Determine if __array_wrap__ on ndarray subclass is called from unary ufunc or binary ufunc

Question

Is there a way to tell if a ndarray subclass's __array_wrap__ is called with a unary function or a binary function? (another reference)

Answer 1

This is only a partial answer:

The arguments to the ufunc are passed in as a tuple to context. The form is:

(ufunc, ufunc_args, ufunc_domain)

You can check the length of ufunc_args to see if you got 1 argument or 2. As a side note, I have no idea what ufunc_domain is (in my tests, it always seems to be 0)...

import numpy as np
class Tester(np.ndarray):
    def __array_wrap__(self,output,context=None):
        print context[0].__name__,'is binary' if len(context[1]) > 1 else 'is unary'
        return np.ndarray.__array_wrap__(self,output,context)

a = np.zeros(10)
b = a.view(Tester)
print (type(b))

-b
np.sqrt(b)
b+b

I suppose this is how you can tell __array_wrap__ whether it is a binary or unary ufunc. Unfortunately, when I asked the question in the beginning, I was hoping to know if this ufunc call was the result of a unary operator. I didn't think of things like np.abs and np.sqrt as unary functions.