Updates boxplot to allow for custom percentiles

[MattAmos]

Adds the ability for custom percentiles to be passed to boxplots (and grouped box plots). Retains current default behaviour of using quartiles when custom percentiles are omitted.

using Plots
using StatsPlots
using RDatasets

school = RDatasets.dataset("mlmRev", "Hsb82")

Default quartiles

@df school groupedboxplot(string.(:Sector), :MAch, group=string.(:Sx))

@df school groupedboxplot(string.(:Sector), :MAch, notch=true, group=string.(:Sx))

Custom Percentiles

@df school groupedboxplot(string.(:Sector), :MAch, group=string.(:Sx), percentiles=[0.05, 0.25, 0.50, 0.75, 0.95])

@df school groupedboxplot(string.(:Sector), :MAch, group=string.(:Sx), notch=true, percentiles=[0.05, 0.25, 0.50, 0.75, 0.95])

@df school groupedboxplot(string.(:Sector), :MAch, group=string.(:Sx), percentiles=[0.05, 0.50, 0.95])

@df school groupedboxplot(string.(:Sector), :MAch, group=string.(:Sx), notch=true, percentiles=[0.05, 0.50, 0.95])

[MattAmos]

Any feedback on this would be greatly appreciated.

[BeastyBlacksmith]

Can you add some tests? Otherwise looks good to me

[cristobaltapia]

@MattAmos thanks for this PR. I was wondering if it would be possible to also consider the placement of the whiskers as percentiles. As I understand this PR, it only allows to modify the extent of the box and add additional lines within it at the specified positions. Could we add an option like whiskerpercentile=[0.05, 0.95] that modifies the extension of the whiskers (currently controlled exclusively by IQR * range)?

[MattAmos]

Thanks @BeastyBlacksmith and @cristobaltapia. I'll work on adding tests and configurable whiskers.

[Ph0non]

Any updates on this?