Library to allow calling Fortran code from Python. Requires gfortran>=8.0, Works with python >= 3.7

Installing locally:

python -m pip install .

or install via pypi

python -m pip install --upgrade --user gfort2py

For a full list of supportetd platforms see the support documentation.

gfort2py has three main aims:

1. Make it trivially easy to call Fortran code from Python
2. Minimise the number of changes needed in the Fortran code to make this work.
3. Support as many Fortran features as possible.

We achieve this by tightly coupling the code to the gfortran compiler, by doing so we can easily embed assumptions about how advanced Fortran features work which makes development easier and minimises the number of changes needed on the Fortran side.

gfort2py use the gfortran mod files to translate your Fortran code's ABI to Python-compatible types using Python's ctype library. By using the mod file we can determine the call signature of all procedures, components of derived types, and the size and shapes of all module-level variables. As long as your code is inside a Fortran module, no other changes are needed to your Fortran code.

The downside to this approach is that we are tightly tied to gfortran's ABI, which means we can not support other non-gfortran compilers and we do not support all versions of gfortran. When gfortran next breaks its ABI (which happens rarely, the last break was gfortran 8) we will re-evaluate our supported gfortran versions.

There are two ways to load Fortran code into Python. Either fFort or compile. The recommended way is via fFort for interfacing with existing code, while compile is more suitable for wrapping short snippets of Fortran code

Your Fortran code must be inside a module and then compiled as a shared library.

On linux:

gfortran -fPIC -shared -c file.f90
gfortran -fPIC -shared -o libfile.so file.f90

On MacOS:

gfortran -dynamiclib -c file.f90
gfortran -dynamiclib -o libfile.dylib file.f90

On Windows:

gfortran -shared -c file.f90
gfortran -shared -o libfile.dll file.f90

If the shared library needs other shared libraries you may need to set the LD_LIBRARY_PATH environment variable, and it is also recommended to run chrpath on the shared libraries so you can access them from anywhere.

import gfort2py as gf

SHARED_LIB_NAME=f'./test_mod.{gf.lib_ext()}' # Handle whether on Linux, Mac, or Windows
MOD_FILE_NAME='tester.mod'

x=gf.fFort(SHARED_LIB_NAME,MOD_FILE_NAME) 

NOTE: The mod data is cached to speed up re-reading the data. To control this pass cache_folder to fFort. A value of False disables caching, a string sets the folder location, while leaving the argument as None defaults to platformdirs user_cache_dir

import gfort2py as gf


fstr = """
            integer function myfunc(x,y)
                integer :: x,y
                myfunc = x+y
            end function myfunc
"""

x  = gf.compile(string=fstr)

The Fortran code can also be in a file in which case:

import gfort2py as gf

x  = gf.compile(file='my_fortran_file.f90')

In either casee the code will be compilied into a Fortran module and then into a shared library. Any Fortran code is valid as long as it can be inserted into a Fortran Module (Its optional whether you need to wrap things in module/end module, if you do not then that is done automatically for you).

Additional options available for compile:

• FC: str Path to gfortran compilier
• FFLAGS: str Additional Fortran compile options. This defaults to -O2.
• LDLIBS: str Any additional libraries needed to be linked in (-l)
• LDFLAGS: str Locations of addtional libraries (-L)
• ouput: str Location to save intermediate files to. Defaults to None which saves files in a temporary location. Otherwise save to the location specified.

NOTE: The interface to compile is currently considered unstable and may change.

x now contains all variables, parameters and procedures from the module (tab completable), and is independant on how the Fortran code was loaded.

y = x.func_name(a,b,c)

Will call the Fortran function with variables a,b,c and returns the result in y.

y will be named tuple which contains (result, args). Where result is a python object for the return value (0 if a subroutine) and where args is a dict containing all arguments passed to the procedure (both those with intent (in) which will be unchanged and intent(inout/out) which may have changed).

x.some_var = 1

Sets a module variable to 1, will attempt to coerce it to the Fortran type

x.some_var

Will return a Python object

Optional arguments that are not present should be passed as a Python None.

Arrays should be passed as a NumPy array of the correct size and shape.

Remember that Fortran by default has 1-based array numbering while Numpy is 0-based.

If a procedure expects an unallocated array, then pass None as the argument, otherwise pass an array of the correct shape.

Derived types can be set with a dict

x.my_dt={'x':1,'y':'abc'}

y=x.my_dt
y['x']
y['y']

If the derived type contains another derived type then you can set a dict in a dict

x.my_dt={'x':1,'y':{'a':1}}

When setting the components of a derived type you do not need to specify all of them at the same time.

If you have an array of derived types

type(my_type), dimension(5) :: my_dt
type(my_type), dimension(5,5) :: my_dt2

Elements can be accessed via an index:

x.my_dt[0]['x']
x.my_dt2[0,0]['x']

You can only access one component at a time (i.e no striding [:]). Allocatable derived types are not yet supported.

Derived types that are dummy arguments to a procedure are returned as a fDT type. This is a dict-like object where the components can only be accessed via the item interface ['x'] and not as attributes .x. This was done so that we do not have a name collision between Python functions (keys, items etc) and any Fortran-derived type components.

You can pass a fDT as an argument to a procedure.

Quad precision (REAL128) variables are not natively supported by Python thus we need a different way to handle them. For now that is the pyQuadp library which can be installed from PyPi with:

python -m pip install pyquadp

or from a git checkout:

python -m pip install .[qaud]

For more details see pyQuadp's documentation, but briefly you can create a quad precision variable from an int, float, or string. On return you will receive a qfloat type. This qfloat type acts like a Python Number, so you can do things like add, multiply, subtract etc this Number with other Numbers (including non-qfloat types).

We currently only support scalar Quad's and scalar complex Quad's. Arrays of quad precision values is planned but not yet supported. Quad values can also not be returned as a function result (this is a limitation in ctypes which we have no control over). Thus a quad precision value can only occur in:

• Module variables
• Parameters
• Procedure arguments

pyQuadp is currently an optional requirement, you must manually install it, it does not get auto-installed when gfort2py is installed. If you try to access a quad precision variable without pyQuadp you should get a TypeError.

To pass a Fortran function as a callback argument to another function then pass the function directly:

y = x.callback_function(1)

y = x.another_function(x.callback_function)

Currently only Fortran functions can be passed. No checking is done to ensure that the callback function has the correct signature to be a callback to the second function.

The callback and also be created in Python at runtime (but must be valid Fortran):

fstr = """
        integer function callback(x)
            integer :: x
            write(*,*) x
            callback = 3*x
        end function callback

        """

f = gf.compile(fstr)


y = x.another_function(f.callback)

python -m pip install .[test]
pytest -v

To run unit tests

• Scalars
• Parameters
• Characters
• Explicit size arrays
• Complex numbers (Scalar and parameters)
• Getting a pointer
• Getting the value of a pointer
• Allocatable arrays
• Derived types
• Nested derived types
• Explicit Arrays of derived types
• Allocatable Arrays of derived types
• Procedure pointers inside derived types
• Derived types with dimension(:) array components (pointer, allocatable, target)
• Allocatable strings (partial)
• Explicit Arrays of strings
• Allocatable arrays of strings
• Classes
• Abstract interfaces
• Common blocks (partial)
• Equivalences
• Namelists
• Quad precision variables
• function overloading

• Basic calling (no arguments)
• Argument passing (scalars)
• Argument passing (strings)
• Argument passing (explicit arrays)
• Argument passing (assumed size arrays)
• Argument passing (assumed shape arrays)
• Argument passing (allocatable arrays)
• Argument passing (derived types)
• Argument intents (in, out, inout and none)
• Passing characters of fixed size (len=10 or len=* etc)
• Functions that return a character as their result
• Allocatable strings (Only for things that do not get altered inside the procedure)
• Explicit arrays of strings
• Allocatable arrays of strings
• Pointer arguments
• Optional arguments
• Value arguments
• Keyword arguments
• Generic/Elemental functions
• Functions as an argument
• Unary operations (arguments that involve an expression to evaluate like dimension(n+1) or dimension((2*n)+1))
• Functions returning an explicit array as their result

There's no direct way to access the common block elements, but if you declare the common block as a module variable you may access the elements by their name:

module my_mod
    implicit none
    
    integer :: a,b,c
    common /comm1/ a,b,c
    

Elements in the common block can thus be accessed as:

x.a
x.b
x.c

For those wanting to explore the module file format, there is a routine mod_info available from the top-level gfort2py module:

module = gf.mod_info('file.mod')

That will parse the mod file and convert it into an intermediate format inside module.

Variables or procedures can be looked up via the item interface (I also recommend using pprint for easier viewing):

from pprint import pprint

pprint(module['a_variable'])

Accessing the list of all available components can be had via module.keys().

You can also do:

module = gf.mod_info('file.mod',json=True)
module['a_variable']

Then when you access each component the return value will be JSON-formatted. Note you can currently only access each component as JSON not the whole module file as JSON at the moment.

Bug reports are of course welcome and PR's should target the main branch.

For those wanting to get more involved, adding Fortran examples to the test suite of currently untested or unsupported features would be helpful. Bonus points if you also provide a Python test case (that can be marked @pytest.mark.skip if it does not work) that demonstrates the proposed interface to the new Fortran feature. Features with test cases will move higher in the order of things I add to the code.

See how to write a test case for details on how to write test cases.

For those wanting to go further and add the new feature themselves open a bug report and we can chat about what needs doing.

Debugging instructions are here