Closed JorySchossau closed 2 years ago
To deter more #denvercoder9 situations, this is for posterity:
let np = pyImport("numpy")
let ctypes = pyImport("ctypes")
var mem = cast[ptr UncheckedArray[cfloat]](alloc0 4*sizeof(cfloat))
var thepointer = ctypes.cast(cast[uint](mem),ctypes.POINTER(nimpy.getAttr(ctypes, "c_float")))
var numpyarray = np.ctypeslib.as_array(thepointer, (4,))
I'm trying to avoid a data copy, and use
numpy.as_array
to create a numpy array from data I already have in the right format, but I don't know how to package a pointer for nimpy to pass along. Is there some fancy way to use PyCapsules for this? I see a previous user has asked, but closed their own issue without any information.