Final code review

[adrhill]

Hi all, I've taken a quick look at your package and have some feedback on what is "good practice" in Julia:

• When writing a package, it's better practice to explicitly import the functions you want to use (e.g. using Interpolations: somefunction, anotherfunction) instead of importing the whole package:
  

  ParallelPlots/src/ParallelPlots.jl

  Lines 3 to 5 in 2c29c9c

  	using CairoMakie
  	using DataFrames
  	using Interpolations

• You currently manually normalize columns in your DataFrame:
  

  ParallelPlots/src/ParallelPlots.jl

  Lines 10 to 11 in 2c29c9c

  	data[!, col] = (data[!, col] .- minimum(data[!, col])) ./
  	(maximum(data[!, col]) - minimum(data[!, col]))

  Instead, you could use transformations from DataFrames.jl:
  https://dataframes.juliadata.org/stable/man/basics/#Basic-Usage-of-Transformation-Functions

• Julia has an isnothing function:

  ParallelPlots/src/ParallelPlots.jl

  Line 19 in 2c29c9c

  if data === nothing

• I don't think this belongs in the docstring:

  ParallelPlots/src/ParallelPlots.jl

  Line 37 in 2c29c9c

  - Julia version: 1.10.5

• The documentation is incomplete:
  

  ParallelPlots/src/ParallelPlots.jl

  Lines 46 to 51 in 2c29c9c

  	- `data::DataFrame`:
  	- `normalize::Bool`:
  	- `color_feature::String || nothing`: select, which axis/feature should be used for the coloring (e.g. 'weight') (default: last)
  	- `title::String`:
  	- `feature_labels::String`:
  	- `curve::Bool`:

• Since you only implement one Makie.plot! method, you probably don't need to specialize on the parametric type of ParallelPlot. function Makie.plot!(pp::ParallelPlot) should also work.
  

  ParallelPlots/src/ParallelPlots.jl

  Line 108 in 2c29c9c

  function Makie.plot!(pp::ParallelPlot{<:Tuple{<:DataFrame}})

• You might want to break down your Makie.plot! method into small subfunctions. This is generally a good pattern in Julia and can help with compiler optimizations and performance. A good candidate for this would be the loop drawing lines from L240 to L287.

• The types in your interpolate function might be too specific:
  

  ParallelPlots/src/ParallelPlots.jl

  Line 374 in 2c29c9c

  function interpolate(last_x::Float64, current_x::Float64, last_y::Float64, current_y::Float64, x::Float64)

  
  If you want to force that all arguments need to have the same type, you could write

  function interpolate(last_x::T, current_x::T, last_y::T, current_y::T, x::T) where {T<:Real}

• It looks like your documentation is not building correctly. If you need help, don't hesitate to contact me. It might be helpful to first build the documentation locally:

  • https://adrianhill.de/julia-ml-course/docs/#building_documentation_locally
  • https://adrianhill.de/julia-ml-course/docs/#looking_at_your_documentation

[yunseopark-kaist]

• Explicitly import the functions you want to use instead of importing the whole package:
• Use transformations from DataFrames.jl instead of normalizing columns manually
• Consider using isnothing function:
• Consider removing it from the docstring:
• Complete documentation
• Remove unnecessary explicit type definition for parameters of Makie.plot!.
• Break down Makie.plot! method into small subfunctions
• Less specific types in interpolate function
• Build documentation correctly

[yunseopark-kaist]

18d998f#diff-418a007f014bb040297d075f1c7a85d0c35ab4945baf1079a3d75940d1a7451dR3

Couldn't resolve dependency for namespace with Makie with an explicit import yet.
I also couldn't find anything to import from interpolations.

resolved both now:)

[leonhaufe]

Interpolation is obsolet, will be removed in next merge :)