lsst.pipe.tasks.setPrimaryFlags Namespace Reference

Classes
class	SetPrimaryFlagsConfig
class	SetPrimaryFlagsTask

Functions
def	getPatchInner (sources, patchInfo)
def	getTractInner (sources, tractInfo, skyMap)
def	getPseudoSources (sources, pseudoFilterList, schema, log)
def	getDeblendPrimaryFlags (sources)

Function Documentation

getDeblendPrimaryFlags()

def lsst.pipe.tasks.setPrimaryFlags.getDeblendPrimaryFlags

(

sources

)

Get flags generated by the deblender

scarlet is different than meas_deblender in that it is not
(necessarily) flux conserving. For consistency in scarlet,
all of the parents with only a single child (isolated sources)
need to be deblended. This creates a question: which type
of isolated source should we make measurements on, the
undeblended "parent" or the deblended child?
For that reason we distinguish between a DeblendedSource,
which is a source that has no children and uses the
isolated parents, and a DeblendedModelSource, which uses
the scarlet models for both isolated and blended sources.
In the case of meas_deblender, DeblendedModelSource is
`None` because it is not contained in the output catalog.

Parameters
----------
sources : `lsst.afw.table.SourceCatalog`
    A sourceCatalog that has already been deblended using
    either meas_extensions_scarlet or meas_deblender.

Returns
-------
fromBlend : array-like of `bool`
    True for each source modeled by the deblender from a `Peak`
    in a parent footprint that contained at least one other `Peak`.
    While these models can be approximated as isolated,
    and measurements are made on them as if that's the case,
    we know deblending to introduce biases in the shape and centroid
    of objects and it is important to know that the sources that these
    models are based on are all bleneded in the true image.
isIsolated : array-like of `bool`
    True for isolated sources, regardless of whether or not they
    were modeled by the deblender.
isDeblendedSource : array-like of `bool`
    True for each source that is a "DeblendedSource" as defined above.
isDeblendedModelSource : array-like of `bool`
    True for each source that is a "DeblendedSourceModel"
    as defined above.

Definition at line 122 of file setPrimaryFlags.py.

def getDeblendPrimaryFlags(sources):
    """Get flags generated by the deblender
 
    scarlet is different than meas_deblender in that it is not
    (necessarily) flux conserving. For consistency in scarlet,
    all of the parents with only a single child (isolated sources)
    need to be deblended. This creates a question: which type
    of isolated source should we make measurements on, the
    undeblended "parent" or the deblended child?
    For that reason we distinguish between a DeblendedSource,
    which is a source that has no children and uses the
    isolated parents, and a DeblendedModelSource, which uses
    the scarlet models for both isolated and blended sources.
    In the case of meas_deblender, DeblendedModelSource is
    `None` because it is not contained in the output catalog.
 
    Parameters
    ----------
    sources : `lsst.afw.table.SourceCatalog`
        A sourceCatalog that has already been deblended using
        either meas_extensions_scarlet or meas_deblender.
 
    Returns
    -------
    fromBlend : array-like of `bool`
        True for each source modeled by the deblender from a `Peak`
        in a parent footprint that contained at least one other `Peak`.
        While these models can be approximated as isolated,
        and measurements are made on them as if that's the case,
        we know deblending to introduce biases in the shape and centroid
        of objects and it is important to know that the sources that these
        models are based on are all bleneded in the true image.
    isIsolated : array-like of `bool`
        True for isolated sources, regardless of whether or not they
        were modeled by the deblender.
    isDeblendedSource : array-like of `bool`
        True for each source that is a "DeblendedSource" as defined above.
    isDeblendedModelSource : array-like of `bool`
        True for each source that is a "DeblendedSourceModel"
        as defined above.
    """
    nChildKey = "deblend_nChild"
    nChild = sources[nChildKey]
    parent = sources["parent"]
 
    if "deblend_scarletFlux" in sources.schema:
        # The number of peaks in the sources footprint.
        # This (may be) different than nChild,
        # the number of deblended sources in the catalog,
        # because some peaks might have been culled during deblending.
        nPeaks = sources["deblend_nPeaks"]
        parentNChild = sources["deblend_parentNChild"]
        # It is possible for a catalog to contain a hierarchy of sources,
        # so we mark the leaves (end nodes of the hierarchy tree with no
        # children).
        isLeaf = nPeaks == 1
        fromBlend = parentNChild > 1
        isIsolated = isLeaf & ((parent == 0) | parentNChild == 1)
        isDeblendedSource = (fromBlend & isLeaf) | (isIsolated & (parent == 0))
        isDeblendedModelSource = (parent != 0) & isLeaf
    else:
        # Set the flags for meas_deblender
        fromBlend = parent != 0
        isIsolated = (nChild == 0) & (parent == 0)
        isDeblendedSource = nChild == 0
        isDeblendedModelSource = None
    return fromBlend, isIsolated, isDeblendedSource, isDeblendedModelSource
 
 

getPatchInner()

def lsst.pipe.tasks.setPrimaryFlags.getPatchInner	(		sources,
			patchInfo
	)

Set a flag for each source if it is in the innerBBox of a patch.

Parameters
----------
sources : `lsst.afw.table.SourceCatalog`
    A sourceCatalog with pre-calculated centroids.
patchInfo : `lsst.skymap.PatchInfo`
    Information about a `SkyMap` `Patch`.

Returns
--------
isPatchInner : array-like of `bool`
    `True` for each source that has a centroid
    in the inner region of a patch.

Definition at line 29 of file setPrimaryFlags.py.

def getPatchInner(sources, patchInfo):
    """Set a flag for each source if it is in the innerBBox of a patch.
 
    Parameters
    ----------
    sources : `lsst.afw.table.SourceCatalog`
        A sourceCatalog with pre-calculated centroids.
    patchInfo : `lsst.skymap.PatchInfo`
        Information about a `SkyMap` `Patch`.
 
    Returns
    --------
    isPatchInner : array-like of `bool`
        `True` for each source that has a centroid
        in the inner region of a patch.
    """
    # Extract the centroid position for all the sources
    x = sources["slot_Centroid_x"]
    y = sources["slot_Centroid_y"]
    centroidFlag = sources["slot_Centroid_flag"]
 
    # set inner flags for each source and set primary flags for
    # sources with no children (or all sources if deblend info not available)
    innerFloatBBox = Box2D(patchInfo.getInnerBBox())
    inInner = innerFloatBBox.contains(x, y)
 
    # When the centroider fails, we can still fall back to the peak,
    # but we don't trust that quite as much -
    # so we use a slightly smaller box for the patch comparison.
    shrunkInnerFloatBBox = Box2D(innerFloatBBox)
    shrunkInnerFloatBBox.grow(-1)
    inShrunkInner = shrunkInnerFloatBBox.contains(x, y)
 
    # Flag sources contained in the inner region of a patch
    isPatchInner = (centroidFlag & inShrunkInner) | (~centroidFlag & inInner)
    return isPatchInner
 
 

getPseudoSources()

def lsst.pipe.tasks.setPrimaryFlags.getPseudoSources	(		sources,
			pseudoFilterList,
			schema,
			log
	)

Get a flag that marks pseudo sources.

Some categories of sources, for example sky objects,
are not really detected sources and should not be considered primary
sources.

Parameters
----------
sources : `lsst.afw.table.SourceCatalog`
    The catalog of sources for which to identify "pseudo"
    (e.g. sky) objects.
pseudoFilterList : `list` of `str`
    Names of filters which should never be primary

Returns
-------
isPseudo : array-like of `bool`
    True for each source that is a pseudo source.
    Note: to remove pseudo sources use `~isPseudo`.

Definition at line 90 of file setPrimaryFlags.py.

def getPseudoSources(sources, pseudoFilterList, schema, log):
    """Get a flag that marks pseudo sources.
 
    Some categories of sources, for example sky objects,
    are not really detected sources and should not be considered primary
    sources.
 
    Parameters
    ----------
    sources : `lsst.afw.table.SourceCatalog`
        The catalog of sources for which to identify "pseudo"
        (e.g. sky) objects.
    pseudoFilterList : `list` of `str`
        Names of filters which should never be primary
 
    Returns
    -------
    isPseudo : array-like of `bool`
        True for each source that is a pseudo source.
        Note: to remove pseudo sources use `~isPseudo`.
    """
    # Filter out sources that should never be primary
    isPseudo = np.zeros(len(sources), dtype=bool)
    for filt in pseudoFilterList:
        try:
            pseudoFilterKey = schema.find("merge_peak_%s" % filt).getKey()
            isPseudo |= sources[pseudoFilterKey]
        except KeyError:
            log.warning("merge_peak is not set for pseudo-filter %s", filt)
    return isPseudo
 
 

getTractInner()

def lsst.pipe.tasks.setPrimaryFlags.getTractInner	(		sources,
			tractInfo,
			skyMap
	)

Set a flag for each source that the skyMap includes in tractInfo.

Parameters
----------
sources : `lsst.afw.table.SourceCatalog`
    A sourceCatalog with pre-calculated centroids.
tractInfo : `lsst.skymap.TractInfo`
    Tract object
skyMap : `lsst.skymap.BaseSkyMap`
    Sky tessellation object

Returns
-------
isTractInner : array-like of `bool`
    True if the skyMap.findTract method returns
    the same tract as tractInfo.

Definition at line 67 of file setPrimaryFlags.py.

def getTractInner(sources, tractInfo, skyMap):
    """Set a flag for each source that the skyMap includes in tractInfo.
 
    Parameters
    ----------
    sources : `lsst.afw.table.SourceCatalog`
        A sourceCatalog with pre-calculated centroids.
    tractInfo : `lsst.skymap.TractInfo`
        Tract object
    skyMap : `lsst.skymap.BaseSkyMap`
        Sky tessellation object
 
    Returns
    -------
    isTractInner : array-like of `bool`
        True if the skyMap.findTract method returns
        the same tract as tractInfo.
    """
    tractId = tractInfo.getId()
    isTractInner = np.array([skyMap.findTract(s.getCoord()).getId() == tractId for s in sources])
    return isTractInner