[CORRECT] FIX: Consider list of karabo_da and test_balance_sequences

Summary:

This MR introduces a fix for a reported bug during testing offline calibration on max-exfl016 on Friday.

data is processed but we do not have any report and slurm output is still there. run 511 from p900145

The run has 39 JF sequences, but about 76 sequences for slow data. balance_sequences didn't check for a list of karabo_das, and all 76 sequences were being corrected.

There was also a bug to split the number of sequences properly on max-nodes. (e.i. if the number of sequences is 9, the nodes used were only 5.)

I have added bit of validation for error prone cases and a test_balance_sequences

Tests:

xfel-calibrate JUNGFRAU CORRECT \
--in-folder /gpfs/exfel/exp/SPB/202031/p900145/raw \
--run 511 \
--out-folder /gpfs/exfel/data/scratch/ahmked/test/JF

output ->

Split concurrency into [redacted]

xfel-calibrate AGIPD CORRECT \
--out-folder /gpfs/exfel/data/scratch/ahmedk/AGIPD_ \
--run 205 \
--in-folder /gpfs/exfel/exp/MID/202002/p002718/raw \
--sequences 0

output ->

Split concurrency into [redacted]

Reviewers:

@danilevc @hammerd @roscar