Doc/warning improvements

[dopplershift]

Description Of Changes

• Fix incorrect note in lcl doc (lcl works on n-dimensional arrays #2313)
• Link to the units tutorial from the units API page
• Tweak our sphinx config to use recently added support to use __all__ as the definitive source for what to include
• Improve the sidebar headings to be consistent and include a nice "English" heading as well as the full module name
• Make the error message on a calculation adjust the advice depending on the input data type (Incorrect instruction for assigning units in check_unit error message #2334)

With the last one:

import metpy.calc as mpcalc
import numpy as np

mpcalc.wind_speed(np.ma.array([1]), np.ma.array([2]))

gives:

ValueError: `wind_speed` given arguments with incorrect units: `u` requires "[speed]" but given "none", `v` requires "[speed]" but given "none"
A masked array `m` can be assigned a unit as follows:
    from metpy.units import units
    m = units.Quantity(m, "m/s")
For more information see the Units Tutorial: https://unidata.github.io/MetPy/latest/tutorials/unit_tutorial.html

whereas with a numpy array it gives:

ValueError: `wind_speed` given arguments with incorrect units: `u` requires "[speed]" but given "none", `v` requires "[speed]" but given "none"
A xarray DataArray or numpy array `x` can be assigned a unit as follows:
    from metpy.units import units
    x = x * units("m/s")
For more information see the Units Tutorial: https://unidata.github.io/MetPy/latest/tutorials/unit_tutorial.html

Checklist

• Closes Incorrect instruction for assigning units in check_unit error message #2334
• Closes lcl works on n-dimensional arrays #2313
• Tests added
• Fully documented

[dopplershift]

Current 1.2 docs (note the weird metpy.io in the sidebar):

With this PR:

[dcamron]

I like these changes, and final product seems to come together consistently on my machine and on actions.

[jthielen]

Looks good to me!