Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Enhance ranking code #589

Merged
merged 7 commits into from
Feb 1, 2021
Merged

Conversation

alyst
Copy link
Member

@alyst alyst commented Aug 9, 2020

Small cleanups in the ranking code. Added @inbounds annotations for speed.
Common code is moved into _rank() method, which allows to:

  • correctly support n-dim arrays (n>1). If generic array input was not intended, I can tighten signatures to only allow x::AbstractVector
  • sortperm() kwargs are just passed through, so, in addition to lt and rev, ranking also supports by (so it could be applied to custom types) and alg
  • no function-generating macros are required to support missing values

_rank() helper provides:
1) correct support for n-dim, n>1, input arrays
2) minimizes code duplication
3) passthrough of sortperm() args
4) macro-less support for missing values
Copy link
Member

@nalimilan nalimilan left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sorry for the very slow reponse...

src/ranking.jl Outdated
return rks
end


"""
ordinalrank(x; lt = isless, rev::Bool = false)
ordinalrank(x; sortkwargs...)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Repeating the list of supported arguments is more user-friendly IMO, and it doesn't take a lot of space.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I've expanded lt, by and rev, because they are more commonly used, and alg and order definitions are too long.
The docstring now explicitly refers to the sort(x) keywords in the description.

src/ranking.jl Outdated
Comment on lines 27 to 30
xv = Vector{T}(undef, length(inds))
@inbounds for (i, ind) in enumerate(inds)
xv[i] = x[ind]
end
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why don't we just do xv = view(x, inds)? If making a copy is faster, then xv = disallowmissing(view(x, inds)) as in the old code or xv = convert(AbstractArray{T}, x) is both shorter and much more efficient e.g. for CategoricalArray.

n = _check_randparams(rks, x, p)

if n > 0
@inbounds if n > 0
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

While you're at it, maybe replace these while loops with for i in 2:n and so on? They are hard to read currently.

src/ranking.jl Outdated
("1234" ranking) of an array. The `lt` keyword allows providing a custom "less
than" function; use `rev=true` to reverse the sorting order.
All items in `x` are given distinct, successive ranks based on their
("1234" ranking) of an array. Supports the same keyword arguments as `sort(x; sortkwargs...)`
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

sortkwargs isn't defined now. Same below.

Suggested change
("1234" ranking) of an array. Supports the same keyword arguments as `sort(x; sortkwargs...)`
("1234" ranking) of an array. Supports the same keyword arguments as the `sort`

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actually, in L48 I sort-of introduce the sortkwargs, because otherwise later in the docstring it would not be entirely clear that the keyword args are passed through to sort to define the element order. I think from the context (supports the same keyword args as ...) it should be clear that sortkwargs refers to the keyword arguments of ordinalranks(). But if you have a better suggestion, I'm eager to fix it.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

OK. I would just have said "in the sorted vector" in the next sentence.

As you prefer, but if you keep this then please replace sort(x) with just sort below.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, that sounds good. I also simplified a bit the language in the other ranking docstrings.

@alyst
Copy link
Member Author

alyst commented Oct 5, 2020

CI failed on nightly, but it's not ranking-related.

@nalimilan Thanks for the review! There's also #590 if you have time :)

src/ranking.jl Outdated
function _rank(f!, x::AbstractArray{>: Missing}, R::Type=Int; sortkwargs...)
inds = findall(!ismissing, vec(x))
isempty(inds) && return missings(R, size(x))
T = nonmissingtype(eltype(x))
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

just noticed that T is not used anymore, will remove this line

@nalimilan nalimilan closed this Feb 1, 2021
@nalimilan nalimilan reopened this Feb 1, 2021
@nalimilan nalimilan merged commit 4620545 into JuliaStats:master Feb 1, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants