Indexer cleanup

[donald-e-boyce]

Overview

See #908, unintentionally closed. This cleans up documentation and typing in indexer.py. The latest changes the ParamMP variable to a TypedDict.

[ZackAttack614]

This is definitely better! I'll propose that we add in typing for the elements of numpy arrays, as well. We've done that in most places, e.g. here. That allows us to statically analyze the script and know whether the data type is an int, float64, bool, or str - all of which are matrix element types we've worked with in the HEXRD ecosystem.

[donald-e-boyce]

So, in your file, what is the difference between np.integer and np.signedinteger? And is np.floatingpreferred for any float that is not explicitly something more specific? In other words, is np.floating, np.integer, and np.bool enough? And why are there numpy types for non-NDArrays? e.g. tTheta: np.floating -- couldn't that just be float?

class HKLData(TypedDict):
    hklID: int
    hkl: NDArray[np.integer]
    tTheta: np.floating
    dSpacings: np.floating
    tThetaLo: np.floating
    tThetaHi: np.floating
    latPlnNrmls: NDArray[np.floating]
    symHKLs: NDArray[np.signedinteger]
    centrosym: bool

[donald-e-boyce]

So, technically, it looks like it's wrong to use numpy datatypes for regular floats and ints.

In [2]: x = 10.0

In [3]: isinstance(x, np.floating)
Out[3]: False

[donald-e-boyce]

OH, I see. It's a typedict. I thought they were type hints.
never mind

[donald-e-boyce]

OK, I added the numpy scalar types and pushed everything up. Should be good now.