FgcmBuildStarsTask¶
FgcmBuildStarsTask
finds all the single-visit sources in a repository (or a subset based on command-line parameters) and extracts all the potential photometric calibration stars for input into fgcm.
This task additionally uses fgcm to match star observations into unique stars, and performs as much cleaning of the input catalog as possible.
A tutorial on the steps of running fgcmcal
are found in the cookbook.
This fgcmcal
task runs on calexp source catalogs from visits constrained by the --id
parameter on the command line.
Best results are obtained when fgcmcal
is run with full visits.
If sourceTable_visit
parquet tables are available, use of FgcmBuildStarsTableTask is recommended in place of this task.
In Gen2, due to limitations of the Gen2 Butler, optimal performance is obtained by specifying a single “reference” ccd on the command line (e.g. ccd=13
) and setting the config variable checkAllCcds = True
(which is the default).
The alternative is to specify all the desired CCDs and set checkAllCcds = False
, e.g., ccd=0..8^10..103
.
However, this is slower than the first option, and the improvement in speed in the first option is greater the more visits are specified.
If instead you want to process all the visits in a rerun selected by filter, field, or some other dataid field, then by using a reference ccd and setting checkAllCcds = True
you can speed things up by a factor of approximately 100 relative to the alternative (naming CCDs specifically).
Be aware that if a visit does not have a calexp
available with the given reference CCD then it will be skipped.
If this is possibly an issue, multiple reference ccds can be specified on the command line, although performance will degrade the more are specified.
At the current time, fgcmcal
does not support Gen3.
This is the second task in a typical fgcmcal
processing chain.
The first is FgcmMakeLutTask, the third is FgcmFitCycleTask, and the fourth is FgcmOutputProductsTask.
FgcmBuildStarsTask
is available as a command-line task, fgcmBuildStars.py.
Processing summary¶
FgcmBuildStarsTask
runs this sequence of operations:
Finds unique visits and collates visit metadata, including exposure time, pointing, typical psf size, background level.
Reads in all sources, selecting good stars according to the chosen source selector.
Matches sources internally across bands to get a unique multi-band list of possible calibration stars.
Matches possible calibration stars to a reference catalog.
All results are stored in the output repo
fgcm-process
directory.
fgcmBuildStars.py command-line interface¶
fgcmBuildStars.py REPOPATH [@file [@file2 ...]] [--output OUTPUTREPO | --rerun RERUN] [--id] [other options]
Key arguments:
REPOPATH
The input Butler repository’s URI or file path.
Key options:
--id
:The data IDs to process.
See also
See Command-line task argument reference for details and additional options.
Python API summary¶
from lsst.fgcmcal.fgcmBuildStars import FgcmBuildStarsTask
-
class
(butler=None, initInputs=None, **kwargs)FgcmBuildStarsTask
Build stars for the FGCM global calibration, using src catalogs
...
-
attribute
config
Access configuration fields and retargetable subtasks.
-
method
(butler, dataRefs)runDataRef
Cross-match and make star list for FGCM Input
...
See also
See the FgcmBuildStarsTask
API reference for complete details.
Butler datasets¶
When run as the fgcmBuildStars.py
command-line task, or directly through the runDataRef
method, FgcmBuildStarsTask
obtains datasets from the input Butler data repository and persists outputs to the output Butler data repository.
Note that configurations for FgcmBuildStarsTask
, and its subtasks, affect what datasets are persisted and what their content is.
Input datasets¶
src
Full-depth source catalog (
lsst.afw.table
) produced byProcessCcdTask
calexp
Calibrated exposure produced by
ProcessCcdTask
(for exposure metadata)fgcmLookupTable
FGCM look-up table produced by FgcmMakeLutTask
Output datasets¶
fgcmVisitCatalog
Catalog (
lsst.afw.table
) of visit metadatafgcmStarObservations
Catalog of star observations
fgcmStarIds
Catalog of unique star ids, positions, and number of observations
fgcmStarIndices
Catalog of indices linking unique star ids to star observations
fgcmReferenceStars
Catalog of reference stars matched to unique star ids.
Retargetable subtasks¶
fgcmLoadReferenceCatalog¶
- Default
lsst.fgcmcal.fgcmLoadReferenceCatalog.FgcmLoadReferenceCatalogTask
- Field type
FGCM reference object loader
sourceSelector¶
- Default
'science'
- Field type
Single-selection
RegistryField
- Choices
'science'
lsst.meas.algorithms.sourceSelector.ScienceSourceSelectorTask
'references'
lsst.meas.algorithms.sourceSelector.ReferenceSourceSelectorTask
'objectSize'
lsst.meas.algorithms.objectSizeStarSelector.ObjectSizeStarSelectorTask
'flagged'
lsst.meas.algorithms.flaggedSourceSelector.FlaggedSourceSelectorTask
'astrometry'
lsst.meas.algorithms.astrometrySourceSelector.AstrometrySourceSelectorTask
'matcher'
lsst.meas.algorithms.matcherSourceSelector.MatcherSourceSelectorTask
'diaCatalog'
lsst.ip.diffim.diaCatalogSourceSelector.DiaCatalogSourceSelectorTask
How to select sources
Configuration fields¶
apertureInnerInstFluxField¶
Full name of instFlux field that contains inner aperture flux for aperture correction proxy
apertureOuterInstFluxField¶
Full name of instFlux field that contains outer aperture flux for aperture correction proxy
ccdDataRefName¶
dataRef name for the ‘ccd’ field, usually ‘ccd’ or ‘detector’.
checkAllCcds¶
Check repo for all CCDs for each visit specified. To be used when the full set of ids (visit/ccd) are not specified on the command line. For Gen2, specifying one ccd and setting checkAllCcds=True is significantly faster than the alternatives.
coarseNside¶
Healpix coarse Nside for partitioning matches
densityCutMaxPerPixel¶
Density cut number of stars per pixel
densityCutNside¶
Density cut healpix nside
doApplyWcsJacobian¶
Apply the jacobian of the WCS to the star observations prior to fit?
doModelErrorsWithBackground¶
Model flux errors with background term?
doReferenceMatches¶
Match reference catalog as additional constraint on calibration
doSubtractLocalBackground¶
Subtract the local background before performing calibration? This is only supported for circular aperture calibration fluxes.
filterMap¶
Mapping from ‘filterName’ to band. Deprecated: This field is no longer used, and has been deprecated by DM-28088. It will be removed after v22. Use physicalFilterMap instead.
instFluxField¶
Faull name of the source instFlux field to use, including ‘instFlux’. The associated flag will be implicitly included in badFlags
isolationRadius¶
Isolation radius (arcseconds)
localBackgroundFluxField¶
Full name of the local background instFlux field to use.
matchNside¶
Healpix Nside for matching
matchRadius¶
Match radius (arcseconds)
minPerBand¶
Minimum observations per band
nVisitsPerCheckpoint¶
Number of visits read between checkpoints
physicalFilterMap¶
Mapping from ‘physicalFilter’ to band.
primaryBands¶
Bands for ‘primary’ star matches. A star must be observed in one of these bands to be considered as a calibration star.
psfCandidateName¶
Name of field with psf candidate flag for propagation
referenceCCD¶
Reference CCD for scanning visits
requiredBands¶
Bands required for each star
visitDataRefName¶
dataRef name for the ‘visit’ field, usually ‘visit’.
Examples¶
See the cookbook for worked examples.