lsst.cp.pipe._lookupStaticCalibration Namespace Reference

Functions
def	lookupStaticCalibration (datasetType, registry, quantumDataId, collections)

Function Documentation

lookupStaticCalibration()

def lsst.cp.pipe._lookupStaticCalibration.lookupStaticCalibration	(		datasetType,
			registry,
			quantumDataId,
			collections
	)

A lookup function override for QuantumGraph generation that allows a
PipelineTask to have an input dataset (usually a camera) that is formally a
calibration with a validity range, without having a temporal data ID for
the lookup, by asserting that there is in fact only dataset for all time.

Parameters
----------
datasetType : `lsst.daf.butler.DatasetType`
    Dataset type to look up.
registry : `lsst.daf.butler.Registry`
    Registry for the data repository being searched.
quantumDataId : `lsst.daf.butler.DataCoordinate`
    Data ID for the quantum of the task this dataset will be passed to.
    This must include an "instrument" key, and should also include any
    keys that are present in ``datasetType.dimensions``.  If it has an
    ``exposure`` or ``visit`` key, that's a sign that this function is
    not actually needed, as those come with the temporal information that
    would allow a real validity-range lookup.
collections : `lsst.daf.butler.registry.CollectionSearch`
    Collections passed by the user when generating a QuantumGraph.  Ignored
    by this function (see notes below).

Returns
-------
refs : `list` [ `DatasetRef` ]
    A zero- or single-element list containing the matching dataset, if one
    was found.

Notes
-----
This works by looking in the `~CollectionType.RUN` collection
that `lsst.obs.base.Instrument.writeCuratedCalibrations` (currently!) uses,
instead of the collections passed into it.  This may be considered
surprising by users (but will usually go unnoticed because the dataset
returned _is_ actually in those given input colllections, too).  It may
stop working entirely once we have data repositories with multiple
calibration collections; a better workaround or a more principled change
to the PipelineTasks that use this function (which are by definition asking
for something ill-defined) will ultimately be needed.

Definition at line 27 of file _lookupStaticCalibration.py.

def lookupStaticCalibration(datasetType, registry, quantumDataId, collections):
    """A lookup function override for QuantumGraph generation that allows a
    PipelineTask to have an input dataset (usually a camera) that is formally a
    calibration with a validity range, without having a temporal data ID for
    the lookup, by asserting that there is in fact only dataset for all time.
 
    Parameters
    ----------
    datasetType : `lsst.daf.butler.DatasetType`
        Dataset type to look up.
    registry : `lsst.daf.butler.Registry`
        Registry for the data repository being searched.
    quantumDataId : `lsst.daf.butler.DataCoordinate`
        Data ID for the quantum of the task this dataset will be passed to.
        This must include an "instrument" key, and should also include any
        keys that are present in ``datasetType.dimensions``.  If it has an
        ``exposure`` or ``visit`` key, that's a sign that this function is
        not actually needed, as those come with the temporal information that
        would allow a real validity-range lookup.
    collections : `lsst.daf.butler.registry.CollectionSearch`
        Collections passed by the user when generating a QuantumGraph.  Ignored
        by this function (see notes below).
 
    Returns
    -------
    refs : `list` [ `DatasetRef` ]
        A zero- or single-element list containing the matching dataset, if one
        was found.
 
    Notes
    -----
    This works by looking in the `~CollectionType.RUN` collection
    that `lsst.obs.base.Instrument.writeCuratedCalibrations` (currently!) uses,
    instead of the collections passed into it.  This may be considered
    surprising by users (but will usually go unnoticed because the dataset
    returned _is_ actually in those given input colllections, too).  It may
    stop working entirely once we have data repositories with multiple
    calibration collections; a better workaround or a more principled change
    to the PipelineTasks that use this function (which are by definition asking
    for something ill-defined) will ultimately be needed.
    """
    instrument = Instrument.fromName(quantumDataId["instrument"], registry)
    unboundedCollection = instrument.makeUnboundedCalibrationRunName()
    ref = registry.findDataset(datasetType, dataId=quantumDataId, collections=[unboundedCollection])
    if ref is None:
        return []
    else:
        return [ref]