Enable individual indexing for merged lookups

[asinghvi17]

ds = rand(X(1:10), Y(1:10), Ti(1:10))
m = mergedims(ds, (X, Y) => :space)

m[X(At(1))]

previously threw an error, with this PR it just works

What this does is that it introduces a new AbstractMergedLookup supertype that requires that the lookup have dims. this way we can find all lookups that might be secretly multidimensional (currently just merged lookup and geometry lookup) and pass all matching dims into them directly. That gets us back a vector index like thing that we can replace the placeholder Colon with that we receive from the split_dims.

This enables geometry lookups in Rasters to work like this as well, and also selecting by extents, crop, etc.

This isn't actually breaking in any way so can be merged for a non breaking release, I think.

Needs tests and benchmarks to check that this does not impact the getindex time, ie it should just compile away

[rafaqz]

Had a thought: what happens if e.g. X and and Y are in the regular dimensions, and also in the merged dimensions

[asinghvi17]

In the current implementation the merged dimensions are ignored, because X and Y don't make it out of otherdims

We could in theory duplicate those. But it seems fragile and pretty ambiguous if your array is already fully specified.

[rafaqz]

Yeah, Its probably best. I just wondered about the behaviour where there are two merged dims and X/Y working on both of them, but not if there is also regular X/Y. That may be slightly confusing.

[codecov]

Codecov Report

Attention: Patch coverage is 33.33333% with 18 lines in your changes missing coverage. Please review.

Project coverage is 75.81%. Comparing base (20b42df) to head (4c3d727).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/Dimensions/dimension.jl	0.00%	17 Missing ⚠️
src/Dimensions/indexing.jl	88.88%	1 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (20b42df) and HEAD (4c3d727). Click for more details.

HEAD has 5 uploads less than BASE

Flag	BASE (20b42df)	HEAD (4c3d727)
	6	1

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #980      +/-   ##
==========================================
- Coverage   83.44%   75.81%   -7.64%     
==========================================
  Files          54       54              
  Lines        5154     5168      +14     
==========================================
- Hits         4301     3918     -383     
- Misses        853     1250     +397     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[rafaqz]

I think the MergedLookup docstring need indexing examples so this isn't hidden

[asinghvi17]

ok, all relevant tests are passing.

no clue why plots is failing though..are you using some strange backend?

[rafaqz]

It's just GR. We can ignore it for now

[asinghvi17]

julia> using DimensionalData
imp
julia> import DimensionalData as DD

julia> using DimensionalData.Dimensions: _experimental_extent

julia> using Extents

julia> d = rand(X(1:10), Y(1:10), Z(1:10), Ti(1:10), Dim{:something}(1:2))
┌ 10×10×10×10×2 DimArray{Float64, 5} ┐
├────────────────────────────────────┴────────────────────────────────────────────────────────────────────────────────────────────────── dims ┐
  ↓ X         Sampled{Int64} 1:10 ForwardOrdered Regular Points,
  → Y         Sampled{Int64} 1:10 ForwardOrdered Regular Points,
  ↗ Z         Sampled{Int64} 1:10 ForwardOrdered Regular Points,
  ⬔ Ti        Sampled{Int64} 1:10 ForwardOrdered Regular Points,
  ◩ something Sampled{Int64} 1:2 ForwardOrdered Regular Points
└─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘
[:, :, 1, 1, 1]
 ⋮      ⋱  

julia> dm = mergedims(d, (X, Y) => :space, (Z, Ti) => :valu)
┌ 2×100×100 DimArray{Float64, 3} ┐
├────────────────────────────────┴────────────────────────────────────────────────────────────────────────────────────────────────────── dims ┐
  ↓ something Sampled{Int64} 1:2 ForwardOrdered Regular Points,
  → space     MergedLookup{Tuple{Int64, Int64}} [(1, 1), (2, 1), …, (9, 10), (10, 10)] d↓ X, → Y,
  ↗ dvalu      MergedLookup{Tuple{Int64, Int64}} [(1, 1), (2, 1), …, (9, 10), (10, 10)] ↓ Z, → Ti
└─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘
[:, :, 1]
 ⋮      ⋱  

julia> ddm = dims(dm)
(↓ something Sampled{Int64} 1:2 ForwardOrdered Regular Points,
→ space     MergedLookup{Tuple{Int64, Int64}} [(1, 1), (2, 1), …, (9, 10), (10, 10)] (↓ X, → Y),
↗ valu      MergedLookup{Tuple{Int64, Int64}} [(1, 1), (2, 1), …, (9, 10), (10, 10)] (↓ Z, → Ti))