Suggestion for __getitem__/__setitem__ support

The main idea is, to allow add slicing functionality to the SaQC obejct, so that we can additionally express the following:

qc = SaQC(data)
qc = qc.flagRange(field="a")
qc = qc.flagRange(field="b", target="c")

as

qc = SaQC(data)
qc["a"] = qc["a"].flagRange()
qc["c"] = qc["b"].flagRange()

To get this rolling, (I think) the following changes are needed:

• __getitem__ with signature (str | Sequence[str]) -> SaQC to 'slice' the SaQC-Object into a subset of variables
• __setitem__ with signature (str | Sequence[str], SaQC) -> None to 'merge' to SaQC objects into one
• make field an optional parameter and default to 'process-all-columns'. While this might not be strictly necessary, something like the following is IMO a requirement to make the entire feature usable:

  qc["c", "d"] = qc["a", "b"].flagRange()

The code below is only there to express the ideas given above in code...What do you think @palmb ?

import numpy as np
import pandas as pd


class Prototype:
    def __init__(self, data) -> None:
        self._data = data

    def __getitem__(self, key):
        return Prototype({key: self._data[key]})

    def __repr__(self):
        return f"Test({self._data})"

    def __setitem__(self, key: str, value: "Prototype"):
        self._data[key] = value._data[tuple(value._data.keys())[0]]

    def flagRange(self, field=None, min=-np.inf, max=np.inf):
        if field is None:
            field = self._data.keys()

        if np.isscalar(field):
            field = [field]

        for k, v in self._data.items():
            self._data[k] = self._data[k].clip(lower=min, upper=max)

        return self


if __name__ == "__main__":

    p = Prototype(data={"SM1": pd.Series([1, 5, 10, 15]), "SM2": pd.Series([2, 18, 34, 64])})

    p["SM3"] = p["SM1"].flagRange(min=2, max=19)
    p = p.flagRange("SM1", target="SM3", min=2, max=19)
    print(p)