Current implementation:
thresholds = {'presence_ratio': {'min': 0.9}, # some places say this threshold is 0.8 but official documentation says 0.9
'isi_violations_ratio': {'max': 0.5},
'amplitude_cutoff': {'max': 0.1}
}
aind_labels = sc.threshold_metrics_label_units(metrics, thresholds, column_name='aind_label')
I don't know if it's just me but thinking about 'max' or 'min' requires some mental gymnastics and I have to keep double checking. Why not use 'greater'/'less'? E.g.,
thresholds = {'presence_ratio': {'greater': 0.9},
'isi_violations_ratio': {'less': 0.5},
'amplitude_cutoff': {'less': 0.1}
}
it reads nicely and It's early enough that a change would not require backwards-compatibility with max/min (only perhaps changing docs & examples).
Current implementation:
I don't know if it's just me but thinking about 'max' or 'min' requires some mental gymnastics and I have to keep double checking. Why not use 'greater'/'less'? E.g.,
it reads nicely and It's early enough that a change would not require backwards-compatibility with max/min (only perhaps changing docs & examples).