Plots fail when used in Pluto notebooks

[cncastillo]

Apparently the plots need to be fixed in Pluto. Not sure if this is a problem with PlotlyJS and Pluto or if we can fix it somehow.

PlutoNotebook.zip

[cncastillo]

@beorostica I moved this to high priority, as I believe it is fundamental to be compatible with the language tools.

This can be done creating a new struct that has the Plotly plot inside and define what the show function does.

Like PlutoPlotlyPlot in this file and this HTML example. Or even simpler:

struct PlutoPlotlyPlot
    plot::PlotlyJS.SyncPlot
end

This should be a package extension so it will only be defined if you imported Pluto.

using KomaMRIPlots
p = plot_seq(...) #p::PlotlyJS.SyncPlot

and

using KomaMRIPlots, Pluto
p = plot_seq(...) #p::KomaMRIPlots.PlutoPlotlyPlot

The problem I see with this approach is that we could loose the subplotting functionality p_combined = [p1 p2]. To avoid this we can define what the followting operations do:

https://github.com/JuliaPlots/PlotlyJS.jl/blob/5562c6ca71054cac9c13fcb248e698e3c67b172c/src/util.jl#L7-L10

Base.hcat(sps::SyncPlot...) = SyncPlot(hcat(Plot[sp.plot for sp in sps]...))
Base.vcat(sps::SyncPlot...) = SyncPlot(vcat(Plot[sp.plot for sp in sps]...))
Base.hvcat(rows::Tuple{Vararg{Int}}, sps::SyncPlot...) =
    SyncPlot(hvcat(rows, Plot[sp.plot for sp in sps]...))

So

#Example with hcat
import Base.hcat
Base.hcat(ppps::PlutoPlotlyPlot...) = PlutoPlotlyPlot(hcat([ppp.plot for ppp in ppps]...))

[cncastillo]

Just realized that is even easier! someone already did the job for us. PlutoPlotly.jl. We just neet to replace the plot function in the plotting functions to use PlutoPlotly.plot.

[cncastillo]

This is an example that seems to work, just need to do this automatically when in Pluto.

notebook.zip

[cncastillo]

I think I fixed it in

• KomaMRIPlots Pluto compatibility #195

Example:

notebook (1).zip