[TASK] Analyze PySAT implementation for bottlenecks

## Task

Need to profile and read what PySAT actually does to call solver.

## Current state

Obliviously using PySAT for solver calling.

## Problems

- `add_clause` in PySAT sometimes is heavy (maybe optimizations are called right after adding new clause)
- ensure there is no overhead on solver instance creation.