How can I parse a numpydoc docstring and access co

问题:

I'd like to parse a numpydoc docstring and access each component programatically.

For example:

def foobar(a, b):
   '''Something something

   Parameters
   ----------
   a : int, default: 5
        Does something cool
   b : str
        Wow
'''

What I'd like to do is:

parsed = magic_parser(foobar)
parsed.text  # Something something
parsed.a.text  # Does something cool
parsed.a.type  # int
parsed.a.default  # 5

I've been searching around and found things like numpydoc and napoleon but I haven't found any good leads for how to use them in my own program. I'd appreciate any help.

回答1:

You can use NumpyDocString from numpydoc to parse docstrings into a Python-friendly structure.

This is an example of how to use it:

from numpydoc.docscrape import NumpyDocString


class Photo():
    """
    Array with associated photographic information.


    Parameters
    ----------
    x : type
        Description of parameter `x`.
    y
        Description of parameter `y` (with type not specified)

    Attributes
    ----------
    exposure : float
        Exposure in seconds.

    Methods
    -------
    colorspace(c='rgb')
        Represent the photo in the given colorspace.
    gamma(n=1.0)
        Change the photo's gamma exposure.

    """

    def __init__(x, y):
        print("Snap!")

doc = NumpyDocString(Photo.__doc__)
print(doc["Summary"])
print(doc["Parameters"])
print(doc["Attributes"])
print(doc["Methods"])

However, this won't work with the example you gave (nor a lot of the code I want to run this on) for reasons I don't understand. Instead, you need to use the specific FunctionDoc or ClassDoc class, depending on your use-case.

from numpydoc.docscrape import FunctionDoc

def foobar(a, b):
   '''Something something

   Parameters
   ----------
   a : int, default: 5
        Does something cool
   b : str
        Wow
'''

doc = FunctionDoc(foobar)
print(doc["Parameters"])

I figured this all out by looking at this test in their source code. So, this isn't really documented, but hopefully is enough for you to get started.