feat(posfilter): Add Kalman Filtering functionality

[madhavmk]

Overview

Added kalman_filtering operation as proposed by @SquirrelKnight on https://github.com/seafloor-geodesy/gnatss/pull/167. This integration includes re-writing the proposed functionality to a more functional programming with numba to really speed up the performance. A few tweaks has also been made to match the steps of the kalman filtering to the equivalent kfwgpos.F Fortran implementation.

Closes #235

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 21.77419% with 194 lines in your changes are missing coverage. Please review.

:exclamation: No coverage uploaded for pull request base (dev@a8ea7d4). Click here to learn what that means.

Files	Patch %	Lines
src/gnatss/ops/kalman.py	18.97%	158 Missing :warning:
src/gnatss/ops/posfilter.py	9.67%	28 Missing :warning:
src/gnatss/loaders.py	55.55%	4 Missing :warning:
src/gnatss/main.py	0.00%	2 Missing :warning:
src/gnatss/ops/data.py	0.00%	2 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## dev #238 +/- ## ====================================== Coverage ? 44.04% ====================================== Files ? 18 Lines ? 1294 Branches ? 0 ====================================== Hits ? 570 Misses ? 724 Partials ? 0 ```

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

[madhavmk]

Hello @lsetiawan, I made the changes suggested, and incorporated the diagonal values to retrieve the covarianec values. However the cov_xx, cov_yy, and the cov_zz are not close to the gold standard values in the POS_FREED_TRANS_TWTT files. Could you take a look? (at the Optimized_Kalman.ipynb in particular)