Compute the PN log-likelihood in the solver

[nathanaelbosch]

This PR implements computing the PN log-likelihood inside the solver.

Status and issues:

• the implementation depends on an unrelease version of PSDMatrices, which is not yet properly tested
• computing the log-likelihood comes at a cost, so we need to do more profiling to see where it comes from and how to get it down
• currently I only implemented the EK1. EDIT: Turns out the EK0 works too, nice!
• we need tests

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Comparison is base (32bb627) 93.84% compared to head (ae75001) 93.92%. Report is 2 commits behind head on main.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #281 +/- ## ========================================== + Coverage 93.84% 93.92% +0.08% ========================================== Files 36 36 Lines 1593 1614 +21 ========================================== + Hits 1495 1516 +21 Misses 98 98 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.