I have recently started using f2py to call Fortran from Python. I have found this useful for two reasons: speeding up Python scripts by calling compiled Fortran code, and using Python as a unit testing framework for Fortran modules. Unfortunately, the documentation for f2py is rather sparse, and may not be completely up to date. In this note, I will hopefully prevent you from wasting a lot of time figuring out how to pass array arguments, and return array results.
Array arguments
Passing array arguments is a critical when working with numerical algorithms. F2py handles this well, but it is difficult to figure it out how to do it correctly. Here is a simple example with some Fortran 90 routines:
module test
contains
subroutine foo (a)
implicit none
integer, intent(in) :: a
print*, "Hello from Fortran!"
print*, "a=",a
end subroutine foo
function bar (len_a, a)
implicit none
integer, intent(in) :: len_a
real, dimension(len_a), intent(in) :: a
real, dimension(len_a) :: bar
!f2py depend(len_a) a, bar
integer :: i
real, dimension(len_a) :: b
do i=1,len_a
b(i) = 2.0*a(i)
end do
bar = b
end function bar
subroutine sub (a, len_a, a_out)
implicit none
real, dimension(len_a), intent(in) :: a
integer, intent(in) :: len_a
real, dimension(len_a), intent(out) :: a_out
integer :: i
do i=1,len_a
a_out(i) = 2.0*a(i)
end do
end subroutine sub
end module test
Function “foo” is quite straightforward. Function “bar” is a little more complex, because it accepts an array argument and returns an array. If you’re new to Fortran, it will seem strange to pass the length of the array along with the array, but you need that information to declare the output array. f2py needs to know that the input array a and the output array bar both depend on the argument len_a. The special comment line
!f2py depend(len_a) a, bar
is mandatory! It tells f2py that a depends on len_a. If you omit this comment or the corresponding one for, the function will not work correctly. You will get strange errors like
ValueError: failed to create intent(cache|hide)|optional array-- must have defined dimensions but got (0,)
Fortran subroutines
Now look at the subroutine called sub. On the Fortran side, it has three arguments: two inputs and an output. However, Python only has functions, and all non-array arguments are passed by value. How do you reconcile this? When the Fortran subroutine is called from Python, the intent(out) variables are returned as a function result, or a tuple of results if there are more than one. Compare the Fortran code above with the Python call below to see what I mean.
Building
This code can be compiled using the “fast and smart” method described in the docs:
f2py -c -m hello hello.f90
Here is the Python code that calls the Fortran routines:
#!/usr/bin/env python import hello from numpy import * a = arange(0.0, 10.0, 1.0) len_a = len(a) print "foo:" hello.test.foo(len_a) print "bar:" a_out = hello.test.bar(len_a, a) print a_out print "sub:" a_out = hello.test.sub(a, len_a) print a_out
You might also want to check out this rather complicated f2py example.
There is a typo in your f2py command. It should be `f2py -c -m hello hello.f90`. The argument that follows -m is the module name.
Pingback: Error compiling basic Fortran program using f2py
My previous comment isn’t correct, since the command as given *does* work. However, putting the module name immediately after -m seems a bit clearer to me, and that is what is suggested in the help given by `f2py -h`.
Anyway, thanks for your ongoing blog! There’s a lot of very useful stuff here.
Warren, thank you for the suggestion. Putting the module name after -m makes more sense.
This is a great example of how to use f2py. It really helped to get me started.
Thanks
It is useful to blend python and fortran. However, these days I encountered a problem with blending these two languages. The problem is: 1, I compile and execute a fortran code with three simple loops(code will be pasted below), that the program costs just 15-second; 2, in order to be used by python, I change the fortran code to subroutine, compile it by f2py, and execute it with python, but this subroutine cost nearly 50-second. Could you please give me some suggestion? Thank you very much!
grid_nst=0
do i=1,1401
do j=1,1201
glat=(j-1)*0.1
glon=(i-1)*0.1
do ll=1,5000
distance=sqrt((glat-lat(ll))**2+(glon-lon(ll))**2)
if ( distance<= 2.5 ) then
grid_nst(j,i)=grid_nst(j,i)+1
endif
enddo
enddo
enddo
Fortran code for python：subroutine test(nst,lat,lon), the others are same
Python：var=loop.test(lat,lon)
f2py compile: f2py -c tt.f90 -m loop –f90exec=gfortran
SJ, what is the run time if you call the Fortran subroutine from a Fortran program (without Python involved)? Does the Python/Fortran version take 50 seconds every time it runs, or only the first time?
Thank you for replying, CRAIG. The run time is nearly 50 seconds when I call Fortran subroutine compiled by f2py through Python , but is only 15 seconds when call the same subroutine through Fortran program(without Python involved). Unfortunately, the Python version takes 50s every time. I check the performance of the Python version, the result is below:
/———————–\
\———————–/
Overall time spent in …
(a) wrapped (Fortran/C) functions : 49705 msec
(b) f2py interface, 1 calls : 113 msec
(c) call-back (Python) functions : 0 msec
(d) f2py call-back interface, 0 calls : 0 msec
(e) wrapped (Fortran/C) functions (actual) : 49705 msec
Use -DF2PY_REPORT_ATEXIT_DISABLE to disable this message.
Exit status: 0
Thanks again~