the lambert conformal logic a bit, so Simon should take a look.
| pass | ||
|
|
||
|
|
||
| class Grid(base.Grid): |
There was a problem hiding this comment.
I just renamed your LCC grid class @simonbyrne . Intended to be used like earth2grid.projections.Grid.
There was a problem hiding this comment.
Not sure if you want to give it a more specific name (i.e. it assumes the underlying grid is rectilinear, not an unstructured mesh).
There was a problem hiding this comment.
I think "projection" implies rectangle. This is just a style preference for projection.Grid over projection.ProjectionGrid. Same style as healpix.Grid.
| pass | ||
|
|
||
|
|
||
| class Grid(base.Grid): |
There was a problem hiding this comment.
Not sure if you want to give it a more specific name (i.e. it assumes the underlying grid is rectilinear, not an unstructured mesh).
|
|
||
| def vec2ang(x, y, z): | ||
| """convert lon,lat in radians to cartesian coordinates""" | ||
| lat = torch.asin(z) |
There was a problem hiding this comment.
This assumes your inputs are already normalized. If you want it to work with non-normalized inputs, you would need
| lat = torch.asin(z) | |
| lat = torch.atan2(z, torch.hypot(y, x)) |
|
|
||
| @property | ||
| def lat_lon(self): | ||
| mesh_x, mesh_y = np.meshgrid(self.x, self.y, indexing='ij') |
There was a problem hiding this comment.
I don't really have a preference, I think I just copied LatLongGrid which uses Lat (Y) first.
|
|
||
| @property | ||
| def lat_lon(self): | ||
| mesh_x, mesh_y = np.meshgrid(self.x, self.y, indexing='ij') |
There was a problem hiding this comment.
Do you need to update the LCC tests?
This PR adds a new projection based grid, the ying yang grid. It restructures some of the lambert conformal logic a bit, so Simon should take a look.