如何从C调用相同python脚本的多个实例?

时间:2015-10-08 20:36:38

标签: python

我想在C中嵌入python并从C调用相同python脚本的多个实例。我该怎么做?

任何示例代码都会有所帮助.GIL如何影响我?

谢谢..

这是示例程序....哪个给出了错误的结果..因为我只有一个解释器......我需要调用多个解释器......我该怎么做?

-------------------C Program----------------
#include "Python.h"

int main(int argc, char* argv[])
{
    PyObject    *pName, *pModule[2], *pFunc;
   PyObject    *pArgs, *pValue;
   int i = 0;
   int a, b;

   Py_SetProgramName(argv[0]);
   Py_Initialize(); /* initialize the python interpreter */
   for ( i = 0; i < 2; i++) {
       pName = PyString_FromString("scr");
       pModule[i] = PyImport_Import(pName);

    if (pModule[i] != NULL) {
           pFunc = PyObject_GetAttrString(pModule[i], "add");
        if (pFunc && PyCallable_Check(pFunc)) {

            printf ("Enter the numbers to add to run: ");
            scanf ("%d %d", &a, &b);
            pArgs = PyTuple_New(2);
            pValue = PyInt_FromLong (a);
            PyTuple_SetItem(pArgs, 0, pValue);
            Py_DECREF(pValue);

            pValue = PyInt_FromLong (b);
            PyTuple_SetItem(pArgs, 1, pValue);
            Py_DECREF(pValue);

            pValue = PyObject_CallObject(pFunc, pArgs);
            Py_DECREF(pArgs);

            if (pValue != NULL) {
                printf("Result of call: %ld\n", PyInt_AsLong(pValue));
                Py_DECREF(pValue);
            }
            else {
                Py_DECREF(pFunc);
                Py_DECREF(pModule[i]);
                PyErr_Print();
                fprintf(stderr,"Call failed\n");
                return 1;
            }
        }
    }
    else {

        if (PyErr_Occurred())
            PyErr_Print();
        fprintf(stderr, "Cannot find  function \n");
    }
    Py_DECREF(pFunc);
  }

 /* Print result */
  for ( i = 0; i < 2; i++) {
     // pName = PyString_FromString("scr");
    //pModule[i] = PyImport_Import(pName);

    if (pModule[i] != NULL) {
        pFunc = PyObject_GetAttrString(pModule[i], "print_result");
        if (pFunc && PyCallable_Check(pFunc)) {
            pValue = PyObject_CallObject(pFunc, NULL);
            if (pValue != NULL) {
                printf("Call succeeded\n");
            }
            else {
                printf("Call succeeded\n");
                PyErr_Print();
                fprintf(stderr,"Call failed\n");
                return 1;
            }
        }
    }
}

Py_Finalize();
return 0;
}
--------------- Python Script ---------------------
$ cat scr.py
c = 0
a = 0
b = 0

def add(a1,b1):
    global a
    global b
    global c
    a = a1
    b = b1
    print "Python: Will compute ", a,"+", b
    c = a + b
    return c

 def print_result():
     print 'Python: Result of ' , a , ' + ' , b, 'is: ', c
     return
 ---------------- Actual Output-------------------
 Enter the numbers to add to run: 2 3
 Python: Will compute  2 + 3
 Result of call: 5
 Enter the numbers to add to run: 5 5
 Python: Will compute  5 + 5
 Result of call: 10
 Python: Result of  5  +  5 is:  10 <--- PROBLEM
 Call succeeded
 Python: Result of  5  +  5 is:  10
 Call succeeded
 -------------------- Expected Output --------------------
 Enter the numbers to add to run: 2 3
 Python: Will compute  2 + 3
 Result of call: 5
 Enter the numbers to add to run: 5 5
 Python: Will compute  5 + 5
 Result of call: 10
 Python: Result of  2  +  3 is:  5
 Call succeeded
 Python: Result of  5  +  5 is:  10
 Call succeeded
 -------------------------------------------------------

1 个答案:

答案 0 :(得分:0)

通过导入Python.h并调用Py_Initialize,您将创建一个全局Python解释器。这将保留已加载模块的字典:第二次尝试导入同一模块时,该字典将返回与第一次相同的模块。这就是Python(和嵌入式Python)通常需要工作的方式 - 就像在Python脚本中进行第二次导入一样。

可能使用“sub interpreter”api通过调用Py_NewInterpreter()创建一个新的子解释器。这可以用作具有自己的新模块字典等的新解释器。您可以使用PyThreadState_Swap()在子解释器之间交换。

但是,我没有这样做,看起来有点记录不清。有一些文档,特别是一些警告here。显然,副译员并不是完全分开的,尽管它们可能足以满足您的目的。

在你的情况下,我认为你不必过分担心GIL,因为你不打算同时执行两个解释器。

所以我觉得这样的事情应该有用,虽然我没有测试过。

Py_Initialize();

// Do something in first interpreter 

PyThreadState *tstate_1 = PyThreadState_Get();
PyThreadState *tstate_2 = Py_NewInterpreter();

// Do something in second interpreter 

PyThreadState_Swap(tstate_1);

// Go back to using the first interpreter