[Translation] Embed Python into another application -- from the official website

Waste firewood girl in the fishpond Released | November 1, 2018| Programming learning |10934 Browse

<trans data-src="最近因为项目上需要用C++调用pytorch训练好的模型，不得已开始查这方面的资料，正好看到官方文档的第五章[将Python嵌入到另一个应用程序中](">Recently, because the project needs to use C++to call the model trained by pytorch, I had to start checking this information. I just saw Chapter 5 of the official document [Embedding Python into another application]（</trans> <trans data-src="https://docs.python.org/2/extending/embedding.html">https://docs.python.org/2/extending/embedding.html</trans> <trans data-src=")有讲到这方面的知识，正好翻译一下，也贡献给有需要的小伙伴。">）If you have some knowledge about this, just translate it, and also contribute to the little partners in need.</trans> <trans data-src="（注意：本章中主要讲到的是将python嵌入C程序中）">(Note: This chapter mainly talks about embedding python into C programs.)</trans> <trans data-src="ps: 文中部分翻译可能不太恰当，欢迎小伙伴们进行批评指正~">Ps: Some translations in the text may not be appropriate, and you are welcome to criticize and correct~</trans> <trans data-src="另ps：本废柴少女不仅废柴，还懒。">PS: This waste firewood girl is not only waste firewood, but also lazy.</trans> <trans data-src="So直接从我的CSDN账号里原封不动地copy过来了...">So directly copied from my CSDN account intact</trans> <trans data-src="## 5.在另一款应用程序中植入python">##5. Embed python in another application</trans> <trans data-src="在之前的章节中已经讨论了如何扩展python，即，如何给python附加一个C函数库来扩展其功能。">In the previous chapter, we discussed how to extend python, that is, how to attach a C function library to python to extend its functions.</trans> <trans data-src="实际上，反过来做也是可行的：通过植入python来丰富你的C/C++程序。">In fact, it is also feasible to do the opposite: enrich your C/C++programs by embedding python.</trans> <trans data-src="通过植入python你可以在你的程序中实现一些你用python编写的功能，而不是C/C++。">By embedding python, you can implement some functions you wrote in python in your program, instead of C/C++.</trans> <trans data-src="这可以用来满足多种需求，其中一个例子是允许用户根据自己的需要写一些python脚本来定制应用程序。">This can be used to meet a variety of requirements. One example is to allow users to write some python scripts to customize applications according to their needs.</trans> <trans data-src="当然，如果一些功能使用python编写起来要更容易，你也可以这么做。">Of course, if some functions are easier to write in Python, you can do the same.</trans> <trans data-src="嵌入python类似于扩展它，但并不完全如此。">Embedding python is similar to extending it, but not quite.</trans> <trans data-src="区别在于，当你扩展python时，应用程序中的主程序依然是依赖于python解释器；">The difference is that when you extend python, the main program in the application still depends on the python interpreter;</trans> <trans data-src="而当你需要嵌入python时，你的主程序实际上已经和python没什么关系了都——相反，应用程序中的某些部分只会偶尔需要调用python解释器来运行其中的python代码。">When you need to embed python, your main program actually has nothing to do with python -- on the contrary, some parts of the application only occasionally need to call the python interpreter to run its python code.</trans> <trans data-src="因此，如果你正要嵌入python，你得提供你自己的主程序。">Therefore, if you are going to embed python, you have to provide your own main program.</trans> <trans data-src="你的主程序必须要做的事情之一是你必须去初始化python解释器。">One of the things your main program must do is to initialize the python interpreter.</trans> <trans data-src="起码你得调用你的[Py_Initialize（）](">At least you must call your [Py_Initialize()]（</trans> <trans data-src="https://docs.python.org/2/c-api/init.html#c.Py_Initialize">https://docs.python.org/2/c-api/init.html#c.Py_Initialize</trans> <trans data-src=")函数。">）Function.</trans> <trans data-src="还有一些可选的调用将命令行参数传给python。">There are also some optional calls to pass command line parameters to python.</trans> <trans data-src="之后，你就可以在你应用程序的任何部位去调用这个解释器了。">After that, you can call the interpreter anywhere in your application.</trans> <trans data-src="调用这个解释器的途径有很多：你可以传给[PyRun_SimpleString()](">There are many ways to call this interpreter: you can pass it to [PyRun_SimpleString()]（</trans> <trans data-src="https://docs.python.org/2/c-api/veryhigh.html#c.PyRun_SimpleString">https://docs.python.org/2/c-api/veryhigh.html#c.PyRun_SimpleString</trans> <trans data-src=")一个包含python语句的字符串，或者你也可以传给[PyRun_SimpleFile()](">）A string containing python statements, or you can also pass it to [PyRun_SimpleFile()]（</trans> <trans data-src="https://docs.python.org/2/c-api/veryhigh.html#c.PyRun_SimpleFile">https://docs.python.org/2/c-api/veryhigh.html#c.PyRun_SimpleFile</trans> <trans data-src=")一个stdio文件指针和一个文件名（这仅用于识别错误消息）。">）A stdio file pointer and a file name (this is only used to identify error messages).</trans> <trans data-src="此外，你还可以调用前几章中描述得较低级别的操作来构造和使用python对象。">In addition, you can also call the lower level operations described in the previous chapters to construct and use python objects.</trans> <trans data-src="你可以在发行版的源代码下的 ***Demo/embed/***目录中找到一个嵌入python的简单示例。">You can find a simple example of embedding python in the * * * Demo/embedded/* * * directory under the source code of the distribution.</trans> <trans data-src="另请参阅：">See also:</trans> <trans data-src="[Python/C API 参考手册](">[Python/C API Reference Manual]（</trans> <trans data-src="https://docs.python.org/2/c-api/index.html#c">https://docs.python.org/2/c-api/index.html#c</trans> <trans data-src="-api-index)">-api-index)</trans> <trans data-src="本手册提供了Python C接口的详细信息。">This manual provides details of the Python C interface.</trans> <trans data-src="在这里可以找到大量必要的信息。">You can find a lot of necessary information here.</trans> <trans data-src="## 5.1 the very high level 接口">##5.1 the very high level interface</trans> <trans data-src="嵌入python的最简单形式是使用***the very high level *** 接口，这个接口不需要与应用程序直接交互，就可以被用来执行python脚本。">The simplest way to embed python is to use the * * * the very high level * * * interface, which can be used to execute python scripts without direct interaction with applications.</trans> <trans data-src="例如，这可以用于对文件执行某些操作。">For example, this can be used to perform certain operations on files.</trans> <trans data-src="```">```</trans> <trans data-src="#include <Python.h>">#include <Python.h></trans> <trans data-src="int main(int argc, char *argv[])">int main(int argc, char *argv[])</trans> <trans data-src="{">{</trans> <trans data-src="Py_SetProgramName(argv[0]);  /* ">Py_SetProgramName(argv[0]);  /* </trans> <trans data-src="optional but recommended */">optional but recommended */</trans> <trans data-src="Py_Initialize();">Py_Initialize();</trans> <trans data-src="PyRun_SimpleString("from time import time,ctime\n"">PyRun_SimpleString("from time import time,ctime\n"</trans> <trans data-src=""print 'Today is',ctime(time())\n");">"print 'Today is',ctime(time())\n");</trans> <trans data-src="Py_Finalize();">Py_Finalize();</trans> <trans data-src="return 0;">return 0;</trans> <trans data-src="}">}</trans> <trans data-src="```">```</trans> <trans data-src="[ Py_SetProgramName()](">[ Py_SetProgramName()](</trans> <trans data-src="https://docs.python.org/2/c-api/init.html#c.Py_SetProgramName">https://docs.python.org/2/c-api/init.html#c.Py_SetProgramName</trans> <trans data-src=")函数应该在[Py_Initialize（）](">）Function should be in [Py_Initialize()]（</trans> <trans data-src="https://docs.python.org/2/c-api/init.html#c.Py_Initialize">https://docs.python.org/2/c-api/init.html#c.Py_Initialize</trans> <trans data-src=")函数之前被调用，目的是告知解释器python运行时库的路径。">）The function is called before to inform the interpreter of the path of the python runtime library.</trans> <trans data-src="接下来，python解释器会通过[Py_Initialize（）](">Next, the python interpreter will use [Py_Initialize()]（</trans> <trans data-src="https://docs.python.org/2/c-api/init.html#c.Py_Initialize">https://docs.python.org/2/c-api/init.html#c.Py_Initialize</trans> <trans data-src=")函数进行初始化，然后执行硬编码的Python脚本，打印日期和时间。">）Function, and then execute the hard coded Python script to print the date and time.</trans> <trans data-src="随后，调用[Py_Finalize()](">Then, call [Py_Finalize()]（</trans> <trans data-src="https://docs.python.org/2/c-api/init.html#c.Py_Finalize">https://docs.python.org/2/c-api/init.html#c.Py_Finalize</trans> <trans data-src=")函数关闭解释器，然后程序结束。">）The function closes the interpreter and the program ends.</trans> <trans data-src="在实际的程序中，你可能希望从另一个源（可能是文本编辑器例程、文件或数据库）获得python脚本。">In an actual program, you may want to obtain a python script from another source (possibly a text editor routine, file, or database).</trans> <trans data-src="但从文件中获取python代码的更好方式是使用[PyRun_SimpleFile()](">But a better way to get python code from a file is to use [PyRun_SimpleFile()]（</trans> <trans data-src="https://docs.python.org/2/c-api/veryhigh.html#c.PyRun_SimpleFile">https://docs.python.org/2/c-api/veryhigh.html#c.PyRun_SimpleFile</trans> <trans data-src=")函数，这样就省去了分配内存空间和加载文件内容的麻烦。">）Function to save the trouble of allocating memory space and loading file contents.</trans> <trans data-src="## 5.2 超越非常高水平的嵌入:概述">##5.2 Beyond Very High Level Embedding: Overview</trans> <trans data-src="高级接口的确可以使你从应用程序中执行任意的python代码片段，但不得不说，交换数据值相当麻烦。">Advanced interfaces can indeed enable you to execute arbitrary python code fragments from the application, but it must be said that exchanging data values is quite troublesome.</trans> <trans data-src="如果你想要这样做，你应该使用更低级别的调用。">If you want to do this, you should use a lower level call.</trans> <trans data-src="当然，如果你不在乎编写更多更长的C代码，那你几乎可以实现任何目标。">Of course, if you don't care about writing more and longer C code, you can achieve almost any goal.</trans> <trans data-src="值得注意的是，扩展python和嵌入python是几乎完全相同的任务，尽管他们的出发点是不同的。">It is worth noting that extending python and embedding python are almost identical tasks, although their starting points are different.</trans> <trans data-src="前几章讨论过的大多数主题依然有效。">Most of the topics discussed in the previous chapters remain valid.</trans> <trans data-src="为了说明这一点，我们不妨以从python到C的扩展代码为例，看它究竟实际做了什么：">To illustrate this, let's take the extension code from python to C as an example to see what it actually does:</trans> <trans data-src="1）将数据值从python转换为C">1) Convert data values from python to C</trans> <trans data-src="2）使用转换后的值对C例程执行函数调用">2) Use the converted value to make a function call to the C routine</trans> <trans data-src="3）将调用中的数据值从C转换成Python">3) Convert the data value in the call from C to Python</trans> <trans data-src="当嵌入python时，这个接口代码做的工作如下：">When embedding python, the interface code does the following:</trans> <trans data-src="1）将数据值从C转换为python">1) Convert data values from C to python</trans> <trans data-src="2）使用转换后的值对python接口例程执行函数调用">2) Use the converted value to call the python interface routine</trans> <trans data-src="3）将调用中的数据值从python转换为C">3) Convert the data value in the call from python to C</trans> <trans data-src="如你所见，数据转换步骤简单进行了交换以适应跨语言传输时的不同转换方向。">As you can see, the data conversion steps are simply exchanged to adapt to different conversion directions in cross language transmission.</trans> <trans data-src="唯一的区别在于两种数据转换间调用的例程。">The only difference lies in the routines called between the two data transformations.</trans> <trans data-src="当你需要扩展python时，调用C例程；">When you need to extend python, call C routines;</trans> <trans data-src="当你需要嵌入python时，调用python例程。">When you need to embed python, call the python routine.</trans> <trans data-src="本章我们不讨论如何将数据从python转换为C，反之亦然。">In this chapter, we will not discuss how to convert data from python to C, and vice versa.</trans> <trans data-src="此外，正确地使用引用和处理错误是可以理解的。">In addition, it is understandable to use references and handle errors correctly.</trans> <trans data-src="这些方面与扩展解释器并没什么太大区别，你可以参考前面的章节以获得所需的信息。">These aspects are not very different from the extension interpreter. You can refer to the previous chapters for the required information.</trans> <trans data-src="## 5.3 纯粹的嵌入">##5.3 Pure embedding</trans> <trans data-src="下面的第一个程序目的是执行python脚本中的一个功能。">The first program below is intended to execute a function in a python script.</trans> <trans data-src="正如我们在5.1节中提到的，python解释器不会直接与应用程序交互（但这点在下一节中将有所不同）">As we mentioned in section 5.1, the python interpreter will not directly interact with the application (but this will be different in the next section)</trans> <trans data-src="运行python脚本中定义的函数代码如下：">The function code defined in running the python script is as follows:</trans> <trans data-src="```">```</trans> <trans data-src="#include <Python.h>">#include <Python.h></trans> <trans data-src="int main(int argc, char *argv[])">int main(int argc, char *argv[])</trans> <trans data-src="{">{</trans> <trans data-src="PyObject *pName, *pModule, *pFunc;">PyObject *pName, *pModule, *pFunc;</trans> <trans data-src="PyObject *pArgs, *pValue;">PyObject *pArgs, *pValue;</trans> <trans data-src="int i;">int i;</trans> <trans data-src="if (argc < 3) {">if (argc < 3) {</trans> <trans data-src="fprintf(stderr,"Usage: call pythonfile funcname [args]\n");">fprintf(stderr,"Usage: call pythonfile funcname [args]\n");</trans> <trans data-src="return 1;">return 1;</trans> <trans data-src="}">}</trans> <trans data-src="Py_Initialize();">Py_Initialize();</trans> <trans data-src="pName = PyString_FromString(argv[1]);">pName = PyString_FromString(argv[1]);</trans> <trans data-src="/* Error checking of pName left out */">/* Error checking of pName left out */</trans> <trans data-src="pModule = PyImport_Import(pName);">pModule = PyImport_Import(pName);</trans> <trans data-src="Py_DECREF(pName);">Py_DECREF(pName);</trans> <trans data-src="if (pModule != NULL) {">if (pModule != NULL) {</trans> <trans data-src="pFunc = PyObject_GetAttrString(pModule, argv[2]);">pFunc = PyObject_GetAttrString(pModule, argv[2]);</trans> <trans data-src="/* pFunc is a new reference */">/* pFunc is a new reference */</trans> <trans data-src="if (pFunc && PyCallable_Check(pFunc)) {">if (pFunc && PyCallable_Check(pFunc)) {</trans> <trans data-src="pArgs = PyTuple_New(argc - 3);">pArgs = PyTuple_New(argc - 3);</trans> <trans data-src="for (i = 0; i < argc - 3; ++i) {">for (i = 0; i < argc - 3; ++i) {</trans> <trans data-src="pValue = PyInt_FromLong(atoi(argv[i + 3]));">pValue = PyInt_FromLong(atoi(argv[i + 3]));</trans> <trans data-src="if (!pValue) {">if (!pValue) {</trans> <trans data-src="Py_DECREF(pArgs);">Py_DECREF(pArgs);</trans> <trans data-src="Py_DECREF(pModule);">Py_DECREF(pModule);</trans> <trans data-src="fprintf(stderr, "Cannot convert argument\n");">fprintf(stderr, "Cannot convert argument\n");</trans> <trans data-src="return 1;">return 1;</trans> <trans data-src="}">}</trans> <trans data-src="/* pValue reference stolen here: */">/* pValue reference stolen here: */</trans> <trans data-src="PyTuple_SetItem(pArgs, i, pValue);">PyTuple_SetItem(pArgs, i, pValue);</trans> <trans data-src="}">}</trans> <trans data-src="pValue = PyObject_CallObject(pFunc, pArgs);">pValue = PyObject_CallObject(pFunc, pArgs);</trans> <trans data-src="Py_DECREF(pArgs);">Py_DECREF(pArgs);</trans> <trans data-src="if (pValue != NULL) {">if (pValue != NULL) {</trans> <trans data-src="printf("Result of call: %ld\n", PyInt_AsLong(pValue));">printf("Result of call: %ld\n", PyInt_AsLong(pValue));</trans> <trans data-src="Py_DECREF(pValue);">Py_DECREF(pValue);</trans> <trans data-src="}">}</trans> <trans data-src="else {">else {</trans> <trans data-src="Py_DECREF(pFunc);">Py_DECREF(pFunc);</trans> <trans data-src="Py_DECREF(pModule);">Py_DECREF(pModule);</trans> <trans data-src="PyErr_Print();">PyErr_Print();</trans> <trans data-src="fprintf(stderr,"Call failed\n");">fprintf(stderr,"Call failed\n");</trans> <trans data-src="return 1;">return 1;</trans> <trans data-src="}">}</trans> <trans data-src="}">}</trans> <trans data-src="else {">else {</trans> <trans data-src="if (PyErr_Occurred())">if (PyErr_Occurred())</trans> <trans data-src="PyErr_Print();">PyErr_Print();</trans> <trans data-src="fprintf(stderr, "Cannot find function \"%s\"\n", argv[2]);">fprintf(stderr, "Cannot find function \"%s\"\n", argv[2]);</trans> <trans data-src="}">}</trans> <trans data-src="Py_XDECREF(pFunc);">Py_XDECREF(pFunc);</trans> <trans data-src="Py_DECREF(pModule);">Py_DECREF(pModule);</trans> <trans data-src="}">}</trans> <trans data-src="else {">else {</trans> <trans data-src="PyErr_Print();">PyErr_Print();</trans> <trans data-src="fprintf(stderr, "Failed to load \"%s\"\n", argv[1]);">fprintf(stderr, "Failed to load \"%s\"\n", argv[1]);</trans> <trans data-src="return 1;">return 1;</trans> <trans data-src="}">}</trans> <trans data-src="Py_Finalize();">Py_Finalize();</trans> <trans data-src="return 0;">return 0;</trans> <trans data-src="}">}</trans> <trans data-src="```">```</trans> <trans data-src="这个代码块使用命令行参数 ***argv[1]***来载入python脚本，并且调用参数***argv[2]***中命名的函数。">This code block uses the command line parameter * * * argv [1] * * * to load the python script, and calls the function named in the parameter * * * argv [2] * * *.</trans> <trans data-src="它的整数参数是argv数组的其他值。">Its integer parameters are other values of argv array.</trans> <trans data-src="如果你编译并链接这个程序(我们不妨将完成的可执行程序称之为 ***call*** ），并且使用它来执行python脚本，例如：">If you compile and link this program (let's call the completed executable program * * * call * * *), and use it to execute python scripts, for example:</trans> <trans data-src="```">```</trans> <trans data-src="def multiply(a,b):">def multiply(a,b):</trans> <trans data-src="print "Will compute", a, "times", b">print "Will compute", a, "times", b</trans> <trans data-src="c = 0">c = 0</trans> <trans data-src="for i in range(0, a):">for i in range(0, a):</trans> <trans data-src="c = c + b">c = c + b</trans> <trans data-src="return c">return c</trans> <trans data-src="```">```</trans> <trans data-src="随后将得到：">Then you will get:</trans> <trans data-src="```">```</trans> <trans data-src="$ call multiply multiply 3 2">$ call multiply multiply 3 2</trans> <trans data-src="Will compute 3 times 2">Will compute 3 times 2</trans> <trans data-src="Result of call: 6">Result of call: 6</trans> <trans data-src="```">```</trans> <trans data-src="尽管该程序的功能块看起来非常庞大，但其实大部分都是用于python和C之间的数据转换，以及错误报告。">Although the function blocks of this program look very large, in fact, most of them are used for data conversion between python and C, as well as error reporting.</trans> <trans data-src="关于嵌入python，下面开始才是真正有趣的部分：">The following is the really interesting part about embedding python:</trans> <trans data-src="```">```</trans> <trans data-src="Py_Initialize();">Py_Initialize();</trans> <trans data-src="pName = PyString_FromString(argv[1]);">pName = PyString_FromString(argv[1]);</trans> <trans data-src="/* Error checking of pName left out */">/* Error checking of pName left out */</trans> <trans data-src="pModule = PyImport_Import(pName);">pModule = PyImport_Import(pName);</trans> <trans data-src="```">```</trans> <trans data-src="在对解释器进行初始化之后，使用[PyImport_Import()](">After initializing the interpreter, use [PyImport_Import()]（</trans> <trans data-src="https://docs.python.org/2/c-api/import.html#c.PyImport_Import">https://docs.python.org/2/c-api/import.html#c.PyImport_Import</trans> <trans data-src=")加载脚本。">）Load script.</trans> <trans data-src="这个例程需要一个python字符串作为它的参数，该参数是通过使用[ PyString_FromString()](">This routine needs a python string as its parameter, which is determined by using [PyString_FromString()]（</trans> <trans data-src="https://docs.python.org/2/c-api/string.html#c.PyString_FromString">https://docs.python.org/2/c-api/string.html#c.PyString_FromString</trans> <trans data-src=")来构造的。">）To construct.</trans> <trans data-src="```">```</trans> <trans data-src="pFunc = PyObject_GetAttrString(pModule, argv[2]);">pFunc = PyObject_GetAttrString(pModule, argv[2]);</trans> <trans data-src="/* pFunc is a new reference */">/* pFunc is a new reference */</trans> <trans data-src="if (pFunc && PyCallable_Check(pFunc)) {">if (pFunc && PyCallable_Check(pFunc)) {</trans> <trans data-src="...">...</trans> <trans data-src="}">}</trans> <trans data-src="Py_XDECREF(pFunc);">Py_XDECREF(pFunc);</trans> <trans data-src="```">```</trans> <trans data-src="一旦脚本被载入，我们使用[ PyObject_GetAttrString()](">Once the script is loaded, we use [PyObject_GetAttrString()]（</trans> <trans data-src="https://docs.python.org/2/c-api/object.html#c.PyObject_GetAttrString">https://docs.python.org/2/c-api/object.html#c.PyObject_GetAttrString</trans> <trans data-src=")检索我们需要查找的名称。">）Retrieve the name we need to find.</trans> <trans data-src="如果名称存在，并且其返回对象是可调用的，我们就大可以假设它是一个函数。">If the name exists and its return object is callable, we can assume that it is a function.</trans> <trans data-src="然后，程序将按照常规构造一个元组参数，并使用以下命令调用python函数：">Then, the program will construct a tuple parameter as usual, and use the following command to call the python function:</trans> <trans data-src="```">```</trans> <trans data-src="pValue = PyObject_CallObject(pFunc, pArgs);">pValue = PyObject_CallObject(pFunc, pArgs);</trans> <trans data-src="```">```</trans> <trans data-src="函数返回时，***pValue*** 要么为空，要么包含对函数返回值得应用。">When the function returns, * * * pValue * * * is either empty or contains the application of the function return value.</trans> <trans data-src="在检查完值之后，一定要记得释放引用。">After checking the value, be sure to release the reference.</trans> <trans data-src="## 5.4 嵌入式python扩展">##5.4 Embedded python extension</trans> <trans data-src="到目前为止，嵌入式python解释器无法访问应用程序本身的功能，而python API通过扩展嵌入式解释器实现了这一点。">So far, the embedded python interpreter cannot access the functions of the application itself, and the python API implements this by extending the embedded interpreter.</trans> <trans data-src="也就是说，使用应用程序提供的例程扩展嵌入式解释器。">That is, the embedded interpreter is extended using routines provided by the application.</trans> <trans data-src="虽然这听起来很复杂，但其实也没那么糟。">Although it sounds complicated, it's not that bad.</trans> <trans data-src="先暂时忘记应用程序启动python解释器。">Forget to start the python interpreter for the application.</trans> <trans data-src="相反，我们将应用程序视为一组子例程，并编写一些使python能够访问这些例程的粘合代码，就像编写一般的python扩展一样。">Instead, we treat the application as a set of subroutines, and write some glue code that enables python to access these routines, just like writing a general python extension.</trans> <trans data-src="例如：">For example:</trans> <trans data-src="```">```</trans> <trans data-src="static int numargs=0;">static int numargs=0;</trans> <trans data-src="/* Return the number of arguments of the application command line */">/* Return the number of arguments of the application command line */</trans> <trans data-src="static PyObject* emb_numargs(PyObject *self, PyObject *args)">static PyObject* emb_numargs(PyObject *self, PyObject *args)</trans> <trans data-src="{">{</trans> <trans data-src="if(!PyArg_ParseTuple(args, ":numargs"))">if(!PyArg_ParseTuple(args, ":numargs"))</trans> <trans data-src="return NULL;">return NULL;</trans> <trans data-src="return Py_BuildValue("i", numargs);">return Py_BuildValue("i", numargs);</trans> <trans data-src="}">}</trans> <trans data-src="static PyMethodDef EmbMethods[] = {">static PyMethodDef EmbMethods[] = {</trans> <trans data-src="{"numargs", emb_numargs, METH_VARARGS,">{"numargs", emb_numargs, METH_VARARGS,</trans> <trans data-src=""Return the number of arguments received by the process."},">"Return the number of arguments received by the process."},</trans> <trans data-src="{NULL, NULL, 0, NULL}">{NULL, NULL, 0, NULL}</trans> <trans data-src="};">};</trans> <trans data-src="```">```</trans> <trans data-src="在main（）函数之上插入上述代码。">Insert the above code above the main() function.</trans> <trans data-src="另外，在[Py_Initialize（）](">In addition, in [Py_Initialize()]（</trans> <trans data-src="https://docs.python.org/2/c-api/init.html#c.Py_Initialize">https://docs.python.org/2/c-api/init.html#c.Py_Initialize</trans> <trans data-src=") 之后直接插入以下两条语句：">）Then insert the following two statements directly:</trans> <trans data-src="```">```</trans> <trans data-src="numargs = argc;">numargs = argc;</trans> <trans data-src="Py_InitModule("emb", EmbMethods);">Py_InitModule("emb", EmbMethods);</trans> <trans data-src="```">```</trans> <trans data-src="这两行代码初始化了***numargs***变量，并使嵌入式python解释器能够访问***emb.numargs()***函数。">These two lines of code initialize the * * * numargs * * * variable and enable the embedded python interpreter to access the * * * emb. numargs() * * * function.</trans> <trans data-src="有了这些扩展，python脚本可以做以下事情：">With these extensions, python scripts can do the following:</trans> <trans data-src="```">```</trans> <trans data-src="import emb">import emb</trans> <trans data-src="print "Number of arguments", emb.numargs()">print "Number of arguments", emb.numargs()</trans> <trans data-src="```">```</trans> <trans data-src="在实际应用程序中，这些方法将向python公开应用程序的API。">In actual applications, these methods will expose the API of the application to python.</trans> <trans data-src="## 5.5 在C++中嵌入python">##5.5 Embedding python in C++</trans> <trans data-src="在C++程序中嵌入python也是可能地；">It is also possible to embed python in C++programs;</trans> <trans data-src="具体如何实现取决于所使用的C++系统的细节；">How to implement it depends on the details of the C++system used;</trans> <trans data-src="通常，我们需要使用C++编写主程序，并使用C++编译器来编译和链接我们的程序，而不需要使用C++来重新编译python本身。">Usually, we need to use C++to write the main program, and use the C++compiler to compile and link our program, instead of using C++to recompile python itself.</trans> <trans data-src="## 5.6 在类Unix系统下的编译和链接">##5.6 Compilation and linking in Unix like systems</trans> <trans data-src="为了将python解释器嵌入到应用程序中，找到传递给编译器（和链接器）的正确标志并不是件小事，这是因为python需要加载作为C动态扩展来实现（.so文件）的库模块，并将其链接到应用程序中。">In order to embed the python interpreter into the application, it is not a trivial matter to find the correct flag passed to the compiler (and linker). This is because python needs to load the library module that is implemented as a dynamic extension of C (. so file) and link it to the application.</trans> <trans data-src="为了找到所需的编译器和链接器标志，可以执行在安装过程中生成的***pythonX.">To find the required compiler and linker flags, you can execute the * * * pythonX generated during installation</trans> <trans data-src="Y-config*** 脚本（也可以使用***python-config***脚本）。">Y-config * * * script (* * * python config * * * script can also be used).</trans> <trans data-src="这个脚本有好几个选项，下面的选项可能对你有用：">This script has several options. The following options may be useful to you:</trans> <trans data-src="当你编译时***pythonX.">When you compile * * * pythonX</trans> <trans data-src="Y-config --cflags***将给你推荐的标志：">Y-config -- cflags * * * will recommend you the following marks:</trans> <trans data-src="```">```</trans> <trans data-src="$ /opt/bin/python2.7-config --cflags">$ /opt/bin/python2.7-config --cflags</trans> <trans data-src="-I/opt/include/python2.7 -fno-strict-aliasing -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes">-I/opt/include/python2.7 -fno-strict-aliasing -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes</trans> <trans data-src="```">```</trans> <trans data-src="当你链接时，***pythonX.">When you link, * * * pythonX</trans> <trans data-src="Y-config --ldflags***将给你推荐的标志：">Y-config -- ldflags * * * will recommend you:</trans> <trans data-src="```">```</trans> <trans data-src="$ /opt/bin/python2.7-config --ldflags">$ /opt/bin/python2.7-config --ldflags</trans> <trans data-src="-L/opt/lib/python2.7/config -lpthread -ldl -lutil -lm -lpython2.7 -Xlinker -export-dynamic">-L/opt/lib/python2.7/config -lpthread -ldl -lutil -lm -lpython2.7 -Xlinker -export-dynamic</trans> <trans data-src="```">```</trans> <trans data-src="注意：为了避免在几个python安装之间（特别是在系统python和你自己编译的python）产生混淆，建议使用***pythonX.">Note: To avoid confusion between several python installations (especially between system python and your own compiled python), it is recommended to use * * * pythonX</trans> <trans data-src="Y-config*** 的绝对路径，如上例所示。">The absolute path of Y-config * * *, as shown in the above example.</trans> <trans data-src="如果此步骤对你来说并不适用（它并不能保证适用所有的类unix平台；但我们同样欢迎bug报告），你必须阅读系统关于动态链接和/或检查python的Makefiles（使用sysconfig.get_makefile_filename()查找其位置）和编译选项。">If this step is not applicable to you (it does not guarantee that it is applicable to all unix like platforms; however, we also welcome bug reports), you must read the system's Makefiles (use sysconfig. get_makefile_filename() to find its location) and compilation options on dynamic linking and/or checking python.</trans> <trans data-src="在这种情况下，***sysconfig***模块是一个有用的工具，可以通过编程方式提取你想要组合在一起的配置值。">In this case, the * * * sysconfig * * * module is a useful tool that can programmatically extract the configuration values you want to combine.</trans> <trans data-src="例如:">For example:</trans> <trans data-src="```">```</trans> <trans data-src=">>> import sysconfig">>>> import sysconfig</trans> <trans data-src=">>> sysconfig.get_config_var('LIBS')">>>> sysconfig.get_config_var('LIBS')</trans> <trans data-src="'-lpthread -ldl  -lutil'">'-lpthread -ldl  -lutil'</trans> <trans data-src=">>> sysconfig.get_config_var('LINKFORSHARED')">>>> sysconfig.get_config_var('LINKFORSHARED')</trans> <trans data-src="'-Xlinker -export-dynamic'">'-Xlinker -export-dynamic'</trans> <trans data-src="```">```</trans> <trans data-src="--------------------------------------------------------------------------------------------------------这是一条严肃的分割线">--------------------------------------------------------------------------------------------------------This is a serious dividing line</trans> <trans data-src="翻译完也没能搞明白C++是怎么调用Python的，确实是条废柴没错了，哭！">I can't understand how C++calls Python after translation. It's really a piece of rubbish. Cry!</trans>

embedding python

This article was created by bean powder Knowledge Sharing Attribution 4.0 International license agreement. Except for the reprint/source, all articles on this website are original or translated by this website. Please sign your name before reprinting.