Ross Barnowski

There is a lot of overlap here with the [numpydoc validation checks](https://numpydoc.readthedocs.io/en/latest/validation.html#built-in-validation-checks), which I've never really wanted to enable for the same reason :)

FWIW I was recently looking at pydocstyle and noticed that [it no longer seems to be maintained](https://pypi.org/project/pydocstyle/#history), at least not at the pace of other linting tools.

I'm going to go ahead and close this one - AIUI the purpose of the PR is to demonstrate the difficulties in adopting automated doc linting. That's not to say...

Thanks @dschult I'm personally +1 on the idea of modifying `check_create_using` in the proposed way. I very much like the pattern of being able to do the property check (i.e....

I was recently looking at the gexf spec for another PR and given my most recent reading I believe @dschult 's interpretation/explanation is correct. IMO leaving the conversion of "liststrings"...

Thanks @lkk7 for the fix and ping! We'll see if we can't be more diligent about upstreaming issues we get related to `nx_pydot`. Closing as resolved by the pydot 4.0.1...

Adding the `close?` label for now - see https://github.com/networkx/networkx/pull/7965#issuecomment-3053379335 for discussion!

> I really don't get this point. Why is NetworkX using XML modules to serialize GraphML or Gexf, but should not use JSON to serialize Cytoscape? The main difference is...

Thanks for diving into the history @dschult ! While it may look strange in the code, unless it negatively affects the visualization I'm inclined to leave it as-is. @rudyarthur did...

> I can give some example code if you like. This would be great, especially if there's a relatively minimal example that illustrates the behavior! > if all you care...