Wrapping C with Python

Collapse
This topic is closed.
X
X
 
  • Time
  • Show
Clear All
new posts
  • Anish Chapagain

    Wrapping C with Python

    Hi!!
    I tried wrapping a simple C code suing SWIG to Python, but am having
    problem,
    my .c file is,

    Step 1:
    example.c
    --------------
    double val=3.0;
    int fact(int n)
    {
    if(n<=1)
    return 1;
    else
    return n*fact(n-1);
    }

    int mod(int x, int y)
    {
    return (x%y);
    }

    Step 2:
    I then created interface file as.
    example.i
    --------------
    %module example
    %{
    %}
    extern int fact(int n);
    extern int mod(int x,int y);

    Step 3:
    I compiled the .i file with SWIG,
    c:\python25\swi gswig -
    python example.i
    It creates example.py and example_wrap.c

    Step 4:
    I compiled example.c and example_wrap.c individually
    c:\python25\min gw\bin gcc -c example.c (it create
    example.o)
    c:\python25\min gw\bin gcc -c example_wrap.c -Ic:
    \python25\inclu de (it create example_wrap.o)

    Step 5:
    building .pyd for windows,
    c:\python25\min gw\bin gcc -shared *.o -o _example.pyd -Lc:
    \python25\libs -lpython25 (it creates _example.pyd)


    Then, i imported example module by palcing it in python25 directory,
    it get's imported but i'm not able to use function, when i called
    >>>print example.fact(4)
    >>>print example.mod(4,2 )
    It is nither showing any error nor message...
    even i tried simple void fact() by printf inside .c file and is not
    showing ...

    hope for help...
    regard's
    Anish
  • RPM1

    #2
    Re: Wrapping C with Python

    Anish Chapagain wrote:
    Hi!!
    I tried wrapping a simple C code suing SWIG to Python, but am having
    problem,
    I am not too familiar with SWIG, (I have looked at it), but you may want
    to try ctypes in the standard module library. It's very easy to use. I
    use it on Windows with gcc but I believe it works fine on Linux too.
    Basically you just compile your C code as a regular C code dll. ctypes
    then allows you to access the functions in the dll very easily. I was
    concerned that performance might be an issue but that has turned out not
    to be true. I'm writing a chess program and the C dll handles the move
    generation so it is getting called a lot. So far it runs at about
    200,000 calls per second.

    Patrick

    Comment

    • brad

      #3
      Re: Wrapping C with Python

      RPM1 wrote:
      ....
      Basically you just compile your C code as a regular C code dll. ctypes
      then allows you to access the functions in the dll very easily.
      Does that work with C++ code too or just C?

      Comment

      • Anish Chapagain

        #4
        Re: Wrapping C with Python

        On 4 Aug, 14:14, brad <byte8b...@gmai l.comwrote:
        RPM1 wrote:
        >
        ...
        >
        Basically you just compile your C code as a regular C code dll.  ctypes
        then allows you to access the functions in the dll very easily.
        >
        Does that work with C++ code too or just C?
        Hi..
        I havenot tried..before with dll, is there any help material and for
        me my API in C has almost 20 c files, so will it be feasible

        anish

        Comment

        • Stefan Behnel

          #5
          Re: Wrapping C with Python

          Anish Chapagain wrote:
          I tried wrapping a simple C code suing SWIG to Python, but am having
          problem,
          Try Cython. It's a Python-like language between Python and C that compiles to
          C code. It makes it very easy to call into C functions and to hide them behind
          a nice Python module.



          Stefan

          Comment

          • RPM1

            #6
            Re: Wrapping C with Python

            brad wrote:
            RPM1 wrote:
            ...
            >Basically you just compile your C code as a regular C code dll.
            >ctypes then allows you to access the functions in the dll very easily.
            >
            Does that work with C++ code too or just C?
            I believe it does work with C++ although I have not done that. Here's a
            simple example:



            I bet if you google around you'll find what you need.

            Remember there is documentation for ctypes in the Python documentation
            that comes with Python.

            Patrick

            Comment

            • RPM1

              #7
              Re: Wrapping C with Python

              Anish Chapagain wrote:
              On 4 Aug, 14:14, brad <byte8b...@gmai l.comwrote:
              >RPM1 wrote:
              >>
              >...
              >>
              >>Basically you just compile your C code as a regular C code dll. ctypes
              >>then allows you to access the functions in the dll very easily.
              >Does that work with C++ code too or just C?
              >
              Hi..
              I havenot tried..before with dll, is there any help material and for
              me my API in C has almost 20 c files, so will it be feasible
              >
              anish
              It is worth looking into it. I found ctypes to be very easy. I had the
              example up and running in minutes. I have also figured out how to do
              things like return arrays to Python. It isn't that hard. I like the
              fact that I don't have to think about reference counting or PyObjects
              etc. If you are doing simple function calls ctypes is very easy.

              My C code looks like C code. It has no Python stuff at all. My Python
              code has just a few calls to the ctypes module. I guess that's one
              thing I like about it is that my C code looks like C code and my Python
              code looks like Python code, (not much spilling over).

              I'm sure it's not perfect, but it seems a lot easier than SWIG to me.

              Patrick

              Comment

              • Terry Reedy

                #8
                Re: Wrapping C with Python



                brad wrote:
                RPM1 wrote:
                ...
                >Basically you just compile your C code as a regular C code dll.
                >ctypes then allows you to access the functions in the dll very easily.
                >
                Does that work with C++ code too or just C?
                On Windows, You can apparently works either with stdcall or cdecl functions.
                "ctypes tries to protect you from calling functions with the wrong
                number of arguments or the wrong calling convention. Unfortunately this
                only works on Windows. It does this by examining the stack after the
                function returns, so although an error is raised the function has been
                called:" "To find out the correct calling convention you have to look
                into the C header file or the documentation for the function you want to
                call." I do not know if these are only C-isms or also C++-isms.


                Comment

                • Uwe Schmitt

                  #9
                  Re: Wrapping C with Python

                  On 4 Aug., 15:14, brad <byte8b...@gmai l.comwrote:
                  RPM1 wrote:
                  >
                  ...
                  >
                  Basically you just compile your C code as a regular C code dll.  ctypes
                  then allows you to access the functions in the dll very easily.
                  >
                  Does that work with C++ code too or just C?
                  It works if the interface of the DLL is C-style, that is you declare
                  your
                  functions with 'extern "C" { .... }' around them.
                  Inside your modules implementation you can use C++ without any
                  problems.

                  What works fine too, is f2py from numpy. It targets at wrapping
                  Fortran
                  code, but is able to wrap C code too, see



                  Greetings, Uwe

                  Greetings, Uwe

                  Comment

                  Working...