I would like to work on this as my first contribution, is it open? :)

It seems the basic classes are documented, no?

np.float16.__doc__ is None on my machine. Are you testing with PyPy? I think this might be another case of things happening after PyType_Ready.

Thanks @YannickJadoul for knocking out a whole bunch of these!

Hi, can I work on the documentation for the unticked functions? Or have they been done already?

@smellslikekeenspirit you can check by help(), if it returns None then the documentation is missing. If you find they are not missing, please let us know so we can mark them.

Hello. Potential first time contributor here.

I re-ran the script up top and got:

• <function <lambda> at 0x0044D300> (np.deprecate_with_doc) - DOC: Doc for deprecate_with_doc #17852
• <class 'numpy.TooHardError'> (np.TooHardError) - DOC: adding docstring to TooHardError class #21471
• <function __dir__ at 0x00B1BFA8> (np.__dir__)
• <function __getattr__ at 0x009EE780> (np.__getattr__)
• <built-in function int_asbuffer> (np.int_asbuffer)
• <function loads at 0x04E096F0> (np.loads)
• <function show at 0x04D23D20> (np.show_config)
• <class 'numpy.bytes_'> (np.bytes0, np.bytes_, np.string_)
• <class 'numpy.datetime64'> (np.datetime64)
• <class 'numpy.str_'> (np.str0, np.unicode_, np.str_)
• <class 'numpy.timedelta64'> (np.timedelta64)
• <class 'numpy.void'> (np.void, np.void0)

I'm struggling to understand what needs to be accomplished. Any guidance would be appreciated.

Looking at np.datetime64 below. I see that np.datetime64.doc is None. But when checking help(), I see output for that. Can someone explain what the current help is returning, and what it is lacking?

>>> print(np.datetime64.__doc__)
None
>>> help(np.datetime64)
Help on class datetime64 in module numpy:

class datetime64(generic)
 |  Base class for numpy scalar types.
 |  
 |  Class from which most (all?) numpy scalar types are derived.  For
 |  consistency, exposes the same API as `ndarray`, despite many
 |  consequent attributes being either "get-only," or completely irrelevant.
 |  This is the class from which it is strongly suggested users should derive
 |  custom scalar types.
 |  
 |  Method resolution order:
 |      datetime64
[...]

Hi @engineerlisa, welcome! I would suggest to start with np.show_config, that's a straightforward function that is quite useful and not yet documented. What you want to do is:

• write a docstring for the function in the same format as other NumPy function have their docstring. (note, look for generate_config_py in numpy/distutils/misc_util.py for show(), that's what generates the whole __config__.py file)
• insert an entry for show_config in the reference guide. there's not really an obvious place for it, perhaps in https://numpy.org/devdocs/reference/routines.help.html is best. if you see a better place, please suggest that instead

Note that show_config already has a PR to document it at gh-9312, it would be worth building on the previous attempt.

Hmm, that PR looks a little complicated. @engineerlisa please have a look if that PR makes sense to you (@ghost is someone who disappeared from GitHub); the weird formatting stuff in the PR is because the docstring goes through some machinery that generates the __config__.py file so simply """ ... """ content won't work.

@eric-wieser I removed the "good first issue" label, the left overs aren't a matter of writing docs anymore. I think a maintainer should finish the last bits here.

Thank you for getting back to me. I understand some parts of that PR. Based on your removing the tag, I'm going to find another candidate for my first contribution.

Just two left here

• np.deprecate_with_doc, which could do with a docstring similar to np.deprecate
• np.TooHardError, which I think is raised only by https://numpy.org/doc/stable/reference/generated/numpy.shares_memory.html

Both of these should be easy to document, so I'm going to add back the "good first issue" label

Hey, I'm a first time contributor and I would like to take up the remaining two functions if they are still open. Any guidance would be appreciated. Thanks!

Hi @kumudlakara. You should start with the contributor guide for an overview, especially the section on documentation and then the development workflow for the details. We don't assign issues, so be sure to mark your PR with xref gh-10106 to get a cross-reference to show up here, that way people will know you are working on it.

Alright, Thank you @mattip .

Hi, I'm a first time contributor and I would like to work on this issue if they are still open. Any guidance would be appreciated. Thanks!

Hi @Radhika14soni and welcome. I commented above on how to get familiar with the process. Since deprecate_with_doc is underway, the only remaining function lacking documentation is np.TooHardError. It is raised only by https://numpy.org/doc/stable/reference/generated/numpy.shares_memory.html, which should give you a hint as to the needed docstring.

@mattip Hello I'd like to work on it. Am I supposed to make changes in _add_newdocs.py?

Since this is a pure-python class, the change would be to add a docstring to the class definition like for the other exception classes in that file.

Is there any works pending on this issue that I can take ? I am new to numpy contribution and looking for anything to work on

From the comments above it looks like this issue is taken care of, @mugunthanramesh . But feel free to browse other Documentation-related issues and ping me if you need help getting started. Cheers!

Closing then. Please open a new issue if there are still missing docstrings, this one is too difficult to navigate.