New interface - Operation, Algorithm, Manifold #255
Conversation
|
@rafaqz this is my current thought, operations are not so fleshed out but you can see where I'm going with this i think |
|
Looks good! |
| @@ -18,7 +18,13 @@ using Tables | |||
| using DataAPI | |||
|
|
|||
| include("keyword_docs.jl") | |||
| include("types.jl") | |||
There was a problem hiding this comment.
Yep, nice to break this into themes
|
|
||
| abstract type Algorithm{M <: Manifold} end | ||
|
|
||
| abstract type ManifoldIndependentAlgorithm{M <: Manifold} <: Algorithm{M} end |
There was a problem hiding this comment.
| abstract type ManifoldIndependentAlgorithm{M <: Manifold} <: Algorithm{M} end | |
| abstract type ManifoldIndependentAlgorithm{NoManifold} <: Algorithm{M} end |
Or something like that?
There was a problem hiding this comment.
I guess it's not really manifold independent but rather manifold invariant - the manifold goes to the lower level function like point_distance_from_line
There was a problem hiding this comment.
Yeah I think needs a better name, to me it implies the answer won't change between manifolds.
| # specifically for manifolds | ||
| abstract type EllipsoidParametrization end | ||
|
|
||
| struct SemimajorAxisInvFlattening{T} <: EllipsoidParametrization |
There was a problem hiding this comment.
Is this instead of the two fields in Geodesic?
There was a problem hiding this comment.
Yeah I was thinking about that, then you could pass e.g. Proj.geod_geodesic output, or one of the CoordRefSys things, or Geodesy.jl ellipsoids to it and they would just work
No description provided.