ensure_quantity should attempt to wrap arrays and primitives

[mattwthompson]

Here's a rough edge I found while debugging some code in which I can't trust the type of the object I'm interacting with:

In [1]: from openff.units.openmm import ensure_quantity

In [2]: import numpy

In [3]: partial_charges = numpy.linspace(-1, 1, 9, dtype=float)

In [4]: ensure_quantity(partial_charges, "openff")
---------------------------------------------------------------------------
Exception                                 Traceback (most recent call last)
Cell In [4], line 1
----> 1 ensure_quantity(partial_charges, "openff")

File ~/mambaforge/envs/openmmforcefields-test/lib/python3.9/site-packages/openff/units/openmm.py:240, in ensure_quantity(unknown_quantity, type_to_ensure)
    238     return _ensure_openmm_quantity(unknown_quantity)
    239 elif type_to_ensure == "openff":
--> 240     return _ensure_openff_quantity(unknown_quantity)
    241 else:
    242     raise ValueError(
    243         f"Unsupported `type_to_ensure` found. Given {type_to_ensure}, "
    244         "expected 'openff' or 'openmm'."
    245     )

File ~/mambaforge/envs/openmmforcefields-test/lib/python3.9/site-packages/openff/units/openmm.py:230, in _ensure_openff_quantity(unknown_quantity)
    226         raise ValueError(
    227             f"Failed to process input of type {type(unknown_quantity)}."
    228         )
    229 else:
--> 230     raise Exception

Exception:

In #44 I only considered that openmm.unit.Quantity or openff.units.Quantity would be passed through. I think some users would expect ensure_quantity to make some effort to coerce things into those types if asked. I'm less passionate about coercing int/float but the behavior should be similar.

The main downside of this approach is that it the resulting object must be a unitless quantity. Pint handles explicitly unitless quantities in a way that feels naturally to me via carrying around a dimensioness unit. OpenMM does a pretty good job with a similar solution but has a couple of gotchas, i.e. openmm/openmm#2247:

In [1]: from openff.units import unit

In [2]: unit.Quantity(1) / unit.Quantity(2)
Out[2]: 0.5 <Unit('dimensionless')>

In [3]: from openmm import unit

In [4]: unit.Quantity(1) / unit.Quantity(2)
Out[4]: 0.5