In place numpy array operations causing issues with dimensionality of quantities

[domyam97]

In the PlainQuantity class the method dimensionality() only updates the dimensionality of the object when it is previously None. This is causing issues when working with numpy ndarray quantitites and in-place operations.

For example:

import numpy as np
from pint import UnitRegistry

ureg = UnitRegistry()
quantity = ureg.Quantity

# create a numpy array based quantity
x = quantity(np.array([[20.0]],"m")
print(x.dimensionality) # returns {'[length]':1.0}

# divide by 1m to get a dimenstionless quantity using in-place operation
x /= quantity(1,"m")
print(x.dimensionality) # still returns {'[length]':1.0}

Becuase the array was modified in place, a new object was not created and therefore the dimensionality was not reset to None. Subsequent calls to dimensionality do not reflect the changes caused by the division.

[domyam97]

Is there a reason the check that the _dimensionality parameter is None exists? It seems unnecessary and causes issues like this.