Python调用Linux c库：ctypes初学体验

   我在http://jlnsqt.blog.51cto.com/2212965/1405052这篇博客中讲述了匹配URL的一个算法，因项目需要将其封装为动态库，并让python组调用，所以研究了一下ctypes，感觉超级棒，必须记录下来。

   首先介绍一下我的动态库接口。

   动态库结构体：

   

typedef struct _whitelist_tree_node_ {
    uint8_t white_type; //匹配白名单是否结束，代表下一步执行的动作
    uint8_t child_count;                //该节点的子节点个数
    uint8_t child_order[MAX_CHILDS_NUM]; //子节点在childs的位置(起始1)
    struct _whitelist_tree_node_ *childs;   //子节点数组
} whitelist_tree_node;
typedef struct _whitelist_tree_ {
    whitelist_tree_node *root;          //根节点
    unsigned int whitelist_count;       //URL白名单个数
} whitelist_tree, *PUrlWhiteListControl;
                                                                                                                                                                                                                                                                                                                                                    
函数接口：
/**
 * [InitUrlWhiteList 根据配置文件读取数据库中的URL白名单，并初始化到内存中]
 * @param  ini_file   [配置文件路径，一般若无特殊数据库配置可为NULL]
 * @param  p_control  [查询控制器，用完需要关掉]
 * @return            [0: success, other: failure code]
 */
int InitUrlWhiteList(const char *ini_file, PUrlWhiteListControl *p_control);
/**
 * [SearchUrlWhiteList 查询URL是否在白名单中]
 * @param  url        [待查URL]
 * @param  p_control  [查询控制器]
 * @return            [0: find, < 0: run failure code, > 0: not find]
 */
int SearchUrlWhiteList(const char *url, PUrlWhiteListControl p_control);

/**
 * [CloseUrlWhiteList 关闭初始化的查询控制器]
 * @param  control [控制器]
 */
void CloseUrlWhiteList(PUrlWhiteListControl p_control);

   这里不再给出生成动态库方法和各个接口函数的定义，只介绍如何在python中调用。这里假设我动态库的名称为“liburlwhitelist.so”，动态库和python文件在同一目录，或者再引用动态库的时候用绝对路径。

   开始使用ctypes之前，介绍一下ctypes的类型对照：

ctypes type	C type	Python type
c_bool	_Bool	bool (1)
c_char	char	1-character string
c_wchar	wchar_t	1-character unicode string
c_byte	char	int/long
c_ubyte	unsignedchar	int/long
c_short	short	int/long
c_ushort	unsignedshort	int/long
c_int	int	int/long
c_uint	unsignedint	int/long
c_long	long	int/long
c_ulong	unsignedlong	int/long
c_longlong	__int64 or longlong	int/long
c_ulonglong	unsigned__int64 or unsignedlonglong	int/long
c_float	float	float
c_double	double	float
c_longdouble	longdouble	float
c_char_p	char* (NUL terminated)	string or None
c_wchar_p	wchar_t* (NUL terminated)	unicode or None
c_void_p	void*	int/long or None

   首先导入ctypes：

from ctypes import *

   定义最大子节点个数，也即静态数组的大小。    

#max child node number
MAX_NODE_CHILD_NUM = 46

   下面就是重点了，需要用python模拟出Linux C的结构体来。

#define tree node
class whitelist_tree_node(Structure):
    pass
whitelist_tree_node._fields_ = [
    ("white_type", c_ubyte),
    ("child_count", c_ubyte),
    ("child_order", c_ubyte * MAX_NODE_CHILD_NUM),
    ("childs", POINTER(whitelist_tree_node))
]
#define tree
class whitelist_tree(Structure):
    pass
whitelist_tree._fields_ = [
    ("root", POINTER(whitelist_tree_node)),
    ("whitelist_count", c_uint)
]
#define query url whitelist control
PUrlWhiteListControl = POINTER(whitelist_tree)

   为了定义节点指向自己的指针，这里必须先用pass定义一个空的对象，否则会出现“NameError: name 'whitelist_tree_node' is not defined”的错误。取类型的指针用POINTER()函数，而取变量对象的指针用pointer()函数，注意区分。

   导入库，可用绝对路径：    

#load library
url_whitelist_lib = cdll.LoadLibrary("./liburlwhitelist.so")

   引入接口函数，并对接口函数属性进行设置。    

#define init function
InitUrlWhiteList = url_whitelist_lib.InitUrlWhiteList
#define argument types
InitUrlWhiteList.argtypes = [POINTER(PUrlWhiteListControl)]
#define return types. By default functions are assumed to return the C int type
InitUrlWhiteList.restype = c_int
                                                                                                                                                                                        
#define search function
SearchUrlWhiteList = url_whitelist_lib.SearchUrlWhiteList
SearchUrlWhiteList.argtypes = [c_char_p, PUrlWhiteListControl]
SearchUrlWhiteList.restype = c_int
                                                                                                                                                                                        
#define close function
CloseUrlWhiteList = url_whitelist_lib.CloseUrlWhiteList
CloseUrlWhiteList.argtypes = [PUrlWhiteListControl]
CloseUrlWhiteList.restype = None

   定义每个函数第一行“InitUrlWhiteList =url_whitelist_lib.InitUrlWhiteList”是为了减少函数调用。InitUrlWhiteList.argtypes设置函数的参数，为了更好的调用，减少出错。

InitUrlWhiteList.restype设置函数的返回类型，因为ctypes默认的返回类型时C int，我这里还是指出，便于统一和减少出错。

   使用部分例子：

p_control = PUrlWhiteListControl()
ret_code = InitUrlWhiteList(None, pointer(p_control))
if ret_code != 0:
    print "InitUrlWhiteList error code: %d" % ret_code
else:
    print "init url whitelist num: %u" % p_control.contents.whitelist_count

   这里有三个地方需要注意。一、None变量即是C中NULL。二：InitUrlWhiteList参数，因其第二个参数为PUrlWhiteListControl的指针，所以这里用pointer()函数，当然也可用byref()函数，但是在官方文档中指出：

The same effect can be achieved with the pointer() function, although
pointer() does a lot more work since it constructs a real pointer
object, so it is faster to use byref() if you don’t need the pointer
 object in Python itself

   所以这里我们还是用pointer()函数好。三、因PUrlWhiteListControl本身是一个指针，要访问他的内容需要用contents方法。如“p_control.contents.whitelist_count”。

   好了，关于ctypes，先介绍到这里，更详细的请参考官方文档：https://docs.python.org/2/library/ctypes.html。详细代码如下：

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Author: sqt
# @Email:  [email protected]
# @Date:   2014-04-30 09:57:37
# @Desc:   Python测试URL白名单动态库
# @Last Modified by:   sqt
# @Last Modified time: 2014-04-30 13:30:17
               
from ctypes import *
               
#max child node number
MAX_NODE_CHILD_NUM = 46
               
#define tree node
class whitelist_tree_node(Structure):
    pass
               
whitelist_tree_node._fields_ = [
    ("white_type", c_ubyte),
    ("child_count", c_ubyte),
    ("child_order", c_ubyte * MAX_NODE_CHILD_NUM),
    ("childs", POINTER(whitelist_tree_node))
]
               
#define tree
class whitelist_tree(Structure):
    pass
               
whitelist_tree._fields_ = [
    ("root", POINTER(whitelist_tree_node)),
    ("whitelist_count", c_uint)
]
               
#define query url whitelist control
PUrlWhiteListControl = POINTER(whitelist_tree)
#load library
url_whitelist_lib = cdll.LoadLibrary("./liburlwhitelist.so")
               
#simple call function name
#define init function
InitUrlWhiteList = url_whitelist_lib.InitUrlWhiteList
#define argument types
InitUrlWhiteList.argtypes = [POINTER(PUrlWhiteListControl)]
#define return types. By default functions are assumed to return the C int type
InitUrlWhiteList.restype = c_int
               
#define search function
SearchUrlWhiteList = url_whitelist_lib.SearchUrlWhiteList
SearchUrlWhiteList.argtypes = [c_char_p, PUrlWhiteListControl]
SearchUrlWhiteList.restype = c_int
               
#define close function
CloseUrlWhiteList = url_whitelist_lib.CloseUrlWhiteList
CloseUrlWhiteList.argtypes = [PUrlWhiteListControl]
CloseUrlWhiteList.restype = None
               
#sample
if __name__ == "__main__":
    p_control = PUrlWhiteListControl()
    ret_code = InitUrlWhiteList(None, pointer(p_control))
    if ret_code != 0:
        print "InitUrlWhiteList error code: %d" % ret_code
    else:
        print "init url whitelist num: %u" % p_control.contents.whitelist_count
                
    url = ""
    while 1:
        url = raw_input("Input url or exit:")
        if url == "q" or url == "exit":
            break
               
        find_code = SearchUrlWhiteList(url, p_control)
        if find_code == 0:
            print "url: %s find." % url
        else:
            print "url: %s not find." % url
                
    CloseUrlWhiteList(p_control)
    del p_control