DM-36597: Add LATISS filter transmission data #12
Conversation
czwa
commented
Nov 8, 2022
czwa
commented
- Add LATISS filter transmission curves.
- Update existing LSSTCam and TS8 sensor transmission curves to new IntermediateTransmissionCurve format.
* LATISS:
* This involved migrating back to the chip-based path.
* Updated the CALIBCLS to point to the subclasses.
* TS8/LSSTCam:
* Moved to "transmission_sensor" paths.
* Updated metadata. Of the following, the first updates to the
preferred dataset type, and adds a CALIBCLS. The second was
needed to fix that CALIBCLS to fix ingest.
* find . -type f -iname '*ecsv' -exec sed -i.bak 's/QE/transmission_sensor/; s/qe_curve/transmission_sensor/; s/# - {PICKLEFILE/# - {CALIBCLS: lsst.ip.isr.IntermediateTransmissionCurve}\n# - {PICKLEFILE/' {} +
* find . -type f -iname '*ecsv' -exec sed -i.bak2 's/IntermediateTransmissionCurve/IntermediateSensorTransmissionCurve/' {} +
| No negative values were clipped in any band, although they were likely | ||
| caused by random noise during measurement. For consistency with other |
There was a problem hiding this comment.
Whether noise or light leaks, I wonder if we should clip them. Seems like having negative values could be problematic and/or confusing, and while it might not be "right", I feel like it can't be the wrong call, given they're literally unphysical values.
There was a problem hiding this comment.
The negative values are almost certainly sampling noise in the measurements. I did think about clamping everything smaller than 5e-5 to zero, as that would also eliminate the values at that scale that pop-up at the ends of the transmission. I didn't want to do much value processing, to keep the on-disk values as close to the raw measurements as possible, letting the user decide what to do.
There was a problem hiding this comment.
Yeah, I agree, it's probably best to let code using the values deal with these, it's a single line to clip them all, so I think on balance this is indeed the best course here.
| # --- | ||
| # datatype: | ||
| # - {name: wavelength, unit: nm, datatype: float32} | ||
| # - {name: throughput, unit: '', datatype: float32} |
There was a problem hiding this comment.
Files below use - for the dimensionless value, but this one uses '' here and on L22. Does that matter?
There was a problem hiding this comment.
astropy doesn't like anything for dimensionless units (as far as I can tell), so I think these are equivalent.
There was a problem hiding this comment.
That's fine, just thought they should all agree. I'm sure it doesn't actually matter though.
