wrap libtorch rank and shape routines

[TomMelt]

the tensor derived type now supports two methods:

• get_rank
• get_shape

These methods return the shape and rank of the tensor

[jatkinson1000]

Thanks @TomMelt

This all reads correctly to me.
I have looked at the libtorch API, but will clarify that I have not run it myself.
I presume you have tested these routines yourself locally and they perform as expected?

I don't think we can insist on tests here (as they are lacking elsewhere in the code) but as we bring more of these "utility" routines across we should look at tackling #4 with more priority.
I suppose there will be some integration testing from the autograd example.

[jwallwork23]

Looks good, thanks @TomMelt.

[TomMelt]

all reads correctly to me.

Thanks @jatkinson1000 . Spot on.

• I plan to add unit tests in the near future with pFUnit.
• I did check locally by manually modifying the integration test but I will put a proper test in for the autograd work