Division by zero and zero trigger rates from calculate_trigger_rate application
Opened this issue · 1 comments
The integration test for calculate_trigger_rate.py defined in calculate_trigger_rate_one_file_save_table.yml passes the tests, but does not give reasonable results (some of the other integration tests for calculate_trigger_rate also have this problems, but not all):
python simtools/applications/calculate_trigger_rate.py --config tests/integration_tests/config/calculate_trigger_rate_one_file_save_table.yml
INFO::calculate_trigger_rate(l181)::main::Calculating simulated and triggered event rate
INFO::base(l61)::__init__::Opening new file ./tests/resources/run201_proton_za20deg_azm0deg_North_test_layout_test-prod.simtel.zst
INFO::base(l94)::__init__::Found zstd compressed file
{'view_cone': <Quantity [0., 0.] deg>, 'solid_angle': <Quantity 0. sr>, 'total_area': <Quantity 1.25663706e+11 cm2>, 'energy_range': <Quantity [3.0e-03, 3.3e+02] TeV>, 'total_num_simulated_events': 2000.0, 'total_num_triggered_events': 1.0}
INFO::simtel_io_histograms(l285)::_rates_for_each_file::Histogram 1:
INFO::simtel_io_histograms(l286)::_rates_for_each_file::Total number of simulated events: 2000.0 events
INFO::simtel_io_histograms(l290)::_rates_for_each_file::Total number of triggered events: 1.0 events
/workdir/external/simtools/simtools/simtel/simtel_io_histogram.py:607: RuntimeWarning: divide by zero encountered in scalar divide
return (stacked_num_simulated_events / first_estimate) * u.s
INFO::simtel_io_histograms(l298)::_rates_for_each_file::Estimated equivalent observation time corresponding to the number of events simulated: inf s
INFO::simtel_io_histograms(l304)::_rates_for_each_file::Simulated event rate: 0.0000e+00 Hz
INFO::simtel_io_histograms(l310)::_rates_for_each_file::System trigger event rate: 0.0000e+00 ± 0.0000e+00 Hz
File ./tests/resources/run201_proton_za20deg_azm0deg_North_test_layout_test-prod.simtel.zst
System trigger rate (Hz): 0.0000e+00 ± 0.0000e+00 Hz
INFO::calculate_trigger_rate(l205)::main::Writing table 1 to /workdir/external/simtools/simtools-tests/simtools-output/calculate_trigger_rate/application-plots/run201_proton_za20deg_azm0deg_North_test_layout_test-prod.simtel.ecsv
Should at least catch the division by zero in simtel/simtel_io_histogram.py:607.
Maybe worth to add also a check for 'RunTimeWarning's in the output of all integration tests.
The reason for the unreasonable result might be resolved by the fixes I am introducing.
However, I agree that the test should include checking for warnings like this and also that the output is reasonable. The latter I will add to the production application now, so the same can be done here. Perhaps it's worth waiting a few days to have an example to follow.