Computer cohomology instead of homology

[jakipatryk]

Ripser paper (https://arxiv.org/abs/1908.02518) shows how important for performence is computing persistent cohomology instead of persistent homology (which results in the same persistence pairs at the end).

• Redesign the library for cohomology and optimisations described in the Ripser paper #39
• Create the rest of subtasks after design is done