Lighter objects

[PMeira]

Currently, to avoid too much code duplication, we reuse some of the older DSS-Python class infrastructure.

Since the plan is to reimplement some parts, one important aspect is to make the objects lighter in terms of memory. Python is a bit memory heavy anyway, so we could also nudge users to other implementations to complement it (C++ as in dss.hpp, Julia and Rust whenever the impl. is done).

[PMeira]

This is implemented through the new iterate() method on the batch objects since b2 (pip install "altdss>=0.3.0b2"). This function returns a Python iterator, which reuses a single AltDSS Python object. The caveat is that if the user wants to keep a copy, the copy must be explicit, since the temporary object changes with each iteration step, and it is invalidate at the end.

Docs in the PR #12 already recommend this.

I also created #13 to address using this iterator objects to create highly-customizable batches more efficiently.

so we could also nudge users to other implementations to complement it

This will be a future effort to try to present a more unified toolset (which is how we use at Unicamp).