有没有办法从 Python-C API 定义 python 数据模型方法

Question

I've written a simple python module in the python-C API, that defines a class with an integer type and a method to add two of the same class.

static PyObject *IntObject_add(IntObject *self, PyObject *args)
{
    PyObject *other, *rv_args, *rv, *rv_long;
    if(!PyArg_ParseTuple(args, "O", &other))
        return NULL;
    //add type checking here
    long rv_val = ((IntObject *) other)->number + self->number;
    rv_long = PyLong_FromLong(rv_val);
    rv_args = PyTuple_New(1);
    PyTuple_SetItem(rv_args, 0, rv_long);
    rv = PyObject_CallObject((PyObject *) &IntObject_T, rv_args);
    Py_DECREF(rv_long);
    Py_DECREF(rv_args);
    return rv;
}

static PyMethodDef IntObject_methods[] = {
    {"__add__", (PyCFunction) IntObject_add, METH_VARARGS,
    "add two IntObject types"},
    {NULL}
};

忽略缺少类型检查，当调用为object1.__add__(object2)时，它确实返回正确的结果，但是当调用为object1 + object2而不是使用__add__ function 作为+运算符的数据模型方法时，会引发 TypeError。 在 Python-C API 中是否有不同的方法来定义数据模型方法，或者这是一种不正确的方法。