Support Ymsr rotation into local topocentric (or at least PLH) for reporting adj statistics (.adj file; point-by-point)

[jhaasdyk-au]

Brief description
RE: --output-adj-gnss-units 1 (Local [east, north, up)... and similar for Ymsr
Reporting local topocentric corrections, stdevs, N-stat, etc for Gmsr & Xmsr has been very useful.
Please consider making similar reporting available for Ymsr adjustment statistics.
It would be very useful to see adjustment stats for each Ymsr triplet presented in local topocentric (or at least PLH), in particular to help identify if height is a particular problem with the supplied constraint.

Related issue(s)
This request also relates to issue 285(Closed), in that the same sorting options for outputs should apply.

Background
It is common for Ymsr to be applied as (single or a matrix of) constraints in both local and large scale adjustments.

Currently, Ymsr adjustment stats are only reported as supplied (XYZ or LLH), regardless of the use of --output-adj-gnss-units (see example image below). It is not possible to request their rotation.

While Ymsr themselves cannot simply be rotated into a local topo ref frame, their CORRECTIONS could be reported in enu (or at very least in PLH). However, this presents a quandary. Consider either:

1. report the measurement in PLH, and corrections / stats also in PLH? That would be easier to interpret than XYZ, except for the relative contribution of horizontal vs vertical.
2. report the measurement in PLH, but the corrections in enu? That would allow easy interpretation of the horizontal vs vertical stats but it deviates from the usual presentation where measurements and corrections are in similar units. That said, if labelled as 'enu' (vs 'PLH'), at least '-35 56 40.0000' in the 'Measured' column would be easy to interpret.

Ymsr can also be VERY large, with covariances between all stations.
However, despite their size, the Ymsr is reported in the .adj file as (a block of) station-specific triplet(s).
For purposes of presentation only in this context, each triplet could be rotated into its own local frame.

(Aside: While LLH is an acceptable coordinate type for Ymsr, LLh is silently rejected; the default XYZ is assumed.)

Basic requirements

1. For Ymsr, provide a new option to rotate the adjustment file stats into enu (or PLH, or XYZ)
   Probably as a new argument (vs --output-adj-gnss-units)
   0: As measured (default)
   1: Local [CORRECTIONS supplied as east, north up; MEASURED and ADJUSTED supplied as PLH]
   2: PLH
   3: XYZ
2. Report a separate local reference frame for each Ymsr triplet
3. Consider separately sorting each Ymsr triplet as if they are individual points. (e.g. by Nstat).
   (Interestingly, I wouldn't want this for Xmsr, at least not for the purposes that we usually employ them)
   (Perhaps a separate option to allow this?)

Priority
Priority 2 - Medium

Image shows current state with adjust.exe argument --output-adj-gnss-units 1 (i.e. Local [east, north, up)