Making AxisArrays compatible with new core traits.#33
Conversation
|
|
||
| ImageCore.HasDimNames(::Type{A}) where {A<:AxisArray} = HasDimNames{true}() | ||
| ImageCore.namedaxes(a::AxisArray) = NamedTuple{axisnames(a)}(axisvalues(a)) | ||
| Base.names(a::AxisArray) = axisnames(a) |
There was a problem hiding this comment.
This is kind of iffy but it is necessary to really fulfill the intended meaning of HasDimNames.
There was a problem hiding this comment.
Remind me why the name of this function needs to be names? (Link to other package, for example?) The reason I ask:
julia> t = (x = 1, y = 2)
(x = 1, y = 2)
julia> typeof(t)
NamedTuple{(:x, :y),Tuple{Int64,Int64}}
julia> names(t)
ERROR: MethodError: no method matching names(::NamedTuple{(:x, :y),Tuple{Int64,Int64}})
Closest candidates are:
names(::Module; all, imported) at reflection.jl:98
Stacktrace:
[1] top-level scope at REPL[3]:1struct, for example, uses fieldnames rather than just names. axisnames seems like a really good name for this.
There was a problem hiding this comment.
It's what NamedDims uses to access the axis names https://github.com/invenia/NamedDims.jl/blob/master/src/wrapper_array.jl#L77.
But hadn't even considered using a different name. axisnames would certainly be more descriptive.
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #33 +/- ##
==========================================
- Coverage 81.95% 81.61% -0.34%
==========================================
Files 1 1
Lines 133 136 +3
==========================================
+ Hits 109 111 +2
- Misses 24 25 +1 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
2 participants
The goal here is to make AxisArrays compatible with the new traits added to ImageCore. This adds 3 methods to AxisArrays.