Change func name to sort_runs_by_gain, fix docstring, use sorted runs, and fix high gain run check condition

- Use no-cluster-job by default in xfel-calibrate CL
- Introduce --use-slurm to use Slurm for xfel-calibrate
- small refactors e.g. isort
- separate code in functions to reduce main pytest function.

- Decrease number of processed sequences
- Choose random sequence numbers to process
- Refactor test code for readability
- Use threadpool for I/O parallel processing
- Add logs instead of prints
- Only apply find-difference code on a random file.

- Add new configurations for available tests like blc-strips for AGIPD
- Add configs for epix100 and DSSC after changing default values in main notebooks. To test old data.
- Add new Jungfrau and GH2 tests.
- Fix the numerical validation and fixes to `--find-difference`.
- Add arg to only apply numerical validation with slurm sumbitions.