Edits are on u/yusra/
Simon, can you review the obs_lsstSim changes?
Russell, can you review the ip_isr and pipe_tasks changes?
ip_isr/pipe_tasks - Summary of method name discussion: IsrTask has a clearly defined primary method which removes the instrument signature. For this reason, a name such as "removeInstrumentSignature()" or "doISR" is unnecessary since IsrTask is unlikely to be used for its other methods. A simple name such as "run" is appropriate, but "run" should be reserved for commandline tasks. For these reasons, I settled on "apply."
obs_lsstSim - Functionality is untested on a full chip. This should be tested when a full chip of simulated lsst data becomes available. The new unit test operates on one amp in tests/data and does not attempt to assembleCcd.
Creating a unit test on an amp-sized image uncovered unusual behavior in assembleCcd. If amp-sized exposure is passed in as a ccdExposure, assembleCcd duplicates it across the chip. I expected assembleCcd to either:
- Throw an exception
- Place the amp in its proper place on the image and set the rest to nans (which get interpolated to the mean)
This brought up the question of what to do when an amplifier table is missing data. I think this requires some thought, and did not address it in this ticket.