Repository navigation
satOmm2elem skips CelesTrak KVN files (MEAN_ELEMENT_THEORY = SGP/SGP4) #1673
Description
hey, I tried the new OMM reader with gpconf and every record in a CelesTrak KVN file gets skipped with "Unsupported OMM MEAN_ELEMENT_THEORY: 'SGP/SGP4'; expected SGP4." CelesTrak writes MEAN_ELEMENT_THEORY = SGP/SGP4 in their KVN (their XML just says SGP4, so XML reads fine), so on their 565-object analyst file satOmm2elem ends up loading none. would you be ok treating SGP/SGP4 as SGP4? thanks!
Metadata
Metadata
Assignees
Type
Projects
- Status
Activity
schaubh commentedon Nov 10, 2026
Thanks for the feedback. working on a fix.
hneogy commentedon Nov 10, 2026
awesome, thanks! if you want I can re-run the CelesTrak KVN file through gpconf against your branch once it's up
schaubh commentedon Nov 10, 2026
Sure, that would be a great sanity check. I'm checking the omm handling more generally right now to see if there are other things to improve. I'll let you know when the branch is pushed into a PR to run the online checks.
schaubh commentedon Nov 10, 2026
@hneogy , the PR is now running CI tests, see #1675 if you want to pull this branch and test against your code.
hneogy commentedon Nov 10, 2026
ran #1675 (1497db5) through gpconf the same way as before and it does the job: satOmm2elem now loads all 565 records of CelesTrak's analyst KVN, where it loaded none, and the XML and JSON load all 565 too.
across the corpus every KVN file but one reads right now. the one it still refuses is an OMM 3.0 record with no NORAD_CAT_ID. it says why, so your call whether that should be supported. as a bonus
NORAD_CAT_ID = +25544reads now, and a CSV cut off mid-row gets a clean "16 cells; expected 17" instead of a half-built record. nothing that passed before broke. thanks for the quick fix!hneogy commentedon Nov 10, 2026
separate from the OMM side: running the TLE reader through gpconf turned up four things in tleHandling.py, and I have small fixes with tests for each (full suite passes, and they don't touch your #1675 changes):
would you like these as separate issues and PRs? the last one changes behaviour (skip the bad set with a warning instead of raising), so I wanted to check that's the direction you'd want first.
schaubh commentedon Nov 10, 2026
howdy @hneogy , let' make TLE reading a separate issue and PR please. Thanks for the support test driving these new features.
hneogy commentedon Nov 10, 2026
Sounds good, will open a TLE issue and PR shortly.
schaubh commentedon Nov 10, 2026
Just pushed a fix to the NORAD_CAT_ID issue you mentioned. Should be all good now. thanks for checking on your end.