Hello I was wondering if there is an easy way to meld these two (Python3 and sanderling).
I am a machine learning researcher and would like to test the capabilities of building a bot that can learn in game. However my skills are in python using tensorflow, hence my above question.
If anyone can point me in the right direction that’d be great, I searched on the forum but couldn’t find anything.
Hello and welcome!
Yes, integrating Python is rather unusual. It can be more complicated, depending on the data- and control flow you have in mind for your project.
For off-policy learning cases, you could read the screen contents from the files written by apps that integrate Sanderling. This way, you don’t have to learn about the Sanderling interface and can use the files (JSON) as a reference.
If you want to do on-policy stuff, one way would be using the CLI interface, spawning a process with the Sanderling executable file each time you do a reading. However, using a new OS process will cost you some milliseconds. If you want to save those milliseconds, then use the API of the Sanderling .NET assembly. In that case, you would search for calling into .NET Core 6 code from Python.
Where can we find documentation for the functions available in the API as well as proper installation?
@Malkor
For that kind of low-level interface to Sanderling, you would use the API shown at https://github.com/Arcitectus/Sanderling/blob/e046f5a409d45fa038d2720f4b354a9465a19249/implement/read-memory-64-bit/Program.cs
For documentation of the low-level API, you can ask here by use-case or look at how the API is used in the other parts of the program and how it relates to the user interfaces.
You can use the installer from https://dotnet.microsoft.com/en-us/download/dotnet/6.0