Accept any iterable in scalar stats #472
Conversation
Also improve a few docstrings.
| count = 1 | ||
| value, state = y | ||
| y = iterate(x, state) | ||
| # Use Welford algorithm as seen in (among other places) |
There was a problem hiding this comment.
It's too bad I needed to copy the code from var, but if we move this to Statistics it will be easy to share the code in a common internal method.
Codecov Report
@@ Coverage Diff @@
## master #472 +/- ##
==========================================
- Coverage 85.12% 83.76% -1.37%
==========================================
Files 21 21
Lines 2098 2156 +58
==========================================
+ Hits 1786 1806 +20
- Misses 312 350 +38
Continue to review full report at Codecov.
|
|
A solid improvement on the state of the world |
Make a copy when the element type is not concrete to ensure we choose an appropriate eltype.
|
Nice! Is this good to go? I think it's a really important improvement now that |
| function mean_and_var(A::RealArray, w::AbstractWeights; corrected::DepBool=nothing) | ||
| m = mean(A, w) | ||
| v = varm(A, w, m; corrected=depcheck(:mean_and_var, corrected)) | ||
| function mean_and_var(x::RealArray, w::AbstractWeights; corrected::DepBool=nothing) |
There was a problem hiding this comment.
Why do we still expect RealArray here? I thought this should also work with other arrays. E.g. Vector{Union{Missing, Float64}} as long as it does not contain missing (which will simply throw an error at run time if it does). Same questions for methods below.
There was a problem hiding this comment.
Yeah, there are many other things to fix in this file. I'd rather fix them separately in another PR.
There was a problem hiding this comment.
It is OK to leave it for later. I reacted because you explicitly named this PR "Accept any iterable ..." 😄 so I thought this is exactly the change you want to make in this PR.
There was a problem hiding this comment.
Yeah the title is a bit too ambitious.
|
OK to merge then? |
Also improve a few docstrings.
Part of #449. Closes #342. Cc: @oxinabox