New condensed version of Swift BAT Position notice. #1
Conversation
|
Nice! This is a pretty thorough model of the existing GCN schema for Swift as an AVRO packet. Does it parse? |
| {"name": "NOTICE_TYPE", "type": "string", "doc": "type of notice"}, | ||
| {"name": "TRIGGER_NUM", "type": ["int", "string"], "doc": "integer or string of digits uniquely identifying event" }, | ||
| {"name": "SEG_NUM", "type" : "long", "doc": "Swift serial number (for Swift notice types only)" }, | ||
| {"name": "GRB_RA_J2000", "type": "double", "doc": "ra of GRB/Transient in J2000 epoch [deg]"}, |
There was a problem hiding this comment.
- J2000 has been the preferred equinox for a long time. Do we need to specify that in the field name?
- How about just "ra", so that the field name is general to all types of transients, not just GRBs?
There was a problem hiding this comment.
This notice type is specifically only for GRBs, I think BAT would generate a different notice type if it was a non-GRB source. Though for simplicity and comparison to other notice types, and the general use of RA/Dec, it would be easier to parse these streams if not GRB specific.
|
Note that Avro allows you to set aliases for fields. This might be an option to provide new fields with consistent lower-case capitalization while retaining the old names too. |
| @@ -0,0 +1,12 @@ | |||
| { "type": "record", | |||
There was a problem hiding this comment.
This is the referenced schema that would be used in the future enhanced versions.
| @@ -0,0 +1,39 @@ | |||
| {"type": "record", | |||
There was a problem hiding this comment.
This is the referenced schema that would be used in the future enhanced versions.
| "fields": [ | ||
| {"name": "grb-ra-j2000", "type": "double", "doc": "ra of GRB/Transient in J2000 epoch [deg]"}, | ||
| {"name": "grb-ra-current", "type": "double", "doc": "ra of GRB/Transient in Current epoch [deg]"}, | ||
| {"name": "grb-ra-1950", "type": "double", "doc": "ra of GRB/Transient in 1950 epoch [deg]"}, |
There was a problem hiding this comment.
I really do not think that we need to provide precomputed coordinates in epochs other than J2000 (including true of date).
Some tweaks to validation script
No description provided.