Consider the simple fortran function: <div class="highlight highlight-source-fortr

array function error about f90wrap HOT 4 OPEN

jameskermode commented on July 18, 2024

array function error

from f90wrap.

Comments (4)

gfogwill commented on July 18, 2024

Hello,
i'm trying to do something similar, but without success.

My code is:

subroutine test(a, b)

implicit none

INTEGER, PARAMETER           :: real_x = SELECTED_REAL_KIND(p=14, r=300)
INTEGER :: i

REAL(real_x), INTENT(inout)  :: a(10)               
REAL(real_x), INTENT(out)    :: b(10)

do i=1,10
    b(i)=a(i)*2
enddo

end subroutine test

When I call f90wrap, f90wrap_toplevel.f90 is generated:

subroutine f90wrap_test(a, b)
    implicit none
    external test
    
    real(4), intent(inout), dimension(10) :: a
    real(4), dimension(10), intent(inout) :: b
    call test(a, b)
end subroutine f90wrap_test

And my test.py file is:

from __future__ import print_function, absolute_import, division
import _test
import f90wrap.runtime
import logging

def test(a, b):
    """
    test(a, b)
    

    Defined at test.f90 lines 1-13

    Parameters
    ----------
    a : float array
    b : float array
    
    """
    _test.f90wrap_test(a=a, b=b)

So I cannot get the value of 'b' variable.

Is it possible to make this work?

from f90wrap.

jameskermode commented on July 18, 2024

This should work as you expect, I’ll take a look. Thanks for reporting!

from f90wrap.

jameskermode commented on July 18, 2024

I tried your example, I think it's working correctly - if you preallocate arrays from Python you can pass them in and out of Fortran and they are modified in place, e.g:

import test
import numpy as np

a = np.zeros(10, dtype=np.float32)
b = np.zeros(10, dtype=np.float32)
a[:] =  [i for i in range(len(a))]
test.test(a, b)
print(b)