Question

我正在尝试将Python模块反汇编为字节码。

我必须静态或动态导入Python模块才能进行反汇编或检查吗？如果没有，那么（pythonic，便携式）方法是什么？

我想：

在运行时将可用的Python模块的二进制数据加载到内存中：
1. 没有它在sys.modules中显示为可用模块。
2. 我不想执行任何模块的__init__代码，也不想将其添加到任何命名空间。
3. 加载模块应该没有其他副作用。至于在翻译方面，它应该只是一大堆要检查的数据。
拆卸或以其他方式检查模块的类别，功能或数据。
根据需要卸载模块。

我已经搜索过，我看到了许多动态模块导入方法（它具有执行模块__init__代码或其他内联代码的副作用，并且插入到sys.modules中）。但我宁愿不处理这些副作用。

这可能吗？如果是这样，哪种方法最便携/ Pythonic？

Answer 1

我对此进行了一些调查，一个可能的解决方案是使用pyclbr模块。检查它会查看有关类和函数的基本信息，将其加载到字典中以便于访问。这是一个示例运行：

>>> import pyclbr
>>> import sys
>>> info = pyclbr.readmodule_ex('inspect')
>>> info
{'formatargvalues': <pyclbr.Function object at 0x5083e28e50>, 'walktree': <pyclbr.Function object at 0x5083e28b50>, 'getinnerframes': <pyclbr.Function object at 0x5083e29050>, 'indentsize': <pyclbr.Function object at 0x5083e28710>, 'getmodulename': <pyclbr.Function object at 0x5083e28850>, 'formatannotation': <pyclbr.Function object at 0x5083e28d50>, 'ismemberdescriptor': <pyclbr.Function object at 0x5083e283d0>, 'iscode': <pyclbr.Function object at 0x5083e28550>, 'getsource': <pyclbr.Function object at 0x5083e28b10>, 'formatargspec': <pyclbr.Function object at 0x5083e28dd0>, 'getabsfile': <pyclbr.Function object at 0x5083e288d0>, 'getsourcelines': <pyclbr.Function object at 0x5083e28ad0>, '_getfullargs': <pyclbr.Function object at 0x5083e28c10>, 'isabstract': <pyclbr.Function object at 0x5083e28610>, 'isbuiltin': <pyclbr.Function object at 0x5083e28590>, 'getlineno': <pyclbr.Function object at 0x5083e28f10>, 'getcomments': <pyclbr.Function object at 0x5083e28990>, 'getgeneratorstate': <pyclbr.Function object at 0x5083e293d0>, 'getattr_static': <pyclbr.Function object at 0x5083e29390>, 'getframeinfo': <pyclbr.Function object at 0x5083e28ed0>, 'isgenerator': <pyclbr.Function object at 0x5083e28490>, '_static_getmro': <pyclbr.Function object at 0x5083e29190>, 'isframe': <pyclbr.Function object at 0x5083e28510>, 'getouterframes': <pyclbr.Function object at 0x5083e28f90>, 'getclasstree': <pyclbr.Function object at 0x5083e28b90>, 'getfile': <pyclbr.Function object at 0x5083e287d0>, '_shadowed_dict': <pyclbr.Function object at 0x5083e29310>, 'getargvalues': <pyclbr.Function object at 0x5083e28d10>, 'getmembers': <pyclbr.Function object at 0x5083e28650>, 'BlockFinder': <pyclbr.Class object at 0x5083e28a10>, 'isfunction': <pyclbr.Function object at 0x5083e28390>, 'getargspec': <pyclbr.Function object at 0x5083e28c50>, 'currentframe': <pyclbr.Function object at 0x5083e29090>, 'namedtuple': <pyclbr.Function object at 0x5083e1b150>, 'getmoduleinfo': <pyclbr.Function object at 0x5083e28810>, 'trace': <pyclbr.Function object at 0x5083e29110>, 'isclass': <pyclbr.Function object at 0x5083db8950>, '_is_type': <pyclbr.Function object at 0x5083e29290>, 'getcallargs': <pyclbr.Function object at 0x5083e28e90>, 'ismethoddescriptor': <pyclbr.Function object at 0x5083e28310>, 'isgeneratorfunction': <pyclbr.Function object at 0x5083e28450>, 'isroutine': <pyclbr.Function object at 0x5083e285d0>, 'getfullargspec': <pyclbr.Function object at 0x5083e28cd0>, 'getmro': <pyclbr.Function object at 0x5083e286d0>, 'getargs': <pyclbr.Function object at 0x5083e28bd0>, 'stack': <pyclbr.Function object at 0x5083e290d0>, 'getdoc': <pyclbr.Function object at 0x5083e28750>, 'findsource': <pyclbr.Function object at 0x5083e28950>, 'cleandoc': <pyclbr.Function object at 0x5083e28790>, '_check_class': <pyclbr.Function object at 0x5083e29250>, '_check_instance': <pyclbr.Function object at 0x5083e29210>, 'classify_class_attrs': <pyclbr.Function object at 0x5083e28690>, 'ismodule': <pyclbr.Function object at 0x5083db8910>, 'EndOfBlock': <pyclbr.Class object at 0x5083e289d0>, 'isdatadescriptor': <pyclbr.Function object at 0x5083e28350>, 'getmodule': <pyclbr.Function object at 0x5083e28910>, 'formatannotationrelativeto': <pyclbr.Function object at 0x5083e28d90>, 'getsourcefile': <pyclbr.Function object at 0x5083e28890>, 'ismethod': <pyclbr.Function object at 0x5083e282d0>, 'isgetsetdescriptor': <pyclbr.Function object at 0x5083e28410>, 'istraceback': <pyclbr.Function object at 0x5083e284d0>, 'getblock': <pyclbr.Function object at 0x5083e28a50>}
>>> sys.modules['inspect']
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
KeyError: 'inspect'

任何更高级的东西，您都必须开始考虑通过ast module访问抽象语法树。

Answer 2

我为Python 2.7找到了uncompyle2，它包含load a source file and compile it to bytecodes和load a module and compile it to bytecodes的函数而不导入模块。

所以至少看起来它是可行的，但可能涉及在源上调用compile()，或者，如果使用pyc文件，可能不可移植（uncompyle2仅支持带有pyc文件的Python 2.7。）

如何动态加载和卸载Python模块，反汇编和检查它，但不执行init代码或将其添加到sys.modules？

2 个答案: