Quality query falling back to default for ctapipe-optimize-event-selection

[BastienLacave]

when running ctapipe-optimize-event-selection I have the following error:

2025-02-13 09:33:06,171 ERROR [ctapipe.ctapipe-optimize-event-selection] (tool.run): Caught unexpected exception: Error evaluating expression '('multiplicity 4', 'np.count_nonzero(HillasReconstructor_telescopes,axis=1) >= 4')': name 'HillasReconstructor_telescopes' is not defined Traceback (most recent call last): File "/home/blacave/CTLearn/Software/ctapipe/src/ctapipe/core/tool.py", line 439, in run self.start() File "/home/blacave/CTLearn/Software/ctapipe/src/ctapipe/tools/optimize_event_selection.py", line 150, in start events, count, meta = loader.load_preselected_events( File "/home/blacave/CTLearn/Software/ctapipe/src/ctapipe/irf/preprocessing.py", line 227, in load_preselected_events selected = events[self.epp.quality_query.get_table_mask(events)] File "/home/blacave/CTLearn/Software/ctapipe/src/ctapipe/core/qualityquery.py", line 138, in get_table_mask for i, res in enumerate(self.engine(table), start=1): File "/home/blacave/CTLearn/Software/ctapipe/src/ctapipe/core/expression_engine.py", line 44, in __call__ raise ExpressionError( ctapipe.core.expression_engine.ExpressionError: Error evaluating expression '('multiplicity 4', 'np.count_nonzero(HillasReconstructor_telescopes,axis=1) >= 4')': name 'HillasReconstructor_telescopes' is not defined

But in my config I specify

EventPreProcessor:
  energy_reconstructor: "CTLearn"
  gammaness_classifier: "CTLearn"
  geometry_reconstructor: "CTLearn"
  quality_criteria:
    - ["multiplicity", "np.count_nonzero(CTLearn_telescopes,axis=1) == 2"]
    - ["valid classifier", "CTLearn_is_valid"]
    - ["valid geom reco", "CTLearn_is_valid"]
    - ["valid energy reco", "CTLearn_is_valid"]`

public-conf.txt

[Hckjs]

Hey Bastien,
I think you need to set the quality_criteria trait of the EventQualityQuery Component to make the config system work because its not a direct trait of the EventPreprocessor Component, e.g.:

EventPreprocessor: 
    energy_reconstructor: "CTLearn" 
    gammaness_classifier: "CTLearn" 
    geometry_reconstructor: "CTLearn"

    EventQualityQuery:
        quality_criteria: 
            - ["multiplicity", "np.count_nonzero(CTLearn_telescopes,axis=1) == 2"] 
            - ["valid classifier", "CTLearn_is_valid"]
            - ["valid geom reco", "CTLearn_is_valid"] 
            - ["valid energy reco", "CTLearn_is_valid"]

Just like this example