帕克

Paker 是用于从字典和 JSON 格式的文档中导入 Python 包/模块的模块。它的灵感来自httpimporter。

重要提示：由于 v0.6.0支持直接从内存paker导入.pyd和模块。.dll这是通过使用_memimporterpy2exe项目实现的。在 Linux 上导入.so文件仍然需要将它们写入磁盘。

安装

来自 PyPI

pip install paker -U

从源头

git clone https://github.com/desty2k/paker.git
cd paker
pip install .

用法

让我们使用以下目录结构paker转储模块：print_lib

print_lib/
├── __init__.py
├── capitalize.py
└── decapitalize.py

capitalized.py包含Print函数和decapitalized.py包含pRINT函数。__init__.py文件为空。

转储到 JSON

要将库转储到文件使用paker dump命令：

paker dump print_lib --indent 4

Paker 将print_lib.json在当前目录中创建新文件。

{
    "print_lib": {
        "type": "package",
        "extension": "py",
        "code": "",
        "modules": {
            "capitalize": {
                "type": "module",
                "extension": "py",
                "code": "def Print(text: str, *args, **kwargs):\n    print(text.capitalize(), *args, **kwargs)\n"
            },
            "decapitalize": {
                "type": "module",
                "extension": "py",
                "code": "def pRINT(text: str, *args, **kwargs):\n    print(text[0].lower() + text[1:].upper(), *args, **kwargs)\n"
            }
        }
    }
}

列出模块

要列出 JSON 使用paker list命令中的所有模块：

paker list print_lib.json

输出：

P print_lib
M print_lib.capitalize
M print_lib.decapitalize

从 JSON 重新创建包

要从 JSON 使用paker load命令重新创建包。

paker load print_lib.json

print_lib包将在当前目录中创建。设置自定义输出路径使用--output选项。

从导入.json

注意：从原始源代码导入比从 JSON 导入具有更高的优先级。请记住在使用paker.

import paker
import logging

logging.basicConfig(level=logging.NOTSET)

if __name__ == '__main__':
    with open("print_lib.json", "r") as f:
        with paker.load(f) as loader:
            # print_lib will be available only in this context
            from print_lib.capitalize import Print
            from print_lib.decapitalize import pRINT
            Print("hello world!")
            pRINT("hello world!")

此脚本产生以下输出：

DEBUG:jsonimporter:searching for print_lib
INFO:jsonimporter:print_lib has been imported successfully
DEBUG:jsonimporter:searching for print_lib.capitalize
INFO:jsonimporter:print_lib.capitalize has been imported successfully
DEBUG:jsonimporter:searching for print_lib.decapitalize
INFO:jsonimporter:print_lib.decapitalize has been imported successfully

Hello world!
hELLO WORLD!

INFO:jsonimporter:unloaded all modules

从内存导入

您还可以直接从内存中导入 Python 模块。库不仅可以从 Python 对象加载，dict还可以从str和对象加载。bytesbytearray

import paker
import logging

# paker.loads accepts dict, str, bytes and bytearray objects
POW = {"pow": {"type": "module", "extension": "py", "code": "pow = lambda x, y: x**y"}}
SQR = '{"sqr": {"type": "module", "extension": "py", "code": "from pow import pow\\nsqr = lambda x: pow(x, 2)"}}'
TRI = b'{"tri": {"type": "module", "extension": "py", "code": "from pow import pow\\ntri = lambda x: pow(x, 3)"}}'

logging.basicConfig(level=logging.NOTSET)

if __name__ == '__main__':
    # you can use nested loaders
    with paker.loads(POW) as pow_loader:
        # pow will be available only in this context
        with paker.loads(SQR) as sqr_loader:
            # sqr will be available only in this context
            from sqr import sqr
            assert sqr(2), 4
            assert sqr(5), 25
            print("6**2 is {}".format(sqr(6)))

        with paker.loads(TRI) as tri_loader:
            # tri will be available only in this context
            from tri import tri
            assert tri(2), 8
            assert tri(5), 125
            print("6**3 is {}".format(tri(6)))
        print("It works!")

此脚本产生以下输出：

DEBUG:jsonimporter:searching for sqr
INFO:jsonimporter:sqr has been added successfully
DEBUG:jsonimporter:searching for sqr
DEBUG:jsonimporter:searching for pow
INFO:jsonimporter:pow has been imported successfully
INFO:jsonimporter:sqr has been imported successfully

6**2 is 36

INFO:jsonimporter:sqr has been unloaded successfully
DEBUG:jsonimporter:searching for tri
INFO:jsonimporter:tri has been added successfully
DEBUG:jsonimporter:searching for tri
INFO:jsonimporter:tri has been imported successfully

6**3 is 216

INFO:jsonimporter:tri has been unloaded successfully

It works!

INFO:jsonimporter:unloaded all modules

这个怎么运作

当导入模块或包时，Python 会遍历导入器并在每个对象上sys.meta_path调用方法。find_module如果importer返回自己，则表示模块可以被导入，None表示importer没有找到搜索到的包。如果任何导入器已确认能够导入模块，Python 会在其上执行另一种方法 - load_module. Paker 实现了自己的导入jsonimporter器，它不是在目录中搜索模块，而是在 Python 字典中查找它们

为了将模块或包转储到 JSON 文档，Paker 递归地迭代模块并使用每个模块和子模块的代码和类型创建字典（如果对象是包）。