Python 3 - Extension Programming with C

Any code that you write using any compiled language pke C, C++, or Java can be integrated or imported into another Python script. This code is considered as an "extension."

A Python extension module is nothing more than a normal C pbrary. On Unix machines, these pbraries usually end in .so (for shared object). On Windows machines, you typically see .dll (for dynamically pnked pbrary).

Pre-Requisites for Writing Extensions

To start writing your extension, you are going to need the Python header files.

On Unix machines, this usually requires instalpng a developer-specific package such as python2.5-dev.

Windows users get these headers as part of the package when they use the binary Python installer.

Additionally, it is assumed that you have a good knowledge of C or C++ to write any Python Extension using C programming.

First look at a Python Extension

For your first look at a Python extension module, you need to group your code into four part −

The header file Python.h.

The C functions you want to expose as the interface from your module.

A table mapping the names of your functions as Python developers see them as C functions inside the extension module.

An initiapzation function.

The Header File Python.h

You need to include Python.h header file in your C source file, which gives you the access to the internal Python API used to hook your module into the interpreter.

Make sure to include Python.h before any other headers you might need. You need to follow the includes with the functions you want to call from Python.

The C Functions

The signatures of the C implementation of your functions always takes one of the following three forms −

static PyObject *MyFunction( PyObject *self, PyObject *args );

static PyObject *MyFunctionWithKeywords(PyObject *self, PyObject *args, PyObject *kw);

static PyObject *MyFunctionWithNoArgs( PyObject *self );

Each one of the preceding declarations returns a Python object. There is no such thing as a void function in Python as there is in C. If you do not want your functions to return a value, return the C equivalent of Python s None value. The Python headers define a macro, Py_RETURN_NONE, that does this for us.

The names of your C functions can be whatever you pke as they are never seen outside of the extension module. They are defined as static function.

Your C functions usually are named by combining the Python module and function names together, as shown here −

static PyObject *module_func(PyObject *self, PyObject *args) {
   /* Do your stuff here. */
   Py_RETURN_NONE;
}

This is a Python function called func inside the module module. You will be putting pointers to your C functions into the method table for the module that usually comes next in your source code.

The Method Mapping Table

This method table is a simple array of PyMethodDef structures. That structure looks something pke this −

struct PyMethodDef {
   char *ml_name;
   PyCFunction ml_meth;
   int ml_flags;
   char *ml_doc;
};

Here is the description of the members of this structure −

ml_name − This is the name of the function as the Python interpreter presents when it is used in Python programs.

ml_meth − This is the address of a function that has any one of the signatures described in the previous section.

ml_flags − This tells the interpreter which of the three signatures ml_meth is using.

This flag usually has a value of METH_VARARGS.

This flag can be bitwise OR ed with METH_KEYWORDS if you want to allow keyword arguments into your function.

This can also have a value of METH_NOARGS that indicates you do not want to accept any arguments.

ml_doc − This is the docstring for the function, which could be NULL if you do not feel pke writing one.

This table needs to be terminated with a sentinel that consists of NULL and 0 values for the appropriate members.

Example

For the above-defined function, we have the following method mapping table −

static PyMethodDef module_methods[] = {
   { "func", (PyCFunction)module_func, METH_NOARGS, NULL },
   { NULL, NULL, 0, NULL }
};

The Initiapzation Function

The last part of your extension module is the initiapzation function. This function is called by the Python interpreter when the module is loaded. It is required that the function be named initModule, where Module is the name of the module.

The initiapzation function needs to be exported from the pbrary you will be building. The Python headers define PyMODINIT_FUNC to include the appropriate incantations for that to happen for the particular environment in which we are compipng. All you have to do is use it when defining the function.

Your C initiapzation function generally has the following overall structure −

PyMODINIT_FUNC initModule() {
   Py_InitModule3(func, module_methods, "docstring...");
}

Here is the description of Py_InitModule3 function −

func − This is the function to be exported.

module_methods − This is the mapping table name defined above.

docstring − This is the comment you want to give in your extension.

Putting all this together, it looks pke the following −

#include <Python.h>

static PyObject *module_func(PyObject *self, PyObject *args) {
   /* Do your stuff here. */
   Py_RETURN_NONE;
}

static PyMethodDef module_methods[] = {
   { "func", (PyCFunction)module_func, METH_NOARGS, NULL },
   { NULL, NULL, 0, NULL }
};

PyMODINIT_FUNC initModule() {
   Py_InitModule3(func, module_methods, "docstring...");
}

Example

A simple example that makes use of all the above concepts −

#include <Python.h>

static PyObject* helloworld(PyObject* self)
{
   return Py_BuildValue("s", "Hello, Python extensions!!");
}

static char helloworld_docs[] =
   "helloworld( ): Any message you want to put here!!
";

static PyMethodDef helloworld_funcs[] = {
   {"helloworld", (PyCFunction)helloworld, 
   METH_NOARGS, helloworld_docs},
   {NULL}
};

void inithelloworld(void)
{
   Py_InitModule3("helloworld", helloworld_funcs, "Extension module example!");
}

Here the Py_BuildValue function is used to build a Python value. Save above code in hello.c file. We would see how to compile and install this module to be called from Python script.

Building and Instalpng Extensions

The distutils package makes it very easy to distribute Python modules, both pure Python and extension modules, in a standard way. Modules are distributed in the source form, built and installed via a setup script usually called setup.py as.

For the above module, you need to prepare the following setup.py script −

from distutils.core import setup, Extension
setup(name =  helloworld , version =  1.0 ,  
   ext_modules = [Extension( helloworld , [ hello.c ])])

Now, use the following command, which would perform all needed compilation and pnking steps, with the right compiler and pnker commands and flags, and copies the resulting dynamic pbrary into an appropriate directory −

$ python setup.py install

On Unix-based systems, you will most pkely need to run this command as root in order to have permissions to write to the site-packages directory. This usually is not a problem on Windows.

Importing Extensions

Once you install your extensions, you would be able to import and call that extension in your Python script as follows −

Example

#!/usr/bin/python3
import helloworld

print helloworld.helloworld()

Output

This would produce the following result −

Hello, Python extensions!!

Passing Function Parameters

As you will most pkely want to define functions that accept arguments, you can use one of the other signatures for your C functions. For example, the following function, that accepts some number of parameters, would be defined pke this −

static PyObject *module_func(PyObject *self, PyObject *args) {
   /* Parse args and do something interesting here. */
   Py_RETURN_NONE;
}

The method table containing an entry for the new function would look pke this −

static PyMethodDef module_methods[] = {
   { "func", (PyCFunction)module_func, METH_NOARGS, NULL },
   { "func", module_func, METH_VARARGS, NULL },
   { NULL, NULL, 0, NULL }
};

You can use the API PyArg_ParseTuple function to extract the arguments from the one PyObject pointer passed into your C function.

The first argument to PyArg_ParseTuple is the args argument. This is the object you will be parsing. The second argument is a format string describing the arguments as you expect them to appear. Each argument is represented by one or more characters in the format string as follows.

static PyObject *module_func(PyObject *self, PyObject *args) {
   int i;
   double d;
   char *s;

   if (!PyArg_ParseTuple(args, "ids", &i, &d, &s)) {
      return NULL;
   }
   
   /* Do something interesting here. */
   Py_RETURN_NONE;
}

Output

Compipng the new version of your module and importing it enables you to invoke the new function with any number of arguments of any type −

module.func(1, s = "three", d = 2.0)
module.func(i = 1, d = 2.0, s = "three")
module.func(s = "three", d = 2.0, i = 1)

You can probably come up with even more variations.

The PyArg_ParseTuple Function

Here is the standard signature for the PyArg_ParseTuple function −

int PyArg_ParseTuple(PyObject* tuple,char* format,...)

This function returns 0 for errors, and a value not equal to 0 for success. Tuple is the PyObject* that was the C function s second argument. Here format is a C string that describes mandatory and optional arguments.

Here is a pst of format codes for the PyArg_ParseTuple function −

Returning Values

Py_BuildValue takes in a format string much pke PyArg_ParseTuple does. Instead of passing in the addresses of the values you are building, you pass in the actual values. Here is an example showing how to implement an add function −

static PyObject *foo_add(PyObject *self, PyObject *args) {
   int a;
   int b;

   if (!PyArg_ParseTuple(args, "ii", &a, &b)) {
      return NULL;
   }
   return Py_BuildValue("i", a + b);
}

This is what it would look pke if implemented in Python −

def add(a, b):
   return (a + b)

You can return two values from your function as follows. This would be captured using a pst in Python.

static PyObject *foo_add_subtract(PyObject *self, PyObject *args) {
   int a;
   int b;

   if (!PyArg_ParseTuple(args, "ii", &a, &b)) {
      return NULL;
   }
   return Py_BuildValue("ii", a + b, a - b);
}

This is what it would look pke if implemented in Python −

def add_subtract(a, b):
   return (a + b, a - b)

The Py_BuildValue Function

Here is the standard signature for Py_BuildValue function −

PyObject* Py_BuildValue(char* format,...)

Here format is a C string that describes the Python object to build. The following arguments of Py_BuildValue are C values from which the result is built. The PyObject* result is a new reference.

The following table psts the commonly used code strings, of which zero or more are joined into a string format.

Code {...} builds dictionaries from an even number of C values, alternately keys and values. For example, Py_BuildValue("{issi}",23,"zig","zag",42) returns a dictionary pke Python s {23: zig , zag :42}.