调整 yaml.Dumper 以导出数据类及其文档字符串

Question

我想将数据类对象包括它们的文档字符串（如上面的#注释描述）导出到yaml。

到目前为止，我发现我需要一个定制的转储器或调整当前的转储器。

from dataclasses import dataclass

@dataclass
class foo:
"""This would be a nice to have docstring but can be joined"""

   bar: int = 1
   """This is bar"""

# Export

import yaml
class SomeDumper(yaml.Dumper):
   """What is needed here?"""

yaml.dump(foo, "foo.yaml", dumper=SomeDumper)

预期输出：

# -----
# This would be a nice to have docstring but can be joined
# -----

# This is bar
bar : 1

我需要如何调整转储器以保持如上所示的文档字符串？

为了简单起见，假设我有文档字符串，并且我只关心如何调整转储器。

Answer 1

首先，从类中提取文档非常容易。

然后，关于参数注释或参数文档字符串，您需要找到为您执行此操作的良好库（此类库很可能是在代码检查功能之上创建的）。或者更容易的是从使用键入“Annotated”来注释参数开始。这个解决方案对您有用吗？

查看下面的示例，了解如何提取提示信息。

from dataclasses import dataclass
from typing import get_type_hints, Annotated, get_args


@dataclass
class Foo:
    """This would be a nice to have docstring but can be joined"""

    bar: Annotated[int, "This is bar"] = 1
    msg: Annotated[int, "This is msg"] = 1

print (foo.__doc__)

for arg, _type in Foo.__annotations__.items():
    arg_comment = get_args(_type)[1]
    print(f"{arg} -- {arg_comment}")