Dev : new functions for batching + dask compatibility

[gviejo]

1. tsd.get_chunks(window_size, overlap) -> slice object. Helper function to create chunking of time series. Useful for writing batch loop to load memory map chunk and apply some functions.
2. tsd.apply(func, windowsize, overlap) -> new tsd. Batching function that apply func over time series. Calls the method get_chunks.
3. tsd.compute() To be used when passing a dask array. Solely for dask compatibility.
4. tsgroup.groupby("metadata_col") -> dict of tsgroup
5. tsgroup.groupby_apply(metadata_col, nap_function, arg, kwargs) -> results. Group and apply a pynapple high level function.