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

This package needs maintenance #17

Closed
rikhuijzer opened this issue Aug 9, 2021 · 8 comments
Closed

This package needs maintenance #17

rikhuijzer opened this issue Aug 9, 2021 · 8 comments

Comments

@rikhuijzer
Copy link
Collaborator

Hi @harryscholes,

this package looks really great but it's a real struggle to get it to work. I also get some internal errors because the eval in

for (T, f) = [(:CohenD, cohend), (:GlassΔ, glassΔ), (:HedgeG, hedgeg)]
@eval begin
# Normal CI
function $T(
xs::AbstractVector{T},
ys::AbstractVector{T};
quantile::Float64=0.95,
) where T<:Real
es = $f(xs, ys)
ci = ConfidenceInterval(xs, ys, es; quantile=quantile)
$T(es, ci)
end
# Bootstrap CI
function $T(
xs::AbstractVector{T},
ys::AbstractVector{T},
bootstrap::Integer;
quantile::Float64=0.95,
) where T<:Real
es = $f(xs, ys)
ci = BootstrapConfidenceInterval($f, xs, ys, bootstrap; quantile=quantile)
$T(es, ci)
end
end
end
doesn't seem to be evaluated.

Hereby a friendly request to pickup maintance again or give me (or someone else that you like) administration access to this repo. For me, maintaining the repo wouldn't be so much work. I'll leave you as the first author. Or, I could fork your repo and move the EffectSizes.jl name to my fork if you agree 🙂

@harryscholes
Copy link
Owner

Hi, thanks for your issue! Why do you think the @eval block isn't being evaluated? On a fresh installation of the package, I see those constructors that are defined by metaprogramming:

julia> using EffectSizes

julia> methods(CohenD)
# 3 methods for type constructor:
[1] CohenD(xs::AbstractVector{T}, ys::AbstractVector{T}; quantile) where T<:Real in EffectSizes at /Users/harry/.julia/packages/EffectSizes/Qfr3N/src/effect_size.jl:217
[2] CohenD(xs::AbstractVector{T}, ys::AbstractVector{T}, bootstrap::Integer; quantile) where T<:Real in EffectSizes at /Users/harry/.julia/packages/EffectSizes/Qfr3N/src/effect_size.jl:228
[3] CohenD(d::T, ci::CI) where {T<:Real, CI<:AbstractConfidenceInterval{T}} in EffectSizes at /Users/harry/.julia/packages/EffectSizes/Qfr3N/src/effect_size.jl:106

More than happy for you to take over maintenance! I've moved on from statistics, so don't have any time to dedicate to this package anymore.

@rikhuijzer
Copy link
Collaborator Author

I think that I was wrong about it. I got:

ERROR: MethodError: no method matching EffectSizes.ConfidenceInterval(::Vector{Int64}, ::Vector{Int64}, ::Float64; quantile=0.95)
Closest candidates are:
  EffectSizes.ConfidenceInterval(::AbstractVector{T}, ::AbstractVector{T}, ::T; quantile) where T<:Real at /home/rik/.julia/packages/EffectSizes/Qfr3N/src/confidence_interval.jl:76
  EffectSizes.ConfidenceInterval(::T, ::T, ::Float64) where T<:Real at /home/rik/.julia/packages/EffectSizes/Qfr3N/src/confidence_interval.jl:69 got unsupported keyword argument "quantile"
Stacktrace:
 [1] EffectSizes.CohenD(xs::Vector{Int64}, ys::Vector{Int64}; quantile::Float64)
   @ EffectSizes ~/.julia/packages/EffectSizes/Qfr3N/src/effect_size.jl:223
 [2] EffectSizes.CohenD(xs::Vector{Int64}, ys::Vector{Int64})
   @ EffectSizes ~/.julia/packages/EffectSizes/Qfr3N/src/effect_size.jl:222
 [3] es(A::Vector{Int64}, B::Vector{Int64})
   @ Personality ~/git/personality-md/src/tables.jl:5

but cannot reproduce it currently. Indeed, I also get your output for methods. I'll look into it further.

More than happy for you to take over maintenance! I've moved on from statistics, so don't have any time to dedicate to this package anymore.

Awesome! Very nice, it would be awesome if people can keep using the work you've put into the package. Could you make me admin?

@harryscholes
Copy link
Owner

Ahh yes that's because of this method signature

function ConfidenceInterval(
xs::AbstractVector{T},
ys::AbstractVector{T},
es::T;
quantile::AbstractFloat,
) where T<:Real

It might be more helpful if it was:

function ConfidenceInterval(
    xs::AbstractVector{<:Real},
    ys::AbstractVector{<:Real},
    es::Real;
    quantile::AbstractFloat,
)

but I'll leave that up to you :)

I've added you to the repo

@rikhuijzer
Copy link
Collaborator Author

Thank you for your trust 😄!

@rikhuijzer
Copy link
Collaborator Author

Hmm sorry that I have to bother you one more time. What is the name of the environment variable that holds the private key? I think that you have set one in the Travis repository settings.

@harryscholes
Copy link
Owner

Do you mean DOCUMENTER_KEY?

@rikhuijzer
Copy link
Collaborator Author

I just solved the Documenter build by only using GITHUB_TOKEN. For CompatHelper, it would be nice if there is a SSH key with write permissions (https://juliaregistries.github.io/CompatHelper.jl/dev/#Installation,-step-2:-Set-up-the-SSH-deploy-key). Then, for updated packages, CompatHelper will automatically open a PR and run tests on the PR.

But, it will also work without 👍 So, if you can setup a token with the manual from the link above, then that would be great. Otherwise, I'll manage too

@harryscholes
Copy link
Owner

done

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

No branches or pull requests

2 participants