Investigate differences in guiding offsets for two tested modes

From yuki.moritani

Guiding was not stable occasionally in this run: only 1 or 2 cameras were used although other cameras had stars whose spot was detected by agcc.

Learning that agActor uses two options to filter the guide star candidates, we compared how measurement was stable (i.e. consistency of measurements) in two modes, by repeating the error measurement without sending signal to the telescope.

visit 135434--135437 were test data. It seems that measurement was table with only filter out GALAXY object, rather than use the GAIA|PMRA|PMDEC|PARA|ASTROMETRIC|NON_BINARY|PHOTO_SIG filter (which is currently used for guiding). I think this is simply because the number of stars and balance between the cameras.

See also the last page of Yuki's report from 20251202

Look into the above visits to check guiding performance as well as the breakdown of which stars were used for catalog, detection, and eventual matching.