System.TypeLoadException: Could not load type 'Python.Runtime.PyLong' from assembly 'Python.Runtime, Version=3.0.1.0, Culture=neutral, PublicKeyToken=5000fea6cba702dd'.

[dariusz-wozniak]

Using:

• Numpy.Bare Version="3.8.1.28"
• Numpy.Bare.Mono Version="3.8.1.28"
• PackageReference Version="3.0.1"

When initializing code in the Program.cs:

Runtime.PythonDLL = @"C:\Python311\python311.dll";
if (!PythonEngine.IsInitialized) PythonEngine.Initialize();
np.arange(1);

I get exception at the np.arange(1); line:

System.TypeLoadException: Could not load type 'Python.Runtime.PyLong' from assembly 'Python.Runtime, Version=3.0.1.0, Culture=neutral, PublicKeyToken=5000fea6cba702dd'.
   at Numpy.np.ToPython(Object obj)
   at Numpy.np.ToTuple(Array input)
   at Numpy.np.arange(Int32 stop, Int32 step, Dtype dtype)
   at Sennheiser.Features.Linearization.PythonInitialization.Initialize() in 

Any idea how to fix it?

[henon]

Does pylong work when you try directly with pythonnet v3.0.1?

[dariusz-wozniak]

Does pylong work when you try directly with pythonnet v3.0.1?

When I remove all Numpy.Bare.* packages from the solution then:

1. The code works if I replace np.arrange(1) by:

       using (Py.GIL())
        {
            dynamic np = Py.Import("numpy");
            dynamic sin = np.sin;
            var sin5 = sin(5);
        }

2. The code doesn't work if I replace np.arrange(1) by:

        using (Py.GIL())
        {
            dynamic np = Py.Import("numpy");
            dynamic arrange = np.arrange;
            var a1 = arrange(1);
        }

...Then I have an exception on dynamic arrange = np.arrange:

System.Collections.Generic.KeyNotFoundException: The module has no attribute 'arrange'
   at Python.Runtime.PyModule.Get(String name)
   at Python.Runtime.PyModule.TryGetMember(GetMemberBinder binder, Object& result)
   at CallSite.Target(Closure , CallSite , Object , Int32 , Int32 , Double , Object )
   at System.Dynamic.UpdateDelegates.UpdateAndExecute5[T0,T1,T2,T3,T4,TRet](CallSite site, T0 arg0, T1 arg1, T2 arg2, T3 arg3, T4 arg4)

Also tried with the dynamic arrange = np.arrange(1); but got the same exception.

[henon]

arange is with one r

[henon]

I already know what is your problem, I see it now above: you are using Numpy.Bare vor python3.8 with latest python3.11, of course that won't work

[dariusz-wozniak]

True. It solved the issue, thank you 🙏