Error when using callback function with f2py

Question

I'm trying to pass a python function to a f2py generated function. I can not figure out how to do this when the python function's input and output is an array. (I'm new to f2py!)

Here is my fortran code (num_Dg.f):

     subroutine num_Dg(func,U,jacob,n)
        double precision, intent(in) :: U(n)
        double precision, intent(out) :: jacob(n,n)
        double precision Up(n), Um(n)
        double precision g_valp(n),g_valm(n)
        double precision :: rel_eps = 2.e-5
        double precision eps
        integer i, j
        external func

        eps = minval(U)*rel_eps

        do j=1,n
          do i=1,n
            if (i .eq. j) then
            Up(i) = U(i)+(eps)
            Um(i) = U(i)-(eps)
            else
            Up(i) = 0
            Um(i) = 0
            endif
          enddo
          call func(Up,g_valp,n)
          call func(Um,g_valm,n)
          do i=1,n
            jacob(i,j) = 0.0
            jacob(i,j) = (g_valp(i)-g_valm(i))/(2*eps)
          enddo
        enddo
      end subroutine

I use f2py to compile: f2py -c num_Dg.f -m num_Dg . It compiles without errors. But when import the function into python I get this:

>>> from num_Dg import num_dg
>>> print(num_dg.__doc__)
jacob = num_dg(func,u,[n,func_extra_args])

Wrapper for ``num_dg``.

Parameters
----------
func : call-back function
u : input rank-1 array('d') with bounds (n)

Other Parameters
----------------
func_extra_args : input tuple, optional
    Default: ()
n : input int, optional
    Default: len(u)

Returns
-------
jacob : rank-2 array('d') with bounds (n,n)

Notes
-----
Call-back functions::

  def func(up,gvalp,[n]): return 
  Required arguments:
    up : input rank-1 array('d') with bounds (n)
    gvalp : input rank-1 array('d') with bounds (n)
  Optional arguments:
    n : input int, optional
    Default: len(up)

There is a problem with the callback function func . The input should be array U , and the output should be gval . Instead, both U and gval are inputs, with no output.

I'm assuming this is why the following python script returns garbage:

import num_Dg

def func(x):
    return [x[0]**2,x[1],x[2]]

val = [0.2,0.1,0.1]

num_Dg.num_dg(func,val)

How should I change my fortran subroutine so that the callback function works properly? Any advice is appreciated! Thanks!

Answer 1

I ended up figuring it out. I think the reason why it wasn't working was because I was missing the line cf2py intent(out), depend(n) :: g_valp . The script can be compiled in the normal way with f2py: f2py -c num_Dg.f -m num_Dg

      subroutine num_Dg(U,jacob,n)
cf2py   intent(callback) func
        external func
        integer n
        double precision :: U(n)
        double precision :: jacob(n,n)
        double precision Up(n), Um(n)
        double precision g_valp(n),g_valm(n)
        double precision :: rel_eps = 2.e-5
        double precision eps
        integer i, j

cf2py   intent(in,copy), depend(n) :: U
cf2py   intent(hide) :: n
cf2py   intent(out), depend(n) :: jacob
cf2py   intent(out), depend(n) :: g_valp

        eps = minval(U)*rel_eps
        do j=1,n
          do i=1,n
            if (i .eq. j) then
            Up(i) = U(i)+(eps)
            Um(i) = U(i)-(eps)
            else
            Up(i) = 0
            Um(i) = 0
            endif
          enddo

          call func(n,Up,g_valp)
          call func(n,Um,g_valm)

          do i=1,n
            jacob(i,j) = (g_valp(i)-g_valm(i))/(2*eps)

          enddo

        enddo
      end subroutine