The English introduction is placed below the Chinese version.

一个支持并行反编译的Python反编译工具，多进程并行反编译pyc文件中的每个类和函数。
同时支持uncompyle6和decompyle3库。
由于uncompyle6和decompyle3库是不支持并行的，反编译超过20KB的pyc文件会很慢，
而且一旦单个函数反编译失败，整个反编译就会停止。
而这个程序的目的旨在于增强反编译过程的速度和健壮性。

1.命令行

python decompiler.py <要编译的.pyc文件名> <--uncompyle6或--decompyle3开关(可选)>

如果不指定开关，则默认使用uncompyle6库。
运行结束后，会自动在pyc文件所在的目录生成一个同名的.py文件，作为反编译结果，如python decompiler.py test.pyc会生成test.py这个文件。

2.程序的实现原理

pyc文件本身是一个Code(字节码)对象，而这个对象的co_const属性的元组内部又包含了函数和类的Code对象。
程序首先预处理pyc文件，将各个函数和类的字节码分离出来，并将函数和类以简单的return语句替代。
然后将预处理之后的pyc文件反编译成.py源文件，此时的源文件只有大致框架，就像下面这样：

import sys,os

class Test:
    return 123 # 以简单的return语句暂时替代 (虽然并不符合语法)

def test():
    return 125

if __name__=="__main__":test()

接着再使用多进程，并行反编译各个函数和类的字节码文件，最后将结果合并，生成最终的代码。
此外，如果一个函数、类反编译失败，不会影响其他函数、类的反编译。

3.依赖的库

这个反编译工具依赖于pyobject库，尤其是pyobject.code_这个子模块中的Code类。
pyobject库可用pip install pyobject命令安装。

4.兼容性

程序支持所有的Python 3版本，由于不依赖于特定版本的字节码。

5.TODO

由于单个类的方法可能会很多，自己未来会考虑使用递归，将类中的各个方法拆分再并行反编译，以进一步提升速度。

A Python decompiler tool supporting parallel decompilation, designed to decompile each class and function within pyc files using multiple processes. It supports both uncompyle6 and decompyle3 libraries. The uncompyle6 and decompyle3 libraries themselves do not support parallel processing, which can result in slow decompilation for pyc files larger than 20KB and can halt the entire decompilation process if a single function fails to decompile. This program aims to enhance the speed and robustness of the decompilation process.

1. Command Line Usage

python decompiler.py <filename.pyc> [--uncompyle6 | --decompyle3]

If no switch is specified, the uncompyle6 library is used by default. Upon completion, a .py file with the same name as the pyc file will be generated in the pyc file's directory as the decompilation result. For example, running python decompiler.py test.pyc will produce test.py.

2. Implementation Principle

A pyc file is essentially a Code (bytecode) object, and the tuple within its co_consts attribute contains the Code objects for functions and classes. The program first preprocesses the pyc file, separating the bytecode for each function and class, and replaces them with simple return statements. The preprocessed pyc file is then decompiled into a .py source file, which initially contains only a rough framework, as shown below:

import sys
import os

class Test:
    return 123  # Temporarily replaced with a simple return statement (although it is not syntactically valid)

def test():
    return 125

if __name__ == "__main__":
    test()

Next, multiple processes are used to parallelly decompile the bytecode files of each function and class. The results are then merged to generate the final code. Additionally, if decompilation of a single function or class fails, it will not affect the decompilation of other functions or classes.

3. Dependencies

This decompiler tool depends on the pyobject library, particularly the Code class from the pyobject.code_ submodule. The pyobject library can be installed with the command pip install pyobject.

4.Compatibility

The program supports all versions of Python 3 since it doesn't rely on a specific version of bytecode.

5. TODO

Since a class may have many methods, future considerations include using recursion to split and parallelly decompile individual methods within a class to further improve speed.