Consider using the dataclass here instead for simplicity, performance gains, and consistency

The NamedTuple has some drawbacks such as Python allows accidentally compare NamedTuple to a tuple or namedtuple with the same number of fields:

Point3D(2017, 6, 2) == Date(2017, 6, 2)  # evaluates to True
Point2D(1, 10) == (1, 10)  # evaluates to True

That is probably unexpected behaviour, and can result in bugs. Additionally, NamedTuple are less flexible and more brittle than dataclass because:

• There is no option for mutable NamedTuple.
• Cannot specify default values for NamedTuple values.
• No control which fields are used for __init__, __repr__, etc.
• Cannot support combining fields by inheritance.

These drawbacks can be avoided by using a dataclass.