Export OD results in Parquet

[ChristopherRabotin]

• Fix semi minor axis computation (was NaN for non hyperbolic orbits)
• Add a timestamped initializer for ExportCfg (and related TODOs)
• Remove heterogeneous rms_pre/postfit_residual function, use rms_residual_ratios instead
• Remove old code in trigger.rs

• Merge Residual<T> and Estimate Compiler got very unhappy with all permutations I tried of this. This also complicates the fact that Estimate would need to know what kind of measurement is used: that isn't a deal-breaker but it isn't necessarily the most evident idea. For example, this would make it impossible to share a KfEstimate between a filter that runs on Range and one that ingests RangeDoppler.
• Add Kalman gain to Estimate (as an option since it isn't set for a time update) Requires the above.
• Remove predicted: bool from Estimate and change to a pub const function depending on whether there is a residual or not Now irrelevant
• Export the full estimate to parquet (on a separate thread)
• Add ExportCfg to Python interface to control how to export data in parquet

Effects

Closes #134

If this is a new feature or a bug fix ...

• Yes, the branch I'm proposing to merge is called issue-xyz where xyz is the number of the issue.

If this change adds or modifies a validation case

• No.

[codecov]

Codecov Report

Patch coverage: 81.81% and project coverage change: +1.48 🎉

Comparison is base (90e9926) 64.92% compared to head (d8f4c69) 66.40%.

❗ Current head d8f4c69 differs from pull request most recent head 87ae42b. Consider uploading reports for the commit 87ae42b to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #147      +/-   ##
==========================================
+ Coverage   64.92%   66.40%   +1.48%     
==========================================
  Files         126      127       +1     
  Lines       33087    32368     -719     
==========================================
+ Hits        21482    21495      +13     
+ Misses      11605    10873     -732     

Impacted Files	Coverage Δ
src/cosmic/eclipse.rs	67.16% <0.00%> (-1.44%)	⬇️
src/io/trajectory_data.rs	0.00% <ø> (ø)
src/md/mod.rs	0.00% <ø> (ø)
src/md/trajectory/mod.rs	0.00% <ø> (-60.61%)	⬇️
src/md/trajectory/traj_it.rs	30.00% <0.00%> (+2.72%)	⬆️
src/od/filter/mod.rs	30.00% <ø> (ø)
src/od/mod.rs	0.00% <ø> (ø)
src/od/process/rejectcrit.rs	37.50% <ø> (ø)
src/od/process/trigger.rs	88.00% <ø> (+22.32%)	⬆️
tests/orbit_determination/mod.rs	87.09% <ø> (-2.38%)	⬇️
... and 21 more

... and 9 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[ChristopherRabotin]

Cleaning up tasks:

• Change ODP in Python to return the path to the stored parquet file
• Ensure ExportCfg is available in Python
• Update OD plotting scripts for new format, include the ability to scale to covariance and plot as an uncertainty (should be the default)
• Remove NavFormatter (finally replaced by parquet output)
• Remove StateFormatter (replaced entirely by parquet output)

[ChristopherRabotin]

One last thing to make this immediately usable:

• Parse OEM file into a Spacecraft or Orbit trajectory