Possible annotation schema additions

[genghisken]

Just raising this for discussion. We've discussed verbally but would be good to write down so we have a record. Also this shows that we spent time doing this, even if the request is ultimately rejected.

• Our users (fastfinder) wanted to know if we could have a (default null) "summary score" (float) and a "summary score uncertainty" (float) column in the main annotations table to allow indexed selection from the main table. These values can be buried as JSON key/value pairs in the classdict column, but the complex regex required to search it will slow down the SQL query. One issue that we could see is that annotators like 4MOST might want to stick the redshift somewhere that's not buried in the classdict column (e.g. so we can immediately grab all the SN Ia transients within a certain volume). Could re-use summary scores/uncertainties, or have yet another column for redshift (and/or host redshift).
• We also discussed that IF Annotators were have common high classification categories (e.g. KN, SN Ia, etc) could we get them to agree on a common transient label (e.g. as defined by TNS).