DOC: Validation script doesn't report that the docstring is correct

[datapythonista]

When executing for example:

./scripts/validate_docstrings.py pandas.DataFrame.head

The script, after the docstring, should report:

################################################################################
################################## Validation ##################################
################################################################################
Docstring for DataFrame.head correct. :)

But for some reason, the sentence Docstring for DataFrame.head correct. :) is not being printed.

This is the line that is not being executed: https://github.com/pandas-dev/pandas/blob/master/scripts/validate_docstrings.py#L364

[pikamegan]

I might be completely off, but the if and elif statement have the same condition result["errors"] so if it meets the if condition, it just skips the elif statement and thus it will never print the sentence. Perhaps getting rid of the elif statement completely or changing the condition would work?

[miguelmarques1904]

@pikamegan I believe so too. I'm currently testing the docstring by simply replacing the elif statement with an else statement. Are you taking the issue?

[pikamegan]

@miguelmarques1904 I actually don't know how to test the function (new to open source). Tried running ./scripts/validate_docstrings.py pandas.DataFrame.head in git bash and it returned permission denied. Should I be running it in python itself with an actual data frame?

[miguelmarques1904]

@pikamegan Sorry for the delayed response.

You can try typing chmod +x ./scripts/validate_docstrings.py, which adds execute permissions to the ./scripts/validate_docstrings.py file.

Welcome to the community and best of luck.

[MarcoGorelli]

FWIW, I usually run it as python scripts/validate_docstrings.py <method I'd like to validate docstring for>

[pikamegan]

@miguelmarques1904 thanks for the response, it still returns /usr/bin/env: ‘python3’: Permission denied even though I am using the correct version of Python and it is in my path (as various stackoverflow posts suggest).

[pikamegan]

@MarcoGorelli this worked, but now I get the following error:
Traceback (most recent call last): File "scripts/validate_docstrings.py", line 50, in <module> from numpydoc.validate import validate, Docstring # noqa: E402 isort:skip ModuleNotFoundError: No module named 'numpydoc.validate'

Google searched and there isn't such a module, so what is this line about? Would appreciate any help :)

[datapythonista]

Thanks @pikamegan for helping with this. It's unfortunately a bit more complex. What you probably want to do is:

• Write a comment in this issue with just the word take, so the issue is assigned to yourself, and other people don't work on it
• Set up a development environment. You have extensive information here: https://pandas.pydata.org/docs/development/contributing.html And we've also got a simplified version for a sprint, that now is probably outdated in few things, but could be helpful: https://python-sprints.github.io/pandas/guide/pandas_setup.html
• Then you'll be able to test your changes locally
• You'll finally have to submit a PR, if you're not familiar, besides the contributing guide above, you can check this document: https://python-sprints.github.io/pandas/guide/pandas_pr.html

Let me know if you have any question. For the specific error you've got, you'll have to install miniconda (or anaconda), run conda env create, and then activate the environment conda activate pandas-dev. With this, you'll have all the dependencies needed (the error is telling you that you don't have numpydoc installed).

[pikamegan]

@datapythonista Thanks so much for pointing me in the right direction. I'll probably try to familiarise myself with the development environment first, so someone else can take this issue

[rtm010]

Hi all, new to open source. Got my dev environment set up and happy to take this issue 😄

[rtm010]

take