如何获得在计算机上安装的Python模块列表?
当前回答
此解决方案主要基于importlib和pkgutil模块，并与CPython 3.4和CPython 3.5一起工作，但不支持CPython 2。
解释
sys。Builtin_module_names -命名所有内置模块(看我的答案在这里) iter_modules() -返回所有可用模块的信息 Importlib.util.find_spec() -返回关于导入模块的信息，如果存在的话 BuiltinImporter -内置模块(文档)的导入器 SourceFileLoader -标准Python模块的导入器(默认扩展名为*.py) ExtensionFileLoader——作为共享库的模块导入器(用C或c++编写)
完整代码
import sys
import os
import shutil
import pkgutil
import importlib
import collections
if sys.version_info.major == 2:
raise NotImplementedError('CPython 2 is not supported yet')
def main():
# name this file (module)
this_module_name = os.path.basename(__file__).rsplit('.')[0]
# dict for loaders with their modules
loaders = collections.OrderedDict()
# names`s of build-in modules
for module_name in sys.builtin_module_names:
# find an information about a module by name
module = importlib.util.find_spec(module_name)
# add a key about a loader in the dict, if not exists yet
if module.loader not in loaders:
loaders[module.loader] = []
# add a name and a location about imported module in the dict
loaders[module.loader].append((module.name, module.origin))
# all available non-build-in modules
for module_name in pkgutil.iter_modules():
# ignore this module
if this_module_name == module_name[1]:
continue
# find an information about a module by name
module = importlib.util.find_spec(module_name[1])
# add a key about a loader in the dict, if not exists yet
loader = type(module.loader)
if loader not in loaders:
loaders[loader] = []
# add a name and a location about imported module in the dict
loaders[loader].append((module.name, module.origin))
# pretty print
line = '-' * shutil.get_terminal_size().columns
for loader, modules in loaders.items():
print('{0}\n{1}: {2}\n{0}'.format(line, len(modules), loader))
for module in modules:
print('{0:30} | {1}'.format(module[0], module[1]))
if __name__ == '__main__':
main()
使用
对于CPython3.5(截断)
$ python3.5 python_modules_info.py
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
30: <class '_frozen_importlib.BuiltinImporter'>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
_ast | built-in
_codecs | built-in
_collections | built-in
_functools | built-in
_imp | None
_io | built-in
_locale | built-in
_operator | built-in
_signal | built-in
_sre | built-in
_stat | built-in
_string | built-in
_symtable | built-in
_thread | built-in
(****************************truncated*******************************)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
227: <class '_frozen_importlib_external.SourceFileLoader'>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
__future__ | /usr/local/lib/python3.5/__future__.py
_bootlocale | /usr/local/lib/python3.5/_bootlocale.py
_collections_abc | /usr/local/lib/python3.5/_collections_abc.py
_compat_pickle | /usr/local/lib/python3.5/_compat_pickle.py
_compression | /usr/local/lib/python3.5/_compression.py
_dummy_thread | /usr/local/lib/python3.5/_dummy_thread.py
_markupbase | /usr/local/lib/python3.5/_markupbase.py
_osx_support | /usr/local/lib/python3.5/_osx_support.py
_pydecimal | /usr/local/lib/python3.5/_pydecimal.py
_pyio | /usr/local/lib/python3.5/_pyio.py
_sitebuiltins | /usr/local/lib/python3.5/_sitebuiltins.py
(****************************truncated*******************************)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
64: <class '_frozen_importlib_external.ExtensionFileLoader'>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
_bisect | /usr/local/lib/python3.5/lib-dynload/_bisect.cpython-35m-x86_64-linux-gnu.so
_bz2 | /usr/local/lib/python3.5/lib-dynload/_bz2.cpython-35m-x86_64-linux-gnu.so
_codecs_cn | /usr/local/lib/python3.5/lib-dynload/_codecs_cn.cpython-35m-x86_64-linux-gnu.so
_codecs_hk | /usr/local/lib/python3.5/lib-dynload/_codecs_hk.cpython-35m-x86_64-linux-gnu.so
_codecs_iso2022 | /usr/local/lib/python3.5/lib-dynload/_codecs_iso2022.cpython-35m-x86_64-linux-gnu.so
(****************************truncated*******************************)
对于CPython3.4(截断)
$ python3.4 python_modules_info.py
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
54: <class '_frozen_importlib.BuiltinImporter'>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
_ast | built-in
_bisect | built-in
_codecs | built-in
_collections | built-in
_datetime | built-in
_elementtree | built-in
_functools | built-in
_heapq | built-in
_imp | None
_io | built-in
_locale | built-in
_md5 | built-in
_operator | built-in
_pickle | built-in
_posixsubprocess | built-in
_random | built-in
(****************************truncated*******************************)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
246: <class '_frozen_importlib.SourceFileLoader'>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
__future__ | /usr/lib/python3.4/__future__.py
_bootlocale | /usr/lib/python3.4/_bootlocale.py
_collections_abc | /usr/lib/python3.4/_collections_abc.py
_compat_pickle | /usr/lib/python3.4/_compat_pickle.py
_dummy_thread | /usr/lib/python3.4/_dummy_thread.py
_markupbase | /usr/lib/python3.4/_markupbase.py
_osx_support | /usr/lib/python3.4/_osx_support.py
_pyio | /usr/lib/python3.4/_pyio.py
(****************************truncated*******************************)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
44: <class '_frozen_importlib.ExtensionFileLoader'>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
_bz2 | /usr/lib/python3.4/lib-dynload/_bz2.cpython-34m-x86_64-linux-gnu.so
_codecs_cn | /usr/lib/python3.4/lib-dynload/_codecs_cn.cpython-34m-x86_64-linux-gnu.so
_codecs_hk | /usr/lib/python3.4/lib-dynload/_codecs_hk.cpython-34m-x86_64-linux-gnu.so
_codecs_iso2022 | /usr/lib/python3.4/lib-dynload/_codecs_iso2022.cpython-34m-x86_64-linux-gnu.so
_codecs_jp | /usr/lib/python3.4/lib-dynload/_codecs_jp.cpython-34m-x86_64-linux-gnu.so
_codecs_kr | /usr/lib/python3.4/lib-dynload/_codecs_kr.cpython-34m-x86_64-linux-gnu.so
_codecs_tw | /usr/lib/python3.4/lib-dynload/_codecs_tw.cpython-34m-x86_64-linux-gnu.so
_crypt | /usr/lib/python3.4/lib-dynload/_crypt.cpython-34m-x86_64-linux-gnu.so
(****************************truncated*******************************)
其他回答
在正常情况下使用即可
pydoc modules
我在OS x上遇到了一个自定义安装的python 2.7。它需要X11来列出已安装的模块(使用help和pydoc)。
为了能够在不安装X11的情况下列出所有模块，我运行pydoc作为http-server，即:
pydoc -p 12345
然后，可以将Safari定向到http://localhost:12345/以查看所有模块。
对于任何想知道如何从Python程序调用pip list的人，您可以使用以下方法:
import pip
pip.main(['list]) # this will print all the packages
我比较了五种方法来检索已安装的“模块”，所有这些都是我在这个线程中看到的
| iter_modules | help("modules") | builtin_module_names | pip list | working_set | |
|---|---|---|---|---|---|
| Includes distributions | ❌ | ❌ | ❌ | ✔️ | ✔️ |
| Includes modules (No built-in) | ✔️ | ✔️ | ❌ | ❌ | ❌ |
| Includes built-in modules | ❌ | ✔️ | ✔️ | ❌ | ❌ |
| Includes frozen | ✔️ | ✔️ | ❌ | ❌ | ❌ |
| Includes venv | ✔️ | ✔️ | ❌ | ✔️ | ✔️ |
| Includes global | ✔️ | ✔️ | ❌ | ✔️ | ✔️ |
| Includes editable installs | ✔️ | ✔️ | ❌ | ✔️ | ✔️ |
| Includes PyCharm helpers | ✔️ | ❌ | ❌ | ❌ | ❌ |
| Lowers capital letters | ❌ | ❌ | ❌ | ❌ | ✔️ |
| Time taken (665 modules total) | 53.7 msec | 1.03 sec | 577 nsec | 284 msec | 36.2 usec |
总结
PIP list和working_set是针对发行版的，而不是模块的。 Iter_modules和help("modules")非常相似，最大的区别是Iter_modules不包含内置模块。 PIP list和working_set非常相似，唯一不同的是working_set降低了所有大写字母。 内置模块只包含在help("modules")和builtin_module_names中。
相关的说明
发行版、包和模块通常具有相同的名称，因此很容易将它们误认为另一个。 importlib.util。Find_spec用于模块，区分大小写。 sys。Modules只列出导入的模块。
分布
我说分发而不是打包，因为我认为这样会减少误解。一个发行版/包可以包含多个包/模块。
已安装的发行版并不总是可以通过相同的名称导入。例如pip install Pillow是进口PIL。有时，一个发行版甚至可以导入多个模块。
方法(每列按顺序排列)
iter_modules
import pkgutil
{module.name for module in pkgutil.iter_modules()}
帮助("modules")(仅在终端中打印)
help("modules")
builtin_module_names
import sys
set(sys.builtin_module_names)
pip列表(仅在终端打印)
终端PIP清单
working_set
import pkg_resources
{pkg.key for pkg in pkg_resources.working_set}
结论
对于终端，我建议使用help("modules")或python -c "help("modules") "。 在这个帖子中投票最多的答案 对于编程方式，我建议使用iter_modules + builtin_module_names。 这个答案在这个帖子里 然而，这是非常复杂的信息，请看我下面的最小示例。
import sys
import pkgutil
def get_installed_modules_names():
iter_modules = {module.name for module in pkgutil.iter_modules()}
builtin = sys.builtin_module_names
return set.union(iter_modules, builtin)
如果安装了anaconda python发行版，也可以使用
$conda list
除了上述解决方案。
