Unify docstring styles

[sandorkertesz]

The docstrings varies between modules in earthkit-meteo and would be great if a common style could be used. E.g. 2 variants:

def cdf(self, x):
        """Evaluate the cumulative distribution function (CDF).

        Parameters
        ----------
        x: Number | array_like
            Input value.

def mixing_ratio_from_vapour_pressure(e, p, eps=1e-4):
    r"""Compute the mixing ratio from vapour pressure.

    Parameters
    ----------
    e: number or array-like
        Vapour pressure (Pa)

Other questions:

• does "array-like" also mean number or we always have to use something like "number or array-like" to indicate that an arg/kwarg can be both a number and array?
• what typehint should be used for "array-like"?

[oiffrig]

I have a preference for the second one, although I think NumPy uses array_like in its docs.

According to the NumPy glossary, array_like includes scalar. Not sure what is best in that regard. I tend to give hints on array shape when relevant, we could try and come up with a rule. For typing we can use numpy.typing.ArrayLike

As a general rule I try to stick to the NumPy docs style guide: https://numpydoc.readthedocs.io/en/latest/format.html

[sandorkertesz]

I would also choose the second option.

I checked the Xarray (web) docs and it uses array_like. But it also uses dict-like. So it is a bit confusing.

[oiffrig]

I think I have seen file-like, dict-like and such in the Python standard library docs, maybe they are just trying to stick to the conventions already in use. But I agree this is confusing (I wasn't aware that it was written array_like until I went to check the docs yesterday...)