calcat_interface2: Fix require_calibrations() when none of a given calibration type found
Description
When no offset constants are found for a given detector, .require_calibrations() was raising a KeyError, when we expect it to deselect all modules and return an empty CalibrationData object. We have seen this for both LPD & JUNGFRAU.
How Has This Been Tested?
xfel-calibrate jungfrau CORRECT \
--karabo-da JNGFR01 --receiver-template JNGFR01 \
--karabo-id MID_EXP_JF500K1 --run 300 \
--in-folder /gpfs/exfel/exp/MID/202302/p004451/raw \
--out-folder /gpfs/exfel/data/scratch/kluyvert/jf-corr-p4451-r300Now fails with an error from the notebook:
Exception: Could not find offset constants for any modules, will not correct data.A separate fix is likely needed in the JUNGFRAU correction notebook for the fallback to dynamic gain constants, as it's currently catching the wrong error.
Types of changes
- Bug fix (non-breaking change which fixes an issue)
Checklist:
- My code follows the code style of this project.
Reviewers
Merge request reports
Activity
added 1 commit
- f3cf6d4f - Fix fallback to dynamic gain constants in burst mode
I have also now updated the notebook code to fall back to dynamic gain constants for burst-mode data. The run I'm testing with (p4451 r300) still doesn't find any constants - maybe because it's from last year but has a non-default exposure timeout, which was only added as a condition more recently. But I see the warning message before it errors out with no constants:
WARNING:root:No fixed gain offset constants found. Looking for dynamic gain constant. ... Exception: Could not find offset constants for any modules, will not correct data.Edited by Thomas Kluyver276 273 " \"\"\"\n", 277 274 " from datetime import datetime\n", 278 275 "\n", 279 " from cal_tools.calcat_interface import CalCatError\n", 280 "\n", 281 " try:\n", 282 " jungfrau_consts = jungfrau_cal_mdata(gm=1)\n", 283 " except CalCatError as e: # TODO: update with CalibrationError class\n", 276 " jungfrau_consts = jungfrau_cal_mdata(gm=1)\n", 277 " if not jungfrau_consts.aggregator_names:\n", changed this line in version 3 of the diff
-
Good idea. In fact it would already kind of work in a boolean context, because Python falls back to checking
__len__(), but it would do the wrong thing if we have calibration types present but no constants in them.I haven't tested this again after a33d66a7, since it's quite a small change. That probably guarantees that I've made a silly mistake.
added 1 commit
- a33d66a7 - CalibrationData.__bool__ to check if any constants present
changed milestone to %3.16.0
mentioned in commit ac3a2137
mentioned in merge request !1129 (merged)
