IMAP-Lo requested tweaks

[vineetbansal]

Closes #3289, #3270

• modified values of pivot angle thresholds

• get_star_bin_offset guided by STAR_BIN_OFFSET_BY_SYNC dict

• pulled out 2 constants - STAR_END_BINS_TO_EXCLUDE, STAR_MIN_COUNT_THRESHOLD

• lo_angle_offset gone, per issue Lo L1b prostar - remove two degree offset #3270

• New ancillary files for geometric factors. Uncertainty values for H/O are now asymmetric, with the ancillary file having columns:

  • esa_mode
  • incident_E-Step
  • Observed_E-Step
  • Cntr_E
  • Cntr_E_unc
  • GF_Trpl_H
  • GF_Trpl_H_unc_minus
  • GF_Trpl_H_unc_plus

  The value of systematic errors determined in the code (and stored in the CDFs) is by taking a mean of the two systematic errors. This is to avoid major disruptions for now. Another PR (coming soon) will try to tackle this in a coordinated way among the mapper instruments (Lo/Hi/Ultra).

[tmplummer]

Are there instances where the value is not either "DS", or "EN". I'm wondering if this silently handling a None value is a bad thing to do. It seems like a major problem that should fail loudly if the state is not one of the two options.

[vineetbansal]

sounds good - done. Now we fail loudly with a KeyError if the value is not in there (one of DS/EN). I've confirmed that DS/EN are the only two values in the Lo HK files for this attribute (at least going till mid May from mission start).

[vineetbansal]

@tmplummer - the CI was failing on docs because of a completely unrelated thing (spiceypy wheels not available on 3.10), so I've gone ahead and fixed it in a commit here by generating docs on 3.12, just to have a clean CI pass. In case you don't want it here, I can open it in a separate PR.

[tmplummer]

This all looks ok to me.