Now that we can use Dash.NET in F# scripts (plotly/Dash.NET#24), It seems like a perfect candidate for setting up a basic backend for rendering Plotly.NET charts during interactive scripting. The current workflow creates temporary files that get opened in the browser, and this can lead easily to this kind of Tab-apocalypse:

I envision this like the following:

The backend

Just a basic api that accepts plotly.js figure json and passes it to a Dash.NET Graph component on the frontend.

This could be either a new package e.g. Plotly.NET.FigureServer or added to Plotly.NET.Interactive

The workflow

Option 1: register in some kind of global config

startFigureServer(Port=XYZ, OpenBrowser=true) //custom port, open browser

PlotlyGlobalConfig.FigureServer.Port // access port on some global object

Option 2: use a record / object for the server

let figureServer = FigureServer.init(Port=XYZ, OpenBrowser=true) //custom port, open browser

figureServer.start()

figureServer.Port// access port on some local instance

sending figures to that backend

Option1: via optional setting in Chart.Show.

This is easier to do when adding backend option 1.

SomeChart
|> Chart.Show(useFigureServer = true) // read necessary info from global config

SomeChart
|> Chart.Show(FigureServer = figureServer) // when using the local object approach (option 2 above)

Why would this be super awesome actually?

Well, this enables us for example to create an extension for VSCode that rneders charts from scripts directly into a plot window. Wouldn't that be amazing?