Bugfixes & documentation

docs/src/rcall.md

@@ -1,51 +1,8 @@
 # Integration with R
 
-If you don't want to use Julia for your data analysis, but you'd like to try out `Mice.jl`, you can use it from R! The R package [`JuliaCall`](https://non-contradiction.github.io/JuliaCall/index.html) [li_juliacall_2019](@cite) and the Julia package [`RCall.jl`](https://juliainterop.github.io/RCall.jl/stable/) allow you to call Julia functions from R and vice-versa.
+If you don't want to use Julia for your data analysis, but you'd like to try out `Mice.jl`, good news - you can call R functions from Julia! The Julia package [`RCall.jl`](https://juliainterop.github.io/RCall.jl/stable/) allows you to do this.
 
-!!! note
-    Using `Mice.jl` from R requires Julia v1.9 or higher to be installed on your computer. If you don't have it, you can install it from [here](https://julialang.org/downloads/). You also need to install the Julia package `RCall.jl` in addition to `Mice.jl` by entering in Julia:
-    ```
-    ] add Mice, RCall
-    ```
-
-Using the combined power of these two packages, you can use `Mice.jl` without leaving R. Here's how:
-
-```r
-> library(JuliaCall)
-
-> julia_setup()
-# Julia version 1.9.4 at location /home/~USERNAME~/julia-1.9.4/bin will be used.
-# Loading setup script for JuliaCall...
-# Finish loading setup script for JuliaCall.
-
-> julia_library("Mice")
-
-> julia_library("RCall")
-
-> julia_library("GLM")
-
-> julia_library("Random")
-
-> data <- read.csv("test/data/cirrhosis.csv")
-
-> data$Stage <- as.factor(data$Stage)
-
-> julia_assign("data", data)
-
-> julia_command("myPredictorMatrix = makePredictorMatrix(data);")
-
-> julia_command('myPredictorMatrix[:, ["ID", "N_Days"]] .= 0;')
-
-> julia_command("Random.seed!(1234);") # Set random seed for reproducibility
-
-> julia_command("imputedData = mice(data, predictorMatrix = myPredictorMatrix);")
-
-> julia_command("analyses = with(imputedData, data -> lm(@formula(N_Days ~ Drug + Age + Stage + Bilirubin), data));")
-
-> julia_command("results = pool(analyses);")
-```
-
-You can also start in Julia, perform your data wrangling in R, use `Mice.jl` and then send your Mids object back to R and continue analysing it there. For example:
+You can start in Julia, perform your data wrangling in R, use `Mice.jl` and then send your Mids object back to R and continue analysing it there. For example:
 
 ```r
 julia> using Mice, Random, RCall

src/makeFunctions.jl

@@ -111,9 +111,7 @@ function initialiseWorkingData(
 
             # Convert to non-missing type
             if workingData[var][1] isa CategoricalArray
-                for j in eachindex(workingData[var])
-                    workingData[var][j] = convert(CategoricalArray{nonmissingtype(eltype(workingData[var][1]))}, workingData[var][j])
-                end
+                workingData[var] = [CategoricalArray{nonmissingtype(eltype(workingData[var][1]))}(workingData[var][j]) for j in 1:m]
             else
                 workingData[var] = convert(Vector{Vector{nonmissingtype(eltype(workingData[var][1]))}}, workingData[var])
             end

src/sampler.jl

@@ -2,7 +2,7 @@
 function sampler!(
     workingData::AxisVector{Vector},
     workingDataPacified::AxisVector,
-    workingDataLevels::AxisVector{Vector},
+    workingDataLevels::AxisVector,
     meanTraces::Vector{Matrix{Float64}},
     varTraces::Vector{Matrix{Float64}},
     imputeWhere::AxisVector{Vector{Bool}},