Exception handling inside core calibrator class

Summary

Need to have common method for exception handling inside the core calibrator class.

1. Figure out all possible exceptions

• Optical throughput:

  • FailedThroughputCalibration (can be as parent or for any unexpected case)
  • CorruptedMuonData (problems with input data):
    • No muons at all
    • Not enough good muons
    • Bad format of the data

• Atmospheric

  • External agents are not available (server down, protocol/format changed, authentication error etc.)
  • No Contemporary MDP
  • Wrong time period configuration
  • Lack of DAS data
  • Discontinuity of weather/climate data

• Camera Calibration

  • Bad input data
    • Not enough
    • Bad quality
  • CalibPipe fails to calculate for a nonnegligible N pixels
  • Memory issue

• Related to DB and .ecsv table storage

  • Database connection is not established
  • DB table does not exist
  • Not consistent container vs. DB table
  • Other unexpected exception related to the DB/ecsv storage

2. Exceptions should have structure of the tree based on the type and severity

• Type:

  • Bad input data
  • Bad config parameters
  • Bad output
  • Unable to store the output

• Severity:

  • Pipeline fails, can be relaunched with same data, other config
  • Pipeline fails, can be relaunched with other data, same config
  • Pipeline fails, can't be relaunched
  • Pipeline fails, should be used alternative method

3. Need to decide on the exit code value (exit_status) for each of the above mentioned options

4. Logic of whole procedure, from generating exception to the finishing of the program with the appropriate exit code:

• Exception should be generated inside the calibrate method, processed in the handle_exception method and transferred as exit_status value to the tool, where we have inherited method self.exit(exit_status) that will be received by the Workload