mergeMeasurements.py

Go to the documentation of this file.

#!/usr/bin/env python
#
# LSST Data Management System
# Copyright 2008-2015 AURA/LSST.
#
# This product includes software developed by the
# LSST Project (http://www.lsst.org/).
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the LSST License Statement and
# the GNU General Public License along with this program.  If not,
# see <https://www.lsstcorp.org/LegalNotices/>.
#
import numpy

from .multiBandUtils import (MergeSourcesRunner, _makeGetSchemaCatalogs, makeMergeArgumentParser,
                             getInputSchema, getShortFilterName, readCatalog)

import lsst.afw.table as afwTable
import lsst.pex.config as pexConfig
import lsst.pipe.base as pipeBase


class MergeMeasurementsConfig(pipeBase.PipelineTaskConfig):
    """!
    @anchor MergeMeasurementsConfig_

    @brief Configuration parameters for the MergeMeasurementsTask
    """
    # Gen 3 options
    inputSchema = pipeBase.InitInputDatasetField(
        doc="Schema for the input measurement catalogs.",
        nameTemplate="{inputCoaddName}Coadd_meas_schema",
        storageClass="SourceCatalog",
    )
    outputSchema = pipeBase.InitOutputDatasetField(
        doc="Schema for the output merged measurement catalog.",
        nameTemplate="{outputCoaddName}Coadd_ref_schema",
        storageClass="SourceCatalog",
    )
    catalogs = pipeBase.InputDatasetField(
        doc="Input catalogs to merge.",
        nameTemplate="{inputCoaddName}Coadd_meas",
        scalar=False,
        storageClass="SourceCatalog",
        dimensions=["AbstractFilter", "SkyMap", "Tract", "Patch"],
    )
    mergedCatalog = pipeBase.OutputDatasetField(
        doc="Output merged catalog.",
        nameTemplate="{outputCoaddName}Coadd_ref",
        scalar=True,
        storageClass="SourceCatalog",
        dimensions=["SkyMap", "Tract", "Patch"],
    )
    # Task configuration options
    pseudoFilterList = pexConfig.ListField(
        dtype=str,
        default=["sky"],
        doc="Names of filters which may have no associated detection\n"
        "(N.b. should include MergeDetectionsConfig.skyFilterName)"
    )
    snName = pexConfig.Field(
        dtype=str,
        default="base_PsfFlux",
        doc="Name of flux measurement for calculating the S/N when choosing the reference band."
    )
    minSN = pexConfig.Field(
        dtype=float,
        default=10.,
        doc="If the S/N from the priority band is below this value (and the S/N "
        "is larger than minSNDiff compared to the priority band), use the band with "
        "the largest S/N as the reference band."
    )
    minSNDiff = pexConfig.Field(
        dtype=float,
        default=3.,
        doc="If the difference in S/N between another band and the priority band is larger "
        "than this value (and the S/N in the priority band is less than minSN) "
        "use the band with the largest S/N as the reference band"
    )
    flags = pexConfig.ListField(
        dtype=str,
        doc="Require that these flags, if available, are not set",
        default=["base_PixelFlags_flag_interpolatedCenter", "base_PsfFlux_flag",
                 "ext_photometryKron_KronFlux_flag", "modelfit_CModel_flag", ]
    )
    priorityList = pexConfig.ListField(
        dtype=str,
        default=[],
        doc="Priority-ordered list of bands for the merge."
    )
    coaddName = pexConfig.Field(
        dtype=str,
        default="deep",
        doc="Name of coadd"
    )

    def validate(self):
        super().validate()
        if len(self.priorityList) == 0:
            raise RuntimeError("No priority list provided")

    def setDefaults(self):
        super().setDefaults()
        self.formatTemplateNames({"inputCoaddName": "deep",
                                  "outputCoaddName": "deep"})
        self.quantum.dimensions = ("SkyMap", "Tract", "Patch")




class MergeMeasurementsTask(pipeBase.PipelineTask, pipeBase.CmdLineTask):
    r"""!
    @anchor MergeMeasurementsTask_

    @brief Merge measurements from multiple bands

    @section pipe_tasks_multiBand_Contents Contents

      - @ref pipe_tasks_multiBand_MergeMeasurementsTask_Purpose
      - @ref pipe_tasks_multiBand_MergeMeasurementsTask_Initialize
      - @ref pipe_tasks_multiBand_MergeMeasurementsTask_Run
      - @ref pipe_tasks_multiBand_MergeMeasurementsTask_Config
      - @ref pipe_tasks_multiBand_MergeMeasurementsTask_Debug
      - @ref pipe_tasks_multiband_MergeMeasurementsTask_Example

    @section pipe_tasks_multiBand_MergeMeasurementsTask_Purpose Description

    Command-line task that merges measurements from multiple bands.

    Combines consistent (i.e. with the same peaks and footprints) catalogs of sources from multiple filter
    bands to construct a unified catalog that is suitable for driving forced photometry. Every source is
    required to have centroid, shape and flux measurements in each band.

      @par Inputs:
        deepCoadd_meas{tract,patch,filter}: SourceCatalog
      @par Outputs:
        deepCoadd_ref{tract,patch}: SourceCatalog
      @par Data Unit:
        tract, patch

    MergeMeasurementsTask subclasses @ref CmdLineTask_ "CmdLineTask".

    @section pipe_tasks_multiBand_MergeMeasurementsTask_Initialize       Task initialization

    @copydoc \_\_init\_\_

    @section pipe_tasks_multiBand_MergeMeasurementsTask_Run       Invoking the Task

    @copydoc run

    @section pipe_tasks_multiBand_MergeMeasurementsTask_Config       Configuration parameters

    See @ref MergeMeasurementsConfig_

    @section pipe_tasks_multiBand_MergeMeasurementsTask_Debug           Debug variables

    The @link lsst.pipe.base.cmdLineTask.CmdLineTask command line task@endlink interface supports a
    flag @c -d to import @b debug.py from your @c PYTHONPATH; see @ref baseDebug for more about @b debug.py
    files.

    MergeMeasurementsTask has no debug variables.

    @section pipe_tasks_multiband_MergeMeasurementsTask_Example A complete example
    of using MergeMeasurementsTask

    MergeMeasurementsTask is meant to be run after deblending & measuring sources in every band.
    The purpose of the task is to generate a catalog of sources suitable for driving forced photometry in
    coadds and individual exposures.
    Command-line usage of MergeMeasurementsTask expects a data reference to the coadds to be processed. A list
    of the available optional arguments can be obtained by calling mergeCoaddMeasurements.py with the `--help`
    command line argument:
    @code
    mergeCoaddMeasurements.py --help
    @endcode

    To demonstrate usage of the DetectCoaddSourcesTask in the larger context of multi-band processing, we
    will process HSC data in the [ci_hsc](https://github.com/lsst/ci_hsc) package. Assuming one has finished
    step 7 at @ref pipeTasks_multiBand, one may merge the catalogs generated after deblending and measuring
    as follows:
    @code
    mergeCoaddMeasurements.py $CI_HSC_DIR/DATA --id patch=5,4 tract=0 filter=HSC-I^HSC-R
    @endcode
    This will merge the HSC-I & HSC-R band catalogs. The results are written in
    `$CI_HSC_DIR/DATA/deepCoadd-results/`.
    """
    _DefaultName = "mergeCoaddMeasurements"
    ConfigClass = MergeMeasurementsConfig
    RunnerClass = MergeSourcesRunner
    inputDataset = "meas"
    outputDataset = "ref"
    getSchemaCatalogs = _makeGetSchemaCatalogs("ref")

    @classmethod
    def _makeArgumentParser(cls):
        return makeMergeArgumentParser(cls._DefaultName, cls.inputDataset)

    def getInputSchema(self, butler=None, schema=None):
        return getInputSchema(self, butler, schema)

    def getInitOutputDatasets(self):
        return {"outputSchema": afwTable.SourceCatalog(self.schema), }

    def adaptArgsAndRun(self, inputData, inputDataIds, outputDataIds, butler):
        catalogDict = {dataId['abstract_filter']: cat for dataId, cat in zip(inputDataIds['catalogs'],
                                                                             inputData['catalogs'])}
        inputData['catalogs'] = catalogDict

        return super().adaptArgsAndRun(inputData, inputDataIds, outputDataIds, butler)

    def __init__(self, butler=None, schema=None, initInputs=None, **kwargs):
        """!
        Initialize the task.

        @param[in] schema: the schema of the detection catalogs used as input to this one
        @param[in] butler: a butler used to read the input schema from disk, if schema is None

        The task will set its own self.schema attribute to the schema of the output merged catalog.
        """
        super().__init__(**kwargs)

        if initInputs is not None:
            inputSchema = initInputs['inputSchema'].schema
        else:
            inputSchema = self.getInputSchema(butler=butler, schema=schema)
        self.schemaMapper = afwTable.SchemaMapper(inputSchema, True)
        self.schemaMapper.addMinimalSchema(inputSchema, True)
        self.instFluxKey = inputSchema.find(self.config.snName + "_instFlux").getKey()
        self.instFluxErrKey = inputSchema.find(self.config.snName + "_instFluxErr").getKey()
        self.fluxFlagKey = inputSchema.find(self.config.snName + "_flag").getKey()

        self.flagKeys = {}
        for band in self.config.priorityList:
            short = getShortFilterName(band)
            outputKey = self.schemaMapper.editOutputSchema().addField(
                "merge_measurement_%s" % short,
                type="Flag",
                doc="Flag field set if the measurements here are from the %s filter" % band
            )
            peakKey = inputSchema.find("merge_peak_%s" % short).key
            footprintKey = inputSchema.find("merge_footprint_%s" % short).key
            self.flagKeys[band] = pipeBase.Struct(peak=peakKey, footprint=footprintKey, output=outputKey)
        self.schema = self.schemaMapper.getOutputSchema()

        self.pseudoFilterKeys = []
        for filt in self.config.pseudoFilterList:
            try:
                self.pseudoFilterKeys.append(self.schema.find("merge_peak_%s" % filt).getKey())
            except Exception as e:
                self.log.warn("merge_peak is not set for pseudo-filter %s: %s" % (filt, e))

        self.badFlags = {}
        for flag in self.config.flags:
            try:
                self.badFlags[flag] = self.schema.find(flag).getKey()
            except KeyError as exc:
                self.log.warn("Can't find flag %s in schema: %s" % (flag, exc,))

    def runDataRef(self, patchRefList):
        """!
        @brief Merge coadd sources from multiple bands. Calls @ref `run`.
        @param[in] patchRefList list of data references for each filter
        """
        catalogs = dict(readCatalog(self, patchRef) for patchRef in patchRefList)
        mergedCatalog = self.run(catalogs).mergedCatalog
        self.write(patchRefList[0], mergedCatalog)

    def run(self, catalogs):
        """!
        Merge measurement catalogs to create a single reference catalog for forced photometry

        @param[in] catalogs: the catalogs to be merged

        For parent sources, we choose the first band in config.priorityList for which the
        merge_footprint flag for that band is is True.

        For child sources, the logic is the same, except that we use the merge_peak flags.
        """
        # Put catalogs, filters in priority order
        orderedCatalogs = [catalogs[band] for band in self.config.priorityList if band in catalogs.keys()]
        orderedKeys = [self.flagKeys[band] for band in self.config.priorityList if band in catalogs.keys()]

        mergedCatalog = afwTable.SourceCatalog(self.schema)
        mergedCatalog.reserve(len(orderedCatalogs[0]))

        idKey = orderedCatalogs[0].table.getIdKey()
        for catalog in orderedCatalogs[1:]:
            if numpy.any(orderedCatalogs[0].get(idKey) != catalog.get(idKey)):
                raise ValueError("Error in inputs to MergeCoaddMeasurements: source IDs do not match")

        # This first zip iterates over all the catalogs simultaneously, yielding a sequence of one
        # record for each band, in priority order.
        for orderedRecords in zip(*orderedCatalogs):

            maxSNRecord = None
            maxSNFlagKeys = None
            maxSN = 0.
            priorityRecord = None
            priorityFlagKeys = None
            prioritySN = 0.
            hasPseudoFilter = False

            # Now we iterate over those record-band pairs, keeping track of the priority and the
            # largest S/N band.
            for inputRecord, flagKeys in zip(orderedRecords, orderedKeys):
                parent = (inputRecord.getParent() == 0 and inputRecord.get(flagKeys.footprint))
                child = (inputRecord.getParent() != 0 and inputRecord.get(flagKeys.peak))

                if not (parent or child):
                    for pseudoFilterKey in self.pseudoFilterKeys:
                        if inputRecord.get(pseudoFilterKey):
                            hasPseudoFilter = True
                            priorityRecord = inputRecord
                            priorityFlagKeys = flagKeys
                            break
                    if hasPseudoFilter:
                        break

                isBad = any(inputRecord.get(flag) for flag in self.badFlags)
                if isBad or inputRecord.get(self.fluxFlagKey) or inputRecord.get(self.instFluxErrKey) == 0:
                    sn = 0.
                else:
                    sn = inputRecord.get(self.instFluxKey)/inputRecord.get(self.instFluxErrKey)
                if numpy.isnan(sn) or sn < 0.:
                    sn = 0.
                if (parent or child) and priorityRecord is None:
                    priorityRecord = inputRecord
                    priorityFlagKeys = flagKeys
                    prioritySN = sn
                if sn > maxSN:
                    maxSNRecord = inputRecord
                    maxSNFlagKeys = flagKeys
                    maxSN = sn

            # If the priority band has a low S/N we would like to choose the band with the highest S/N as
            # the reference band instead.  However, we only want to choose the highest S/N band if it is
            # significantly better than the priority band.  Therefore, to choose a band other than the
            # priority, we require that the priority S/N is below the minimum threshold and that the
            # difference between the priority and highest S/N is larger than the difference threshold.
            #
            # For pseudo code objects we always choose the first band in the priority list.
            bestRecord = None
            bestFlagKeys = None
            if hasPseudoFilter:
                bestRecord = priorityRecord
                bestFlagKeys = priorityFlagKeys
            elif (prioritySN < self.config.minSN and (maxSN - prioritySN) > self.config.minSNDiff and
                  maxSNRecord is not None):
                bestRecord = maxSNRecord
                bestFlagKeys = maxSNFlagKeys
            elif priorityRecord is not None:
                bestRecord = priorityRecord
                bestFlagKeys = priorityFlagKeys

            if bestRecord is not None and bestFlagKeys is not None:
                outputRecord = mergedCatalog.addNew()
                outputRecord.assign(bestRecord, self.schemaMapper)
                outputRecord.set(bestFlagKeys.output, True)
            else:  # if we didn't find any records
                raise ValueError("Error in inputs to MergeCoaddMeasurements: no valid reference for %s" %
                                 inputRecord.getId())

        # more checking for sane inputs, since zip silently iterates over the smallest sequence
        for inputCatalog in orderedCatalogs:
            if len(mergedCatalog) != len(inputCatalog):
                raise ValueError("Mismatch between catalog sizes: %s != %s" %
                                 (len(mergedCatalog), len(orderedCatalogs)))

        return pipeBase.Struct(
            mergedCatalog=mergedCatalog
        )

    def write(self, patchRef, catalog):
        """!
        @brief Write the output.

        @param[in]  patchRef   data reference for patch
        @param[in]  catalog    catalog

        We write as the dataset provided by the 'outputDataset'
        class variable.
        """
        patchRef.put(catalog, self.config.coaddName + "Coadd_" + self.outputDataset)
        # since the filter isn't actually part of the data ID for the dataset we're saving,
        # it's confusing to see it in the log message, even if the butler simply ignores it.
        mergeDataId = patchRef.dataId.copy()
        del mergeDataId["filter"]
        self.log.info("Wrote merged catalog: %s" % (mergeDataId,))

    def writeMetadata(self, dataRefList):
        """!
        @brief No metadata to write, and not sure how to write it for a list of dataRefs.
        """
        pass