[dsyme]

This adds the F# getting-started samples.

I've only compiled so far, against my on-machine preview of ML.NET

Things to do:

• Depends on ML.NET 0.4.0 with Smoke test for F# machinelearning#600 and Allow use of property-based row classes in ML.NET machinelearning#616
• Make sure the samples run
• Consider using XPlot as it is more normal for F#
• Consider making available as F# scripts rather than just compiled code (these would run on .NET Framework). Adding a pivot "compiled" and "scripting" may make senses, I'm not sure.
• Some "force modules to load" additions may be needed, I'm not sure
• Check the addition of the "csharp" pivot hasn't screwed anything up
• remove stuff that shouldn't be here like the change to nuget.config

[dsyme]

nb. this won't pass CI until the updated ML.NET is available

[isaacabraham]

Linking to #7

[isaacabraham]

@dsyme haven't looked at this yet - what are the folder structures like in comparison to one another? I opted for an fsharp set of folders which had paket etc. in them - they were all self-contained. Has some pros and cons to that approach.

[dsyme]

@dsyme haven't looked at this yet - what are the folder structures like in comparison to one another? I opted for an fsharp set of folders which had paket etc. in them - they were all self-contained. Has some pros and cons to that approach.

csharp and fsharp are parallel. No use of paket here, the projects are compiled for now. I do need to add scripting and there will be a common load script, I'm not sure we'll use paket though, need to think about it

[dsyme]

Closing in favour of #35 and #36