locklosses get mislabeled when guardain state numbers are added or changed

We have seen this problem with the control room lockloss tool before (when we deleted a state number from the guardian, the lockloss tool would not work from the command line if you try lockloss select.) Now it works, both on the web and the control room version, but the locklosses are mislabeled.

For example, yesterday I changed the order of some states in the ISC_LOCK gaurdian, and today we have this: https://ldas-jobs.ligo-wa.caltech.edu/~lockloss/index.cgi/event/1230678186

The lockloss is labeled in the command line tool as being from DOWN - to LOCKLOSS, but it really was a transition from DARM_TO_DC_READOUT (501) to LOCKLOSS.

Here's the output of the command line tool:

  3: 2019-01-04_22:23:52Z  ISC_LOCK  ACQUIRE_DRMI_1F -> LOCKLOSS
  2: 2019-01-04_22:36:37Z  ISC_LOCK  DRMI_LOCKED_CHECK_ASC -> LOCKLOSS_DRMI
  1: 2019-01-04_22:36:41Z  ISC_LOCK  ACQUIRE_DRMI_1F -> LOCKLOSS
  0: 2019-01-04_23:02:48Z  ISC_LOCK  DOWN -> LOCKLOSS
select event by index [0]: 0

When we had this problem before (and the command line tool didn't work at all) we eventually stopped and restarted the guardain (or maybe destroyed and re-created it?) which fixed the problem, so we will try this next time we are unlocked.