Tags

- [JF] Only emit a warning if all sequences of a job are empty
- [JF] Fix storing multiple ROIs for a single module
- [JF] Replicate a raw dataset in processed data for legacy analysis code

- [DataFile] Add support for older INDEX and METADATA versions.
- [LPD] Fix histogram plotting if selected cell's data is identical for trains.
- [JF] Avoid unclear error messages when the run has no trains.
- [AGIPD] Change notebook-default setting of max-task-per-worker to 1.
- [REMI] Allow pulse separation in REMI/DLD reconstruction to ignore FEL pulses.
- [webservice] Fix timestamps in the overview page.
- [Webservice] Check run type more directly in myMdC response.

- [JF] Add support for strixel sensors
- [JF] Shorter summary notebook for darks by default
- [JF] Always align corrected data properly
- [JF] Harden against 0 adc data in dark characterization
- [REMI] Add support for quad nodes and various pulse separation improvements
- [AGIPD] Allow manual photon energy input for photonization and add consistency plots
- Grant priority partitions for set-up time

* [EPIX100] Add Common mode corrections to dark characterization
* [EPIX100] Use a single value for temperature
* [JUNGFRAU] Fix bad pixel masking
* [JUNGFRAU] Limit number of plotted trains to 500 during corrections
* [AGIPD] Don't fail when there are no trains in sequence file
* [REMI] Use DataFile API and support asymmetric pump-probe patterns
* Skip corrections for runs which match specific types

* Significant improvements to AGIPG LitFrameFinder implementation, including offline support
* [AGIPD] Fix misalignment of data and gain/mask when common mode is used with manual pulse slicing

Deployment before the bring-up tests for SPB

Test tag 3.6.0

Cherry picked changes from below branch:
- fix/bias_voltage_integer
- fix/LPD-correct-single-pulse
- lpd-vcxi-require-4-mods

Minor release with improvements to LPD corrections for users and various fixes

Test Gotthard2 offline correction on FXE

Beta release for use at HED to run xfel-calibrate on ONC