A rotated NaN is still a NaN#2901
A rotated NaN is still a NaN#2901maxinelasp merged 2 commits intoIMAP-Science-Operations-Center:devfrom
Conversation
Fix rotations outside of spice so they maintan FILL_VAL during rotation Before this we were introducing nearly but not quite NaN/FILL_VAL meaning very bad field data.
|
@maxinelasp please can you also review and release this ASAP please? And then regen all L2 BM. This issue means all our L2 BM science has very dodgy values like -0.999999999999e30 and not the proper FILL_VAL. I am away for a few days so please push this forward in my absence if possible. Thank you. |
maxinelasp
left a comment
maxinelasp
left a comment
There was a problem hiding this comment.
Great catch! I'll get this released ASAP and re-process.
maxinelasp
merged commit cdfd0ae
into
IMAP-Science-Operations-Center:dev
| constants.FILLVAL, | ||
| result, | ||
| ) | ||
|
|
There was a problem hiding this comment.
Oh, no! This is MAG specific code in the common geometry function. If the input is a NaN, this function already behaves as expected. If the input is a MAG specific FILLVAL, then that case should be handled explicitly in the MAG code, not the general geometry module. @maxinelasp please revert this.
There was a problem hiding this comment.
Shoot, good catch Tim! I assumed this was in the rotate frames function in the mag code and totally missed the file. Sorry! I will revert and put up a proper fix.
3 participants
Fix rotations outside of spice so they maintain FILL_VAL during rotation.
Before this we were introducing nearly but not quite NaN/FILL_VAL meaning very bad field data.
After this is merged and released please can the SDC regenerate new versions of all MAG L2 BURST science (not normal - as it has no NaNs).
Testing
Added failing unit test to repro the issue. We see this when publishing L2 science data.