Muon Calibrator refactoring

Summary

I plan to refactor the whole muon throughput routine to be more efficient and to satisfy our current goals.

Main changes:

• First of all, I propose slightly refactoring again the throughput_calibrator. Let's make all calibration work including handling exceptions and DB/ecsv population in the MuonCalibrator. So throughput_calibrator should (see also figure below):

  • Initialise MuonCalibrator with the correct config and input file
  • Call the method calibrate
  • Receive the information about exceptions handled by Calibrator and return to the Workload correct exit status.

• Refactor calibpipe.throughput.throughput_containers because they contain redundant containers, i.e.

  • In fact, we need only one container, that will contain only fields that will be uploaded to DB, and regarding current naming, this is TelescopeOpticalThoughtputContainer. This container contains information about obs_id and tel_id so I would delete the other two containers related to obs_id and tel_id, and subsequently rename this container to just OpticalThroughputContainer.

• Now let's deal with a MuonCalibrator. I think it is working not efficiently and with a redundant number of methods. For example:

  • It performs calculations for every telescope in every event, then stores not full information in some intermediate table, then checks for what telescope we can calculate efficiency (good_muon_data), then for each telescope calculates mean values and stores them in another container, and then uploads child container to DB.
  • My idea is
    • to perform calculations only for the telescopes that have good_muon_data (in that way drastically reduce the number of calculations for multi-telescope event files)
    • make the intermediate table that stores results to look the same as in DB (in that way it will be easier to upload the data to DB - we will only need to perform mean/std/etc. operations to all rows in the table for each telescope)
    • overally reduce the number of methods and make the code much clearer

What is the expected correct behavior?

Relevant logs and/or screenshots

Schematic view of how the calibration pipeline should look from my view: