Closed filtron closed 3 months ago
Attention: Patch coverage is 20.00000%
with 8 lines
in your changes missing coverage. Please review.
Project coverage is 91.93%. Comparing base (
44e6544
) to head (7bf0ecd
). Report is 7 commits behind head on main.
Files | Patch % | Lines |
---|---|---|
src/distributions/uvnormal.jl | 0.00% | 7 Missing :warning: |
src/covariance_parameter.jl | 0.00% | 1 Missing :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
I had one new comment. In addition, as you say that this fixes #101, should we maybe add a test that ensures that this is acutally the case? Otherwise if you didn't check that this actually works now, I would suggest not closing #101 yet.
I added a test to make sure you can add constructor methods for normals with PSDmatrix now.
Fixes #101 and #102
What has been done:
Normal(::Number, ::Number)
UvNormal{T}
forNormal{T,T,T} where {T<:Number}
UvNormal
AbstractNormal
methods more generic, so they work forUvNormal
as well.Some things to consider:
UvNormal
store the covariance as a real type? i.e. should it beUvNormal{T,V}
forNormal{T,T,V} where {T<:Number,V<:real(T)}
. In this case, the first point is irrelevant?Future work:
AffineMap
so that corresponding kernels can be constructed.NormalKernel
andDiracKernel
for the aforementioned affine maps.