For now, calibration models are keyed off of exposure times. The mission (including SPOC products) have only every used one 2D bias model, which is just scaled to the appropriate exposure times.
However, we check the 2D bias every year (usually in April) to look for changes in the fixed pattern noise. In principle, this could lead to different versions of the 2D bias model used by TICA for a given exposure time.
The user is mostly responsible for setting the calibration model, so this is not a big development change. But it is good to have guardrails in case the wrong models are specified, or at least a warning to say "this 2D bias model differs from what was used by the misison pipeline".
Likely this needs a dictionary keyed of sectors, which can look up the appropriate model and check what is being read in twoDbias_ directories.
For now, calibration models are keyed off of exposure times. The mission (including SPOC products) have only every used one 2D bias model, which is just scaled to the appropriate exposure times.
However, we check the 2D bias every year (usually in April) to look for changes in the fixed pattern noise. In principle, this could lead to different versions of the 2D bias model used by TICA for a given exposure time.
The user is mostly responsible for setting the calibration model, so this is not a big development change. But it is good to have guardrails in case the wrong models are specified, or at least a warning to say "this 2D bias model differs from what was used by the misison pipeline".
Likely this needs a dictionary keyed of sectors, which can look up the appropriate model and check what is being read in twoDbias_ directories.