Closed glennmoy closed 3 years ago
Merging #27 (b535fdf) into master (2f1ce6c) will decrease coverage by
0.02%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## master #27 +/- ##
==========================================
- Coverage 98.52% 98.50% -0.03%
==========================================
Files 1 1
Lines 68 67 -1
==========================================
- Hits 67 66 -1
Misses 1 1
Impacted Files | Coverage Δ | |
---|---|---|
src/ObservationDims.jl | 98.50% <100.00%> (-0.03%) |
:arrow_down: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 2f1ce6c...b535fdf. Read the comment docs.
@oxinabox do you have any opinion on the various options in the PR description? and if you think I should do the same for NamedDims while I'm at it?
I do not have strong ones. I think what is done here is acceptable. I do not think you should do the same for NamedDims. AxisArrays is a almost deprecated package, supporting it poorly via hacks feels ok.
I am not in any particular hurry to invert the dependencies. We are only using this package in one place AFAIK. (or is it used in FeatureTransforms.jl?) If it turns out to be a good idea then we will use it in more and then it might be worth inverting the dependencies. But that is for a later day
Related to https://github.com/invenia/ObservationDims.jl/issues/3
Draft status as I want to discuss the below first and then implement the same solution for NamedDims.
There are 3 ways we can do this: