ibllib.qc.oneqc_metrics

Classes

ONEQC

class ONEQC(eid, one=None, bpod_ntrials=None, lazy=False)

Bases: ibllib.qc.base.QC

compute()

load_nDatasetTypes()

17. Proportion of datasetTypes extracted Variable name: nDatasetTypes Metric: len(one.load(eid, offline=True, download_only=True)) / nExpetedDatasetTypes (hardcoded per task?)

load_dstype_qc_metrics(dstype_name: str) → dict

Returns dict to update to metrics or criteria frame Metrics:

_length = number of trials in ONE dstype _count = number of nans in dstype

Criteria:

length / bpod number of trials count / length of dstype

NB: Makes sense for dstypes that should have one value per trial and where nans are informative of failures Other dstypes will have nans because of the contingency of the trial, e.g. if contrastLeft has a nan it means the contrast was on the right.