NumPy: how to check leading dimension of non-owned array?

Question

In NumPy, one can have arrays which are just pointing towards data that was allocated in a different array in order to avoid deep copies. For example:

import numpy as np
a = np.ones((3,4))
b = a[:2]

Now b just contains a pointer to the array in a, as evidenced by the flags:

b.flags

  C_CONTIGUOUS : True
  F_CONTIGUOUS : False
  OWNDATA : False
  WRITEABLE : True
  ALIGNED : True
  WRITEBACKIFCOPY : False
  UPDATEIFCOPY : False

This is all fine as long as one is sticking to NumPy's operations, but now suppose that I have a function that accepts a 2-d array and passes it to some C function, which will index it like this: (i,j) = j + i*ncol (for row-major) or (i,j) = i + j*nrow (for column-major).

If I now try this:

c = a[:2,:2]

Now the leading dimension of the array will not match with the number of columns, so it cannot be indexed like that anymore.

How can I get the leading dimension of an array?

Examples:

import numpy as np
a = np.ascontiguousarray(np.ones((3,4)))
b = np.asfortranarray(a)
c = a[:2,:2] ### LDA is 4
d = b[:2,:2] ### LDA is 3