Coverage for python/lsst/pipe/tasks/postprocess.py: 31%
805 statements
« prev ^ index » next coverage.py v6.4.1, created at 2022-06-14 12:35 +0000
« prev ^ index » next coverage.py v6.4.1, created at 2022-06-14 12:35 +0000
1# This file is part of pipe_tasks
2#
3# Developed for the LSST Data Management System.
4# This product includes software developed by the LSST Project
5# (https://www.lsst.org).
6# See the COPYRIGHT file at the top-level directory of this distribution
7# for details of code ownership.
8#
9# This program is free software: you can redistribute it and/or modify
10# it under the terms of the GNU General Public License as published by
11# the Free Software Foundation, either version 3 of the License, or
12# (at your option) any later version.
13#
14# This program is distributed in the hope that it will be useful,
15# but WITHOUT ANY WARRANTY; without even the implied warranty of
16# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17# GNU General Public License for more details.
18#
19# You should have received a copy of the GNU General Public License
20# along with this program. If not, see <https://www.gnu.org/licenses/>.
22import functools
23import pandas as pd
24from collections import defaultdict
25import logging
26import numpy as np
27import numbers
28import os
30import lsst.geom
31import lsst.pex.config as pexConfig
32import lsst.pipe.base as pipeBase
33import lsst.daf.base as dafBase
34from lsst.obs.base import ExposureIdInfo
35from lsst.pipe.base import connectionTypes
36import lsst.afw.table as afwTable
37from lsst.meas.base import SingleFrameMeasurementTask
38from lsst.pipe.base import CmdLineTask, ArgumentParser, DataIdContainer
39from lsst.coadd.utils.coaddDataIdContainer import CoaddDataIdContainer
40from lsst.daf.butler import DeferredDatasetHandle, DataCoordinate
41from lsst.skymap import BaseSkyMap
43from .parquetTable import ParquetTable
44from .multiBandUtils import makeMergeArgumentParser, MergeSourcesRunner
45from .functors import CompositeFunctor, Column
47log = logging.getLogger(__name__)
50def flattenFilters(df, noDupCols=['coord_ra', 'coord_dec'], camelCase=False, inputBands=None):
51 """Flattens a dataframe with multilevel column index
52 """
53 newDf = pd.DataFrame()
54 # band is the level 0 index
55 dfBands = df.columns.unique(level=0).values
56 for band in dfBands:
57 subdf = df[band]
58 columnFormat = '{0}{1}' if camelCase else '{0}_{1}'
59 newColumns = {c: columnFormat.format(band, c)
60 for c in subdf.columns if c not in noDupCols}
61 cols = list(newColumns.keys())
62 newDf = pd.concat([newDf, subdf[cols].rename(columns=newColumns)], axis=1)
64 # Band must be present in the input and output or else column is all NaN:
65 presentBands = dfBands if inputBands is None else list(set(inputBands).intersection(dfBands))
66 # Get the unexploded columns from any present band's partition
67 noDupDf = df[presentBands[0]][noDupCols]
68 newDf = pd.concat([noDupDf, newDf], axis=1)
69 return newDf
72class WriteObjectTableConnections(pipeBase.PipelineTaskConnections,
73 defaultTemplates={"coaddName": "deep"},
74 dimensions=("tract", "patch", "skymap")):
75 inputCatalogMeas = connectionTypes.Input(
76 doc="Catalog of source measurements on the deepCoadd.",
77 dimensions=("tract", "patch", "band", "skymap"),
78 storageClass="SourceCatalog",
79 name="{coaddName}Coadd_meas",
80 multiple=True
81 )
82 inputCatalogForcedSrc = connectionTypes.Input(
83 doc="Catalog of forced measurements (shape and position parameters held fixed) on the deepCoadd.",
84 dimensions=("tract", "patch", "band", "skymap"),
85 storageClass="SourceCatalog",
86 name="{coaddName}Coadd_forced_src",
87 multiple=True
88 )
89 inputCatalogRef = connectionTypes.Input(
90 doc="Catalog marking the primary detection (which band provides a good shape and position)"
91 "for each detection in deepCoadd_mergeDet.",
92 dimensions=("tract", "patch", "skymap"),
93 storageClass="SourceCatalog",
94 name="{coaddName}Coadd_ref"
95 )
96 outputCatalog = connectionTypes.Output(
97 doc="A vertical concatenation of the deepCoadd_{ref|meas|forced_src} catalogs, "
98 "stored as a DataFrame with a multi-level column index per-patch.",
99 dimensions=("tract", "patch", "skymap"),
100 storageClass="DataFrame",
101 name="{coaddName}Coadd_obj"
102 )
105class WriteObjectTableConfig(pipeBase.PipelineTaskConfig,
106 pipelineConnections=WriteObjectTableConnections):
107 engine = pexConfig.Field(
108 dtype=str,
109 default="pyarrow",
110 doc="Parquet engine for writing (pyarrow or fastparquet)"
111 )
112 coaddName = pexConfig.Field(
113 dtype=str,
114 default="deep",
115 doc="Name of coadd"
116 )
119class WriteObjectTableTask(CmdLineTask, pipeBase.PipelineTask):
120 """Write filter-merged source tables to parquet
121 """
122 _DefaultName = "writeObjectTable"
123 ConfigClass = WriteObjectTableConfig
124 RunnerClass = MergeSourcesRunner
126 # Names of table datasets to be merged
127 inputDatasets = ('forced_src', 'meas', 'ref')
129 # Tag of output dataset written by `MergeSourcesTask.write`
130 outputDataset = 'obj'
132 def __init__(self, butler=None, schema=None, **kwargs):
133 # It is a shame that this class can't use the default init for CmdLineTask
134 # But to do so would require its own special task runner, which is many
135 # more lines of specialization, so this is how it is for now
136 super().__init__(**kwargs)
138 def runDataRef(self, patchRefList):
139 """!
140 @brief Merge coadd sources from multiple bands. Calls @ref `run` which must be defined in
141 subclasses that inherit from MergeSourcesTask.
142 @param[in] patchRefList list of data references for each filter
143 """
144 catalogs = dict(self.readCatalog(patchRef) for patchRef in patchRefList)
145 dataId = patchRefList[0].dataId
146 mergedCatalog = self.run(catalogs, tract=dataId['tract'], patch=dataId['patch'])
147 self.write(patchRefList[0], ParquetTable(dataFrame=mergedCatalog))
149 def runQuantum(self, butlerQC, inputRefs, outputRefs):
150 inputs = butlerQC.get(inputRefs)
152 measDict = {ref.dataId['band']: {'meas': cat} for ref, cat in
153 zip(inputRefs.inputCatalogMeas, inputs['inputCatalogMeas'])}
154 forcedSourceDict = {ref.dataId['band']: {'forced_src': cat} for ref, cat in
155 zip(inputRefs.inputCatalogForcedSrc, inputs['inputCatalogForcedSrc'])}
157 catalogs = {}
158 for band in measDict.keys():
159 catalogs[band] = {'meas': measDict[band]['meas'],
160 'forced_src': forcedSourceDict[band]['forced_src'],
161 'ref': inputs['inputCatalogRef']}
162 dataId = butlerQC.quantum.dataId
163 df = self.run(catalogs=catalogs, tract=dataId['tract'], patch=dataId['patch'])
164 outputs = pipeBase.Struct(outputCatalog=df)
165 butlerQC.put(outputs, outputRefs)
167 @classmethod
168 def _makeArgumentParser(cls):
169 """Create a suitable ArgumentParser.
171 We will use the ArgumentParser to get a list of data
172 references for patches; the RunnerClass will sort them into lists
173 of data references for the same patch.
175 References first of self.inputDatasets, rather than
176 self.inputDataset
177 """
178 return makeMergeArgumentParser(cls._DefaultName, cls.inputDatasets[0])
180 def readCatalog(self, patchRef):
181 """Read input catalogs
183 Read all the input datasets given by the 'inputDatasets'
184 attribute.
186 Parameters
187 ----------
188 patchRef : `lsst.daf.persistence.ButlerDataRef`
189 Data reference for patch
191 Returns
192 -------
193 Tuple consisting of band name and a dict of catalogs, keyed by
194 dataset name
195 """
196 band = patchRef.get(self.config.coaddName + "Coadd_filter", immediate=True).bandLabel
197 catalogDict = {}
198 for dataset in self.inputDatasets:
199 catalog = patchRef.get(self.config.coaddName + "Coadd_" + dataset, immediate=True)
200 self.log.info("Read %d sources from %s for band %s: %s",
201 len(catalog), dataset, band, patchRef.dataId)
202 catalogDict[dataset] = catalog
203 return band, catalogDict
205 def run(self, catalogs, tract, patch):
206 """Merge multiple catalogs.
208 Parameters
209 ----------
210 catalogs : `dict`
211 Mapping from filter names to dict of catalogs.
212 tract : int
213 tractId to use for the tractId column
214 patch : str
215 patchId to use for the patchId column
217 Returns
218 -------
219 catalog : `pandas.DataFrame`
220 Merged dataframe
221 """
223 dfs = []
224 for filt, tableDict in catalogs.items():
225 for dataset, table in tableDict.items():
226 # Convert afwTable to pandas DataFrame
227 df = table.asAstropy().to_pandas().set_index('id', drop=True)
229 # Sort columns by name, to ensure matching schema among patches
230 df = df.reindex(sorted(df.columns), axis=1)
231 df['tractId'] = tract
232 df['patchId'] = patch
234 # Make columns a 3-level MultiIndex
235 df.columns = pd.MultiIndex.from_tuples([(dataset, filt, c) for c in df.columns],
236 names=('dataset', 'band', 'column'))
237 dfs.append(df)
239 catalog = functools.reduce(lambda d1, d2: d1.join(d2), dfs)
240 return catalog
242 def write(self, patchRef, catalog):
243 """Write the output.
245 Parameters
246 ----------
247 catalog : `ParquetTable`
248 Catalog to write
249 patchRef : `lsst.daf.persistence.ButlerDataRef`
250 Data reference for patch
251 """
252 patchRef.put(catalog, self.config.coaddName + "Coadd_" + self.outputDataset)
253 # since the filter isn't actually part of the data ID for the dataset we're saving,
254 # it's confusing to see it in the log message, even if the butler simply ignores it.
255 mergeDataId = patchRef.dataId.copy()
256 del mergeDataId["filter"]
257 self.log.info("Wrote merged catalog: %s", mergeDataId)
259 def writeMetadata(self, dataRefList):
260 """No metadata to write, and not sure how to write it for a list of dataRefs.
261 """
262 pass
265class WriteSourceTableConnections(pipeBase.PipelineTaskConnections,
266 defaultTemplates={"catalogType": ""},
267 dimensions=("instrument", "visit", "detector")):
269 catalog = connectionTypes.Input(
270 doc="Input full-depth catalog of sources produced by CalibrateTask",
271 name="{catalogType}src",
272 storageClass="SourceCatalog",
273 dimensions=("instrument", "visit", "detector")
274 )
275 outputCatalog = connectionTypes.Output(
276 doc="Catalog of sources, `src` in Parquet format. The 'id' column is "
277 "replaced with an index; all other columns are unchanged.",
278 name="{catalogType}source",
279 storageClass="DataFrame",
280 dimensions=("instrument", "visit", "detector")
281 )
284class WriteSourceTableConfig(pipeBase.PipelineTaskConfig,
285 pipelineConnections=WriteSourceTableConnections):
286 pass
289class WriteSourceTableTask(CmdLineTask, pipeBase.PipelineTask):
290 """Write source table to parquet
291 """
292 _DefaultName = "writeSourceTable"
293 ConfigClass = WriteSourceTableConfig
295 def runQuantum(self, butlerQC, inputRefs, outputRefs):
296 inputs = butlerQC.get(inputRefs)
297 inputs['ccdVisitId'] = butlerQC.quantum.dataId.pack("visit_detector")
298 result = self.run(**inputs).table
299 outputs = pipeBase.Struct(outputCatalog=result.toDataFrame())
300 butlerQC.put(outputs, outputRefs)
302 def run(self, catalog, ccdVisitId=None, **kwargs):
303 """Convert `src` catalog to parquet
305 Parameters
306 ----------
307 catalog: `afwTable.SourceCatalog`
308 catalog to be converted
309 ccdVisitId: `int`
310 ccdVisitId to be added as a column
312 Returns
313 -------
314 result : `lsst.pipe.base.Struct`
315 ``table``
316 `ParquetTable` version of the input catalog
317 """
318 self.log.info("Generating parquet table from src catalog ccdVisitId=%s", ccdVisitId)
319 df = catalog.asAstropy().to_pandas().set_index('id', drop=True)
320 df['ccdVisitId'] = ccdVisitId
321 return pipeBase.Struct(table=ParquetTable(dataFrame=df))
324class WriteRecalibratedSourceTableConnections(WriteSourceTableConnections,
325 defaultTemplates={"catalogType": "",
326 "skyWcsName": "jointcal",
327 "photoCalibName": "fgcm"},
328 dimensions=("instrument", "visit", "detector", "skymap")):
329 skyMap = connectionTypes.Input(
330 doc="skyMap needed to choose which tract-level calibrations to use when multiple available",
331 name=BaseSkyMap.SKYMAP_DATASET_TYPE_NAME,
332 storageClass="SkyMap",
333 dimensions=("skymap",),
334 )
335 exposure = connectionTypes.Input(
336 doc="Input exposure to perform photometry on.",
337 name="calexp",
338 storageClass="ExposureF",
339 dimensions=["instrument", "visit", "detector"],
340 )
341 externalSkyWcsTractCatalog = connectionTypes.Input(
342 doc=("Per-tract, per-visit wcs calibrations. These catalogs use the detector "
343 "id for the catalog id, sorted on id for fast lookup."),
344 name="{skyWcsName}SkyWcsCatalog",
345 storageClass="ExposureCatalog",
346 dimensions=["instrument", "visit", "tract"],
347 multiple=True
348 )
349 externalSkyWcsGlobalCatalog = connectionTypes.Input(
350 doc=("Per-visit wcs calibrations computed globally (with no tract information). "
351 "These catalogs use the detector id for the catalog id, sorted on id for "
352 "fast lookup."),
353 name="{skyWcsName}SkyWcsCatalog",
354 storageClass="ExposureCatalog",
355 dimensions=["instrument", "visit"],
356 )
357 externalPhotoCalibTractCatalog = connectionTypes.Input(
358 doc=("Per-tract, per-visit photometric calibrations. These catalogs use the "
359 "detector id for the catalog id, sorted on id for fast lookup."),
360 name="{photoCalibName}PhotoCalibCatalog",
361 storageClass="ExposureCatalog",
362 dimensions=["instrument", "visit", "tract"],
363 multiple=True
364 )
365 externalPhotoCalibGlobalCatalog = connectionTypes.Input(
366 doc=("Per-visit photometric calibrations computed globally (with no tract "
367 "information). These catalogs use the detector id for the catalog id, "
368 "sorted on id for fast lookup."),
369 name="{photoCalibName}PhotoCalibCatalog",
370 storageClass="ExposureCatalog",
371 dimensions=["instrument", "visit"],
372 )
374 def __init__(self, *, config=None):
375 super().__init__(config=config)
376 # Same connection boilerplate as all other applications of
377 # Global/Tract calibrations
378 if config.doApplyExternalSkyWcs and config.doReevaluateSkyWcs:
379 if config.useGlobalExternalSkyWcs:
380 self.inputs.remove("externalSkyWcsTractCatalog")
381 else:
382 self.inputs.remove("externalSkyWcsGlobalCatalog")
383 else:
384 self.inputs.remove("externalSkyWcsTractCatalog")
385 self.inputs.remove("externalSkyWcsGlobalCatalog")
386 if config.doApplyExternalPhotoCalib and config.doReevaluatePhotoCalib:
387 if config.useGlobalExternalPhotoCalib:
388 self.inputs.remove("externalPhotoCalibTractCatalog")
389 else:
390 self.inputs.remove("externalPhotoCalibGlobalCatalog")
391 else:
392 self.inputs.remove("externalPhotoCalibTractCatalog")
393 self.inputs.remove("externalPhotoCalibGlobalCatalog")
396class WriteRecalibratedSourceTableConfig(WriteSourceTableConfig,
397 pipelineConnections=WriteRecalibratedSourceTableConnections):
399 doReevaluatePhotoCalib = pexConfig.Field(
400 dtype=bool,
401 default=True,
402 doc=("Add or replace local photoCalib columns from either the calexp.photoCalib or jointcal/FGCM")
403 )
404 doReevaluateSkyWcs = pexConfig.Field(
405 dtype=bool,
406 default=True,
407 doc=("Add or replace local WCS columns from either the calexp.wcs or or jointcal")
408 )
409 doApplyExternalPhotoCalib = pexConfig.Field(
410 dtype=bool,
411 default=True,
412 doc=("Whether to apply external photometric calibration via an "
413 "`lsst.afw.image.PhotoCalib` object. Uses the "
414 "``externalPhotoCalibName`` field to determine which calibration "
415 "to load."),
416 )
417 doApplyExternalSkyWcs = pexConfig.Field(
418 dtype=bool,
419 default=True,
420 doc=("Whether to apply external astrometric calibration via an "
421 "`lsst.afw.geom.SkyWcs` object. Uses ``externalSkyWcsName`` "
422 "field to determine which calibration to load."),
423 )
424 useGlobalExternalPhotoCalib = pexConfig.Field(
425 dtype=bool,
426 default=True,
427 doc=("When using doApplyExternalPhotoCalib, use 'global' calibrations "
428 "that are not run per-tract. When False, use per-tract photometric "
429 "calibration files.")
430 )
431 useGlobalExternalSkyWcs = pexConfig.Field(
432 dtype=bool,
433 default=False,
434 doc=("When using doApplyExternalSkyWcs, use 'global' calibrations "
435 "that are not run per-tract. When False, use per-tract wcs "
436 "files.")
437 )
439 def validate(self):
440 super().validate()
441 if self.doApplyExternalSkyWcs and not self.doReevaluateSkyWcs:
442 log.warning("doApplyExternalSkyWcs=True but doReevaluateSkyWcs=False"
443 "External SkyWcs will not be read or evaluated.")
444 if self.doApplyExternalPhotoCalib and not self.doReevaluatePhotoCalib:
445 log.warning("doApplyExternalPhotoCalib=True but doReevaluatePhotoCalib=False."
446 "External PhotoCalib will not be read or evaluated.")
449class WriteRecalibratedSourceTableTask(WriteSourceTableTask):
450 """Write source table to parquet
451 """
452 _DefaultName = "writeRecalibratedSourceTable"
453 ConfigClass = WriteRecalibratedSourceTableConfig
455 def runQuantum(self, butlerQC, inputRefs, outputRefs):
456 inputs = butlerQC.get(inputRefs)
457 inputs['ccdVisitId'] = butlerQC.quantum.dataId.pack("visit_detector")
458 inputs['exposureIdInfo'] = ExposureIdInfo.fromDataId(butlerQC.quantum.dataId, "visit_detector")
460 if self.config.doReevaluatePhotoCalib or self.config.doReevaluateSkyWcs:
461 if self.config.doApplyExternalPhotoCalib or self.config.doApplyExternalSkyWcs:
462 inputs['exposure'] = self.attachCalibs(inputRefs, **inputs)
464 inputs['catalog'] = self.addCalibColumns(**inputs)
466 result = self.run(**inputs).table
467 outputs = pipeBase.Struct(outputCatalog=result.toDataFrame())
468 butlerQC.put(outputs, outputRefs)
470 def attachCalibs(self, inputRefs, skyMap, exposure, externalSkyWcsGlobalCatalog=None,
471 externalSkyWcsTractCatalog=None, externalPhotoCalibGlobalCatalog=None,
472 externalPhotoCalibTractCatalog=None, **kwargs):
473 """Apply external calibrations to exposure per configuration
475 When multiple tract-level calibrations overlap, select the one with the
476 center closest to detector.
478 Parameters
479 ----------
480 inputRefs : `lsst.pipe.base.InputQuantizedConnection`, for dataIds of
481 tract-level calibs.
482 skyMap : `lsst.skymap.SkyMap`
483 exposure : `lsst.afw.image.exposure.Exposure`
484 Input exposure to adjust calibrations.
485 externalSkyWcsGlobalCatalog : `lsst.afw.table.ExposureCatalog`, optional
486 Exposure catalog with external skyWcs to be applied per config
487 externalSkyWcsTractCatalog : `lsst.afw.table.ExposureCatalog`, optional
488 Exposure catalog with external skyWcs to be applied per config
489 externalPhotoCalibGlobalCatalog : `lsst.afw.table.ExposureCatalog`, optional
490 Exposure catalog with external photoCalib to be applied per config
491 externalPhotoCalibTractCatalog : `lsst.afw.table.ExposureCatalog`, optional
494 Returns
495 -------
496 exposure : `lsst.afw.image.exposure.Exposure`
497 Exposure with adjusted calibrations.
498 """
499 if not self.config.doApplyExternalSkyWcs:
500 # Do not modify the exposure's SkyWcs
501 externalSkyWcsCatalog = None
502 elif self.config.useGlobalExternalSkyWcs:
503 # Use the global external SkyWcs
504 externalSkyWcsCatalog = externalSkyWcsGlobalCatalog
505 self.log.info('Applying global SkyWcs')
506 else:
507 # use tract-level external SkyWcs from the closest overlapping tract
508 inputRef = getattr(inputRefs, 'externalSkyWcsTractCatalog')
509 tracts = [ref.dataId['tract'] for ref in inputRef]
510 if len(tracts) == 1:
511 ind = 0
512 self.log.info('Applying tract-level SkyWcs from tract %s', tracts[ind])
513 else:
514 ind = self.getClosestTract(tracts, skyMap,
515 exposure.getBBox(), exposure.getWcs())
516 self.log.info('Multiple overlapping externalSkyWcsTractCatalogs found (%s). '
517 'Applying closest to detector center: tract=%s', str(tracts), tracts[ind])
519 externalSkyWcsCatalog = externalSkyWcsTractCatalog[ind]
521 if not self.config.doApplyExternalPhotoCalib:
522 # Do not modify the exposure's PhotoCalib
523 externalPhotoCalibCatalog = None
524 elif self.config.useGlobalExternalPhotoCalib:
525 # Use the global external PhotoCalib
526 externalPhotoCalibCatalog = externalPhotoCalibGlobalCatalog
527 self.log.info('Applying global PhotoCalib')
528 else:
529 # use tract-level external PhotoCalib from the closest overlapping tract
530 inputRef = getattr(inputRefs, 'externalPhotoCalibTractCatalog')
531 tracts = [ref.dataId['tract'] for ref in inputRef]
532 if len(tracts) == 1:
533 ind = 0
534 self.log.info('Applying tract-level PhotoCalib from tract %s', tracts[ind])
535 else:
536 ind = self.getClosestTract(tracts, skyMap,
537 exposure.getBBox(), exposure.getWcs())
538 self.log.info('Multiple overlapping externalPhotoCalibTractCatalogs found (%s). '
539 'Applying closest to detector center: tract=%s', str(tracts), tracts[ind])
541 externalPhotoCalibCatalog = externalPhotoCalibTractCatalog[ind]
543 return self.prepareCalibratedExposure(exposure, externalSkyWcsCatalog, externalPhotoCalibCatalog)
545 def getClosestTract(self, tracts, skyMap, bbox, wcs):
546 """Find the index of the tract closest to detector from list of tractIds
548 Parameters
549 ----------
550 tracts: `list` [`int`]
551 Iterable of integer tractIds
552 skyMap : `lsst.skymap.SkyMap`
553 skyMap to lookup tract geometry and wcs
554 bbox : `lsst.geom.Box2I`
555 Detector bbox, center of which will compared to tract centers
556 wcs : `lsst.afw.geom.SkyWcs`
557 Detector Wcs object to map the detector center to SkyCoord
559 Returns
560 -------
561 index : `int`
562 """
563 if len(tracts) == 1:
564 return 0
566 center = wcs.pixelToSky(bbox.getCenter())
567 sep = []
568 for tractId in tracts:
569 tract = skyMap[tractId]
570 tractCenter = tract.getWcs().pixelToSky(tract.getBBox().getCenter())
571 sep.append(center.separation(tractCenter))
573 return np.argmin(sep)
575 def prepareCalibratedExposure(self, exposure, externalSkyWcsCatalog=None, externalPhotoCalibCatalog=None):
576 """Prepare a calibrated exposure and apply external calibrations
577 if so configured.
579 Parameters
580 ----------
581 exposure : `lsst.afw.image.exposure.Exposure`
582 Input exposure to adjust calibrations.
583 externalSkyWcsCatalog : `lsst.afw.table.ExposureCatalog`, optional
584 Exposure catalog with external skyWcs to be applied
585 if config.doApplyExternalSkyWcs=True. Catalog uses the detector id
586 for the catalog id, sorted on id for fast lookup.
587 externalPhotoCalibCatalog : `lsst.afw.table.ExposureCatalog`, optional
588 Exposure catalog with external photoCalib to be applied
589 if config.doApplyExternalPhotoCalib=True. Catalog uses the detector
590 id for the catalog id, sorted on id for fast lookup.
592 Returns
593 -------
594 exposure : `lsst.afw.image.exposure.Exposure`
595 Exposure with adjusted calibrations.
596 """
597 detectorId = exposure.getInfo().getDetector().getId()
599 if externalPhotoCalibCatalog is not None:
600 row = externalPhotoCalibCatalog.find(detectorId)
601 if row is None:
602 self.log.warning("Detector id %s not found in externalPhotoCalibCatalog; "
603 "Using original photoCalib.", detectorId)
604 else:
605 photoCalib = row.getPhotoCalib()
606 if photoCalib is None:
607 self.log.warning("Detector id %s has None for photoCalib in externalPhotoCalibCatalog; "
608 "Using original photoCalib.", detectorId)
609 else:
610 exposure.setPhotoCalib(photoCalib)
612 if externalSkyWcsCatalog is not None:
613 row = externalSkyWcsCatalog.find(detectorId)
614 if row is None:
615 self.log.warning("Detector id %s not found in externalSkyWcsCatalog; "
616 "Using original skyWcs.", detectorId)
617 else:
618 skyWcs = row.getWcs()
619 if skyWcs is None:
620 self.log.warning("Detector id %s has None for skyWcs in externalSkyWcsCatalog; "
621 "Using original skyWcs.", detectorId)
622 else:
623 exposure.setWcs(skyWcs)
625 return exposure
627 def addCalibColumns(self, catalog, exposure, exposureIdInfo, **kwargs):
628 """Add replace columns with calibs evaluated at each centroid
630 Add or replace 'base_LocalWcs' `base_LocalPhotoCalib' columns in a
631 a source catalog, by rerunning the plugins.
633 Parameters
634 ----------
635 catalog : `lsst.afw.table.SourceCatalog`
636 catalog to which calib columns will be added
637 exposure : `lsst.afw.image.exposure.Exposure`
638 Exposure with attached PhotoCalibs and SkyWcs attributes to be
639 reevaluated at local centroids. Pixels are not required.
640 exposureIdInfo : `lsst.obs.base.ExposureIdInfo`
642 Returns
643 -------
644 newCat: `lsst.afw.table.SourceCatalog`
645 Source Catalog with requested local calib columns
646 """
647 measureConfig = SingleFrameMeasurementTask.ConfigClass()
648 measureConfig.doReplaceWithNoise = False
650 measureConfig.plugins.names = []
651 if self.config.doReevaluateSkyWcs:
652 measureConfig.plugins.names.add('base_LocalWcs')
653 self.log.info("Re-evaluating base_LocalWcs plugin")
654 if self.config.doReevaluatePhotoCalib:
655 measureConfig.plugins.names.add('base_LocalPhotoCalib')
656 self.log.info("Re-evaluating base_LocalPhotoCalib plugin")
657 pluginsNotToCopy = tuple(measureConfig.plugins.names)
659 # Create a new schema and catalog
660 # Copy all columns from original except for the ones to reevaluate
661 aliasMap = catalog.schema.getAliasMap()
662 mapper = afwTable.SchemaMapper(catalog.schema)
663 for item in catalog.schema:
664 if not item.field.getName().startswith(pluginsNotToCopy):
665 mapper.addMapping(item.key)
667 schema = mapper.getOutputSchema()
668 measurement = SingleFrameMeasurementTask(config=measureConfig, schema=schema)
669 schema.setAliasMap(aliasMap)
670 newCat = afwTable.SourceCatalog(schema)
671 newCat.extend(catalog, mapper=mapper)
673 measurement.run(measCat=newCat, exposure=exposure, exposureId=exposureIdInfo.expId)
675 return newCat
678class PostprocessAnalysis(object):
679 """Calculate columns from ParquetTable
681 This object manages and organizes an arbitrary set of computations
682 on a catalog. The catalog is defined by a
683 `lsst.pipe.tasks.parquetTable.ParquetTable` object (or list thereof), such as a
684 `deepCoadd_obj` dataset, and the computations are defined by a collection
685 of `lsst.pipe.tasks.functor.Functor` objects (or, equivalently,
686 a `CompositeFunctor`).
688 After the object is initialized, accessing the `.df` attribute (which
689 holds the `pandas.DataFrame` containing the results of the calculations) triggers
690 computation of said dataframe.
692 One of the conveniences of using this object is the ability to define a desired common
693 filter for all functors. This enables the same functor collection to be passed to
694 several different `PostprocessAnalysis` objects without having to change the original
695 functor collection, since the `filt` keyword argument of this object triggers an
696 overwrite of the `filt` property for all functors in the collection.
698 This object also allows a list of refFlags to be passed, and defines a set of default
699 refFlags that are always included even if not requested.
701 If a list of `ParquetTable` object is passed, rather than a single one, then the
702 calculations will be mapped over all the input catalogs. In principle, it should
703 be straightforward to parallelize this activity, but initial tests have failed
704 (see TODO in code comments).
706 Parameters
707 ----------
708 parq : `lsst.pipe.tasks.ParquetTable` (or list of such)
709 Source catalog(s) for computation
711 functors : `list`, `dict`, or `lsst.pipe.tasks.functors.CompositeFunctor`
712 Computations to do (functors that act on `parq`).
713 If a dict, the output
714 DataFrame will have columns keyed accordingly.
715 If a list, the column keys will come from the
716 `.shortname` attribute of each functor.
718 filt : `str` (optional)
719 Filter in which to calculate. If provided,
720 this will overwrite any existing `.filt` attribute
721 of the provided functors.
723 flags : `list` (optional)
724 List of flags (per-band) to include in output table.
725 Taken from the `meas` dataset if applied to a multilevel Object Table.
727 refFlags : `list` (optional)
728 List of refFlags (only reference band) to include in output table.
730 forcedFlags : `list` (optional)
731 List of flags (per-band) to include in output table.
732 Taken from the ``forced_src`` dataset if applied to a
733 multilevel Object Table. Intended for flags from measurement plugins
734 only run during multi-band forced-photometry.
735 """
736 _defaultRefFlags = []
737 _defaultFuncs = ()
739 def __init__(self, parq, functors, filt=None, flags=None, refFlags=None, forcedFlags=None):
740 self.parq = parq
741 self.functors = functors
743 self.filt = filt
744 self.flags = list(flags) if flags is not None else []
745 self.forcedFlags = list(forcedFlags) if forcedFlags is not None else []
746 self.refFlags = list(self._defaultRefFlags)
747 if refFlags is not None:
748 self.refFlags += list(refFlags)
750 self._df = None
752 @property
753 def defaultFuncs(self):
754 funcs = dict(self._defaultFuncs)
755 return funcs
757 @property
758 def func(self):
759 additionalFuncs = self.defaultFuncs
760 additionalFuncs.update({flag: Column(flag, dataset='forced_src') for flag in self.forcedFlags})
761 additionalFuncs.update({flag: Column(flag, dataset='ref') for flag in self.refFlags})
762 additionalFuncs.update({flag: Column(flag, dataset='meas') for flag in self.flags})
764 if isinstance(self.functors, CompositeFunctor):
765 func = self.functors
766 else:
767 func = CompositeFunctor(self.functors)
769 func.funcDict.update(additionalFuncs)
770 func.filt = self.filt
772 return func
774 @property
775 def noDupCols(self):
776 return [name for name, func in self.func.funcDict.items() if func.noDup or func.dataset == 'ref']
778 @property
779 def df(self):
780 if self._df is None:
781 self.compute()
782 return self._df
784 def compute(self, dropna=False, pool=None):
785 # map over multiple parquet tables
786 if type(self.parq) in (list, tuple):
787 if pool is None:
788 dflist = [self.func(parq, dropna=dropna) for parq in self.parq]
789 else:
790 # TODO: Figure out why this doesn't work (pyarrow pickling issues?)
791 dflist = pool.map(functools.partial(self.func, dropna=dropna), self.parq)
792 self._df = pd.concat(dflist)
793 else:
794 self._df = self.func(self.parq, dropna=dropna)
796 return self._df
799class TransformCatalogBaseConnections(pipeBase.PipelineTaskConnections,
800 dimensions=()):
801 """Expected Connections for subclasses of TransformCatalogBaseTask.
803 Must be subclassed.
804 """
805 inputCatalog = connectionTypes.Input(
806 name="",
807 storageClass="DataFrame",
808 )
809 outputCatalog = connectionTypes.Output(
810 name="",
811 storageClass="DataFrame",
812 )
815class TransformCatalogBaseConfig(pipeBase.PipelineTaskConfig,
816 pipelineConnections=TransformCatalogBaseConnections):
817 functorFile = pexConfig.Field(
818 dtype=str,
819 doc="Path to YAML file specifying Science Data Model functors to use "
820 "when copying columns and computing calibrated values.",
821 default=None,
822 optional=True
823 )
824 primaryKey = pexConfig.Field(
825 dtype=str,
826 doc="Name of column to be set as the DataFrame index. If None, the index"
827 "will be named `id`",
828 default=None,
829 optional=True
830 )
831 columnsFromDataId = pexConfig.ListField(
832 dtype=str,
833 default=None,
834 optional=True,
835 doc="Columns to extract from the dataId",
836 )
839class TransformCatalogBaseTask(pipeBase.PipelineTask):
840 """Base class for transforming/standardizing a catalog
842 by applying functors that convert units and apply calibrations.
843 The purpose of this task is to perform a set of computations on
844 an input `ParquetTable` dataset (such as `deepCoadd_obj`) and write the
845 results to a new dataset (which needs to be declared in an `outputDataset`
846 attribute).
848 The calculations to be performed are defined in a YAML file that specifies
849 a set of functors to be computed, provided as
850 a `--functorFile` config parameter. An example of such a YAML file
851 is the following:
853 funcs:
854 psfMag:
855 functor: Mag
856 args:
857 - base_PsfFlux
858 filt: HSC-G
859 dataset: meas
860 cmodel_magDiff:
861 functor: MagDiff
862 args:
863 - modelfit_CModel
864 - base_PsfFlux
865 filt: HSC-G
866 gauss_magDiff:
867 functor: MagDiff
868 args:
869 - base_GaussianFlux
870 - base_PsfFlux
871 filt: HSC-G
872 count:
873 functor: Column
874 args:
875 - base_InputCount_value
876 filt: HSC-G
877 deconvolved_moments:
878 functor: DeconvolvedMoments
879 filt: HSC-G
880 dataset: forced_src
881 refFlags:
882 - calib_psfUsed
883 - merge_measurement_i
884 - merge_measurement_r
885 - merge_measurement_z
886 - merge_measurement_y
887 - merge_measurement_g
888 - base_PixelFlags_flag_inexact_psfCenter
889 - detect_isPrimary
891 The names for each entry under "func" will become the names of columns in the
892 output dataset. All the functors referenced are defined in `lsst.pipe.tasks.functors`.
893 Positional arguments to be passed to each functor are in the `args` list,
894 and any additional entries for each column other than "functor" or "args" (e.g., `'filt'`,
895 `'dataset'`) are treated as keyword arguments to be passed to the functor initialization.
897 The "flags" entry is the default shortcut for `Column` functors.
898 All columns listed under "flags" will be copied to the output table
899 untransformed. They can be of any datatype.
900 In the special case of transforming a multi-level oject table with
901 band and dataset indices (deepCoadd_obj), these will be taked from the
902 `meas` dataset and exploded out per band.
904 There are two special shortcuts that only apply when transforming
905 multi-level Object (deepCoadd_obj) tables:
906 - The "refFlags" entry is shortcut for `Column` functor
907 taken from the `'ref'` dataset if transforming an ObjectTable.
908 - The "forcedFlags" entry is shortcut for `Column` functors.
909 taken from the ``forced_src`` dataset if transforming an ObjectTable.
910 These are expanded out per band.
913 This task uses the `lsst.pipe.tasks.postprocess.PostprocessAnalysis` object
914 to organize and excecute the calculations.
916 """
917 @property
918 def _DefaultName(self):
919 raise NotImplementedError('Subclass must define "_DefaultName" attribute')
921 @property
922 def outputDataset(self):
923 raise NotImplementedError('Subclass must define "outputDataset" attribute')
925 @property
926 def inputDataset(self):
927 raise NotImplementedError('Subclass must define "inputDataset" attribute')
929 @property
930 def ConfigClass(self):
931 raise NotImplementedError('Subclass must define "ConfigClass" attribute')
933 def __init__(self, *args, **kwargs):
934 super().__init__(*args, **kwargs)
935 if self.config.functorFile:
936 self.log.info('Loading tranform functor definitions from %s',
937 self.config.functorFile)
938 self.funcs = CompositeFunctor.from_file(self.config.functorFile)
939 self.funcs.update(dict(PostprocessAnalysis._defaultFuncs))
940 else:
941 self.funcs = None
943 def runQuantum(self, butlerQC, inputRefs, outputRefs):
944 inputs = butlerQC.get(inputRefs)
945 if self.funcs is None:
946 raise ValueError("config.functorFile is None. "
947 "Must be a valid path to yaml in order to run Task as a PipelineTask.")
948 result = self.run(parq=inputs['inputCatalog'], funcs=self.funcs,
949 dataId=outputRefs.outputCatalog.dataId.full)
950 outputs = pipeBase.Struct(outputCatalog=result)
951 butlerQC.put(outputs, outputRefs)
953 def run(self, parq, funcs=None, dataId=None, band=None):
954 """Do postprocessing calculations
956 Takes a `ParquetTable` object and dataId,
957 returns a dataframe with results of postprocessing calculations.
959 Parameters
960 ----------
961 parq : `lsst.pipe.tasks.parquetTable.ParquetTable`
962 ParquetTable from which calculations are done.
963 funcs : `lsst.pipe.tasks.functors.Functors`
964 Functors to apply to the table's columns
965 dataId : dict, optional
966 Used to add a `patchId` column to the output dataframe.
967 band : `str`, optional
968 Filter band that is being processed.
970 Returns
971 ------
972 `pandas.DataFrame`
974 """
975 self.log.info("Transforming/standardizing the source table dataId: %s", dataId)
977 df = self.transform(band, parq, funcs, dataId).df
978 self.log.info("Made a table of %d columns and %d rows", len(df.columns), len(df))
979 return df
981 def getFunctors(self):
982 return self.funcs
984 def getAnalysis(self, parq, funcs=None, band=None):
985 if funcs is None:
986 funcs = self.funcs
987 analysis = PostprocessAnalysis(parq, funcs, filt=band)
988 return analysis
990 def transform(self, band, parq, funcs, dataId):
991 analysis = self.getAnalysis(parq, funcs=funcs, band=band)
992 df = analysis.df
993 if dataId and self.config.columnsFromDataId:
994 for key in self.config.columnsFromDataId:
995 if key in dataId:
996 df[str(key)] = dataId[key]
997 else:
998 raise ValueError(f"'{key}' in config.columnsFromDataId not found in dataId: {dataId}")
1000 if self.config.primaryKey:
1001 if df.index.name != self.config.primaryKey and self.config.primaryKey in df:
1002 df.reset_index(inplace=True, drop=True)
1003 df.set_index(self.config.primaryKey, inplace=True)
1005 return pipeBase.Struct(
1006 df=df,
1007 analysis=analysis
1008 )
1010 def write(self, df, parqRef):
1011 parqRef.put(ParquetTable(dataFrame=df), self.outputDataset)
1013 def writeMetadata(self, dataRef):
1014 """No metadata to write.
1015 """
1016 pass
1019class TransformObjectCatalogConnections(pipeBase.PipelineTaskConnections,
1020 defaultTemplates={"coaddName": "deep"},
1021 dimensions=("tract", "patch", "skymap")):
1022 inputCatalog = connectionTypes.Input(
1023 doc="The vertical concatenation of the deepCoadd_{ref|meas|forced_src} catalogs, "
1024 "stored as a DataFrame with a multi-level column index per-patch.",
1025 dimensions=("tract", "patch", "skymap"),
1026 storageClass="DataFrame",
1027 name="{coaddName}Coadd_obj",
1028 deferLoad=True,
1029 )
1030 outputCatalog = connectionTypes.Output(
1031 doc="Per-Patch Object Table of columns transformed from the deepCoadd_obj table per the standard "
1032 "data model.",
1033 dimensions=("tract", "patch", "skymap"),
1034 storageClass="DataFrame",
1035 name="objectTable"
1036 )
1039class TransformObjectCatalogConfig(TransformCatalogBaseConfig,
1040 pipelineConnections=TransformObjectCatalogConnections):
1041 coaddName = pexConfig.Field(
1042 dtype=str,
1043 default="deep",
1044 doc="Name of coadd"
1045 )
1046 # TODO: remove in DM-27177
1047 filterMap = pexConfig.DictField(
1048 keytype=str,
1049 itemtype=str,
1050 default={},
1051 doc=("Dictionary mapping full filter name to short one for column name munging."
1052 "These filters determine the output columns no matter what filters the "
1053 "input data actually contain."),
1054 deprecated=("Coadds are now identified by the band, so this transform is unused."
1055 "Will be removed after v22.")
1056 )
1057 outputBands = pexConfig.ListField(
1058 dtype=str,
1059 default=None,
1060 optional=True,
1061 doc=("These bands and only these bands will appear in the output,"
1062 " NaN-filled if the input does not include them."
1063 " If None, then use all bands found in the input.")
1064 )
1065 camelCase = pexConfig.Field(
1066 dtype=bool,
1067 default=False,
1068 doc=("Write per-band columns names with camelCase, else underscore "
1069 "For example: gPsFlux instead of g_PsFlux.")
1070 )
1071 multilevelOutput = pexConfig.Field(
1072 dtype=bool,
1073 default=False,
1074 doc=("Whether results dataframe should have a multilevel column index (True) or be flat "
1075 "and name-munged (False).")
1076 )
1077 goodFlags = pexConfig.ListField(
1078 dtype=str,
1079 default=[],
1080 doc=("List of 'good' flags that should be set False when populating empty tables. "
1081 "All other flags are considered to be 'bad' flags and will be set to True.")
1082 )
1083 floatFillValue = pexConfig.Field(
1084 dtype=float,
1085 default=np.nan,
1086 doc="Fill value for float fields when populating empty tables."
1087 )
1088 integerFillValue = pexConfig.Field(
1089 dtype=int,
1090 default=-1,
1091 doc="Fill value for integer fields when populating empty tables."
1092 )
1094 def setDefaults(self):
1095 super().setDefaults()
1096 self.functorFile = os.path.join('$PIPE_TASKS_DIR', 'schemas', 'Object.yaml')
1097 self.primaryKey = 'objectId'
1098 self.columnsFromDataId = ['tract', 'patch']
1099 self.goodFlags = ['calib_astrometry_used',
1100 'calib_photometry_reserved',
1101 'calib_photometry_used',
1102 'calib_psf_candidate',
1103 'calib_psf_reserved',
1104 'calib_psf_used']
1107class TransformObjectCatalogTask(TransformCatalogBaseTask):
1108 """Produce a flattened Object Table to match the format specified in
1109 sdm_schemas.
1111 Do the same set of postprocessing calculations on all bands
1113 This is identical to `TransformCatalogBaseTask`, except for that it does the
1114 specified functor calculations for all filters present in the
1115 input `deepCoadd_obj` table. Any specific `"filt"` keywords specified
1116 by the YAML file will be superceded.
1117 """
1118 _DefaultName = "transformObjectCatalog"
1119 ConfigClass = TransformObjectCatalogConfig
1121 def run(self, parq, funcs=None, dataId=None, band=None):
1122 # NOTE: band kwarg is ignored here.
1123 dfDict = {}
1124 analysisDict = {}
1125 templateDf = pd.DataFrame()
1127 if isinstance(parq, DeferredDatasetHandle):
1128 columns = parq.get(component='columns')
1129 inputBands = columns.unique(level=1).values
1130 else:
1131 inputBands = parq.columnLevelNames['band']
1133 outputBands = self.config.outputBands if self.config.outputBands else inputBands
1135 # Perform transform for data of filters that exist in parq.
1136 for inputBand in inputBands:
1137 if inputBand not in outputBands:
1138 self.log.info("Ignoring %s band data in the input", inputBand)
1139 continue
1140 self.log.info("Transforming the catalog of band %s", inputBand)
1141 result = self.transform(inputBand, parq, funcs, dataId)
1142 dfDict[inputBand] = result.df
1143 analysisDict[inputBand] = result.analysis
1144 if templateDf.empty:
1145 templateDf = result.df
1147 # Put filler values in columns of other wanted bands
1148 for filt in outputBands:
1149 if filt not in dfDict:
1150 self.log.info("Adding empty columns for band %s", filt)
1151 dfTemp = templateDf.copy()
1152 for col in dfTemp.columns:
1153 testValue = dfTemp[col].values[0]
1154 if isinstance(testValue, (np.bool_, pd.BooleanDtype)):
1155 # Boolean flag type, check if it is a "good" flag
1156 if col in self.config.goodFlags:
1157 fillValue = False
1158 else:
1159 fillValue = True
1160 elif isinstance(testValue, numbers.Integral):
1161 # Checking numbers.Integral catches all flavors
1162 # of python, numpy, pandas, etc. integers.
1163 # We must ensure this is not an unsigned integer.
1164 if isinstance(testValue, np.unsignedinteger):
1165 raise ValueError("Parquet tables may not have unsigned integer columns.")
1166 else:
1167 fillValue = self.config.integerFillValue
1168 else:
1169 fillValue = self.config.floatFillValue
1170 dfTemp[col].values[:] = fillValue
1171 dfDict[filt] = dfTemp
1173 # This makes a multilevel column index, with band as first level
1174 df = pd.concat(dfDict, axis=1, names=['band', 'column'])
1176 if not self.config.multilevelOutput:
1177 noDupCols = list(set.union(*[set(v.noDupCols) for v in analysisDict.values()]))
1178 if self.config.primaryKey in noDupCols:
1179 noDupCols.remove(self.config.primaryKey)
1180 if dataId and self.config.columnsFromDataId:
1181 noDupCols += self.config.columnsFromDataId
1182 df = flattenFilters(df, noDupCols=noDupCols, camelCase=self.config.camelCase,
1183 inputBands=inputBands)
1185 self.log.info("Made a table of %d columns and %d rows", len(df.columns), len(df))
1187 return df
1190class TractObjectDataIdContainer(CoaddDataIdContainer):
1192 def makeDataRefList(self, namespace):
1193 """Make self.refList from self.idList
1195 Generate a list of data references given tract and/or patch.
1196 This was adapted from `TractQADataIdContainer`, which was
1197 `TractDataIdContainer` modifie to not require "filter".
1198 Only existing dataRefs are returned.
1199 """
1200 def getPatchRefList(tract):
1201 return [namespace.butler.dataRef(datasetType=self.datasetType,
1202 tract=tract.getId(),
1203 patch="%d,%d" % patch.getIndex()) for patch in tract]
1205 tractRefs = defaultdict(list) # Data references for each tract
1206 for dataId in self.idList:
1207 skymap = self.getSkymap(namespace)
1209 if "tract" in dataId:
1210 tractId = dataId["tract"]
1211 if "patch" in dataId:
1212 tractRefs[tractId].append(namespace.butler.dataRef(datasetType=self.datasetType,
1213 tract=tractId,
1214 patch=dataId['patch']))
1215 else:
1216 tractRefs[tractId] += getPatchRefList(skymap[tractId])
1217 else:
1218 tractRefs = dict((tract.getId(), tractRefs.get(tract.getId(), []) + getPatchRefList(tract))
1219 for tract in skymap)
1220 outputRefList = []
1221 for tractRefList in tractRefs.values():
1222 existingRefs = [ref for ref in tractRefList if ref.datasetExists()]
1223 outputRefList.append(existingRefs)
1225 self.refList = outputRefList
1228class ConsolidateObjectTableConnections(pipeBase.PipelineTaskConnections,
1229 dimensions=("tract", "skymap")):
1230 inputCatalogs = connectionTypes.Input(
1231 doc="Per-Patch objectTables conforming to the standard data model.",
1232 name="objectTable",
1233 storageClass="DataFrame",
1234 dimensions=("tract", "patch", "skymap"),
1235 multiple=True,
1236 )
1237 outputCatalog = connectionTypes.Output(
1238 doc="Pre-tract horizontal concatenation of the input objectTables",
1239 name="objectTable_tract",
1240 storageClass="DataFrame",
1241 dimensions=("tract", "skymap"),
1242 )
1245class ConsolidateObjectTableConfig(pipeBase.PipelineTaskConfig,
1246 pipelineConnections=ConsolidateObjectTableConnections):
1247 coaddName = pexConfig.Field(
1248 dtype=str,
1249 default="deep",
1250 doc="Name of coadd"
1251 )
1254class ConsolidateObjectTableTask(CmdLineTask, pipeBase.PipelineTask):
1255 """Write patch-merged source tables to a tract-level parquet file
1257 Concatenates `objectTable` list into a per-visit `objectTable_tract`
1258 """
1259 _DefaultName = "consolidateObjectTable"
1260 ConfigClass = ConsolidateObjectTableConfig
1262 inputDataset = 'objectTable'
1263 outputDataset = 'objectTable_tract'
1265 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1266 inputs = butlerQC.get(inputRefs)
1267 self.log.info("Concatenating %s per-patch Object Tables",
1268 len(inputs['inputCatalogs']))
1269 df = pd.concat(inputs['inputCatalogs'])
1270 butlerQC.put(pipeBase.Struct(outputCatalog=df), outputRefs)
1272 @classmethod
1273 def _makeArgumentParser(cls):
1274 parser = ArgumentParser(name=cls._DefaultName)
1276 parser.add_id_argument("--id", cls.inputDataset,
1277 help="data ID, e.g. --id tract=12345",
1278 ContainerClass=TractObjectDataIdContainer)
1279 return parser
1281 def runDataRef(self, patchRefList):
1282 df = pd.concat([patchRef.get().toDataFrame() for patchRef in patchRefList])
1283 patchRefList[0].put(ParquetTable(dataFrame=df), self.outputDataset)
1285 def writeMetadata(self, dataRef):
1286 """No metadata to write.
1287 """
1288 pass
1291class TransformSourceTableConnections(pipeBase.PipelineTaskConnections,
1292 defaultTemplates={"catalogType": ""},
1293 dimensions=("instrument", "visit", "detector")):
1295 inputCatalog = connectionTypes.Input(
1296 doc="Wide input catalog of sources produced by WriteSourceTableTask",
1297 name="{catalogType}source",
1298 storageClass="DataFrame",
1299 dimensions=("instrument", "visit", "detector"),
1300 deferLoad=True
1301 )
1302 outputCatalog = connectionTypes.Output(
1303 doc="Narrower, per-detector Source Table transformed and converted per a "
1304 "specified set of functors",
1305 name="{catalogType}sourceTable",
1306 storageClass="DataFrame",
1307 dimensions=("instrument", "visit", "detector")
1308 )
1311class TransformSourceTableConfig(TransformCatalogBaseConfig,
1312 pipelineConnections=TransformSourceTableConnections):
1314 def setDefaults(self):
1315 super().setDefaults()
1316 self.functorFile = os.path.join('$PIPE_TASKS_DIR', 'schemas', 'Source.yaml')
1317 self.primaryKey = 'sourceId'
1318 self.columnsFromDataId = ['visit', 'detector', 'band', 'physical_filter']
1321class TransformSourceTableTask(TransformCatalogBaseTask):
1322 """Transform/standardize a source catalog
1323 """
1324 _DefaultName = "transformSourceTable"
1325 ConfigClass = TransformSourceTableConfig
1328class ConsolidateVisitSummaryConnections(pipeBase.PipelineTaskConnections,
1329 dimensions=("instrument", "visit",),
1330 defaultTemplates={"calexpType": ""}):
1331 calexp = connectionTypes.Input(
1332 doc="Processed exposures used for metadata",
1333 name="{calexpType}calexp",
1334 storageClass="ExposureF",
1335 dimensions=("instrument", "visit", "detector"),
1336 deferLoad=True,
1337 multiple=True,
1338 )
1339 visitSummary = connectionTypes.Output(
1340 doc=("Per-visit consolidated exposure metadata. These catalogs use "
1341 "detector id for the id and are sorted for fast lookups of a "
1342 "detector."),
1343 name="{calexpType}visitSummary",
1344 storageClass="ExposureCatalog",
1345 dimensions=("instrument", "visit"),
1346 )
1349class ConsolidateVisitSummaryConfig(pipeBase.PipelineTaskConfig,
1350 pipelineConnections=ConsolidateVisitSummaryConnections):
1351 """Config for ConsolidateVisitSummaryTask"""
1352 pass
1355class ConsolidateVisitSummaryTask(pipeBase.PipelineTask, pipeBase.CmdLineTask):
1356 """Task to consolidate per-detector visit metadata.
1358 This task aggregates the following metadata from all the detectors in a
1359 single visit into an exposure catalog:
1360 - The visitInfo.
1361 - The wcs.
1362 - The photoCalib.
1363 - The physical_filter and band (if available).
1364 - The psf size, shape, and effective area at the center of the detector.
1365 - The corners of the bounding box in right ascension/declination.
1367 Other quantities such as Detector, Psf, ApCorrMap, and TransmissionCurve
1368 are not persisted here because of storage concerns, and because of their
1369 limited utility as summary statistics.
1371 Tests for this task are performed in ci_hsc_gen3.
1372 """
1373 _DefaultName = "consolidateVisitSummary"
1374 ConfigClass = ConsolidateVisitSummaryConfig
1376 @classmethod
1377 def _makeArgumentParser(cls):
1378 parser = ArgumentParser(name=cls._DefaultName)
1380 parser.add_id_argument("--id", "calexp",
1381 help="data ID, e.g. --id visit=12345",
1382 ContainerClass=VisitDataIdContainer)
1383 return parser
1385 def writeMetadata(self, dataRef):
1386 """No metadata to persist, so override to remove metadata persistance.
1387 """
1388 pass
1390 def writeConfig(self, butler, clobber=False, doBackup=True):
1391 """No config to persist, so override to remove config persistance.
1392 """
1393 pass
1395 def runDataRef(self, dataRefList):
1396 visit = dataRefList[0].dataId['visit']
1398 self.log.debug("Concatenating metadata from %d per-detector calexps (visit %d)",
1399 len(dataRefList), visit)
1401 expCatalog = self._combineExposureMetadata(visit, dataRefList, isGen3=False)
1403 dataRefList[0].put(expCatalog, 'visitSummary', visit=visit)
1405 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1406 dataRefs = butlerQC.get(inputRefs.calexp)
1407 visit = dataRefs[0].dataId.byName()['visit']
1409 self.log.debug("Concatenating metadata from %d per-detector calexps (visit %d)",
1410 len(dataRefs), visit)
1412 expCatalog = self._combineExposureMetadata(visit, dataRefs)
1414 butlerQC.put(expCatalog, outputRefs.visitSummary)
1416 def _combineExposureMetadata(self, visit, dataRefs, isGen3=True):
1417 """Make a combined exposure catalog from a list of dataRefs.
1418 These dataRefs must point to exposures with wcs, summaryStats,
1419 and other visit metadata.
1421 Parameters
1422 ----------
1423 visit : `int`
1424 Visit identification number.
1425 dataRefs : `list`
1426 List of dataRefs in visit. May be list of
1427 `lsst.daf.persistence.ButlerDataRef` (Gen2) or
1428 `lsst.daf.butler.DeferredDatasetHandle` (Gen3).
1429 isGen3 : `bool`, optional
1430 Specifies if this is a Gen3 list of datarefs.
1432 Returns
1433 -------
1434 visitSummary : `lsst.afw.table.ExposureCatalog`
1435 Exposure catalog with per-detector summary information.
1436 """
1437 schema = self._makeVisitSummarySchema()
1438 cat = afwTable.ExposureCatalog(schema)
1439 cat.resize(len(dataRefs))
1441 cat['visit'] = visit
1443 for i, dataRef in enumerate(dataRefs):
1444 if isGen3:
1445 visitInfo = dataRef.get(component='visitInfo')
1446 filterLabel = dataRef.get(component='filter')
1447 summaryStats = dataRef.get(component='summaryStats')
1448 detector = dataRef.get(component='detector')
1449 wcs = dataRef.get(component='wcs')
1450 photoCalib = dataRef.get(component='photoCalib')
1451 detector = dataRef.get(component='detector')
1452 bbox = dataRef.get(component='bbox')
1453 validPolygon = dataRef.get(component='validPolygon')
1454 else:
1455 # Note that we need to read the calexp because there is
1456 # no magic access to the psf except through the exposure.
1457 gen2_read_bbox = lsst.geom.BoxI(lsst.geom.PointI(0, 0), lsst.geom.PointI(1, 1))
1458 exp = dataRef.get(datasetType='calexp_sub', bbox=gen2_read_bbox)
1459 visitInfo = exp.getInfo().getVisitInfo()
1460 filterLabel = dataRef.get("calexp_filter")
1461 summaryStats = exp.getInfo().getSummaryStats()
1462 wcs = exp.getWcs()
1463 photoCalib = exp.getPhotoCalib()
1464 detector = exp.getDetector()
1465 bbox = dataRef.get(datasetType='calexp_bbox')
1466 validPolygon = exp.getInfo().getValidPolygon()
1468 rec = cat[i]
1469 rec.setBBox(bbox)
1470 rec.setVisitInfo(visitInfo)
1471 rec.setWcs(wcs)
1472 rec.setPhotoCalib(photoCalib)
1473 rec.setValidPolygon(validPolygon)
1475 rec['physical_filter'] = filterLabel.physicalLabel if filterLabel.hasPhysicalLabel() else ""
1476 rec['band'] = filterLabel.bandLabel if filterLabel.hasBandLabel() else ""
1477 rec.setId(detector.getId())
1478 rec['psfSigma'] = summaryStats.psfSigma
1479 rec['psfIxx'] = summaryStats.psfIxx
1480 rec['psfIyy'] = summaryStats.psfIyy
1481 rec['psfIxy'] = summaryStats.psfIxy
1482 rec['psfArea'] = summaryStats.psfArea
1483 rec['raCorners'][:] = summaryStats.raCorners
1484 rec['decCorners'][:] = summaryStats.decCorners
1485 rec['ra'] = summaryStats.ra
1486 rec['decl'] = summaryStats.decl
1487 rec['zenithDistance'] = summaryStats.zenithDistance
1488 rec['zeroPoint'] = summaryStats.zeroPoint
1489 rec['skyBg'] = summaryStats.skyBg
1490 rec['skyNoise'] = summaryStats.skyNoise
1491 rec['meanVar'] = summaryStats.meanVar
1492 rec['astromOffsetMean'] = summaryStats.astromOffsetMean
1493 rec['astromOffsetStd'] = summaryStats.astromOffsetStd
1494 rec['nPsfStar'] = summaryStats.nPsfStar
1495 rec['psfStarDeltaE1Median'] = summaryStats.psfStarDeltaE1Median
1496 rec['psfStarDeltaE2Median'] = summaryStats.psfStarDeltaE2Median
1497 rec['psfStarDeltaE1Scatter'] = summaryStats.psfStarDeltaE1Scatter
1498 rec['psfStarDeltaE2Scatter'] = summaryStats.psfStarDeltaE2Scatter
1499 rec['psfStarDeltaSizeMedian'] = summaryStats.psfStarDeltaSizeMedian
1500 rec['psfStarDeltaSizeScatter'] = summaryStats.psfStarDeltaSizeScatter
1501 rec['psfStarScaledDeltaSizeScatter'] = summaryStats.psfStarScaledDeltaSizeScatter
1503 metadata = dafBase.PropertyList()
1504 metadata.add("COMMENT", "Catalog id is detector id, sorted.")
1505 # We are looping over existing datarefs, so the following is true
1506 metadata.add("COMMENT", "Only detectors with data have entries.")
1507 cat.setMetadata(metadata)
1509 cat.sort()
1510 return cat
1512 def _makeVisitSummarySchema(self):
1513 """Make the schema for the visitSummary catalog."""
1514 schema = afwTable.ExposureTable.makeMinimalSchema()
1515 schema.addField('visit', type='L', doc='Visit number')
1516 schema.addField('physical_filter', type='String', size=32, doc='Physical filter')
1517 schema.addField('band', type='String', size=32, doc='Name of band')
1518 schema.addField('psfSigma', type='F',
1519 doc='PSF model second-moments determinant radius (center of chip) (pixel)')
1520 schema.addField('psfArea', type='F',
1521 doc='PSF model effective area (center of chip) (pixel**2)')
1522 schema.addField('psfIxx', type='F',
1523 doc='PSF model Ixx (center of chip) (pixel**2)')
1524 schema.addField('psfIyy', type='F',
1525 doc='PSF model Iyy (center of chip) (pixel**2)')
1526 schema.addField('psfIxy', type='F',
1527 doc='PSF model Ixy (center of chip) (pixel**2)')
1528 schema.addField('raCorners', type='ArrayD', size=4,
1529 doc='Right Ascension of bounding box corners (degrees)')
1530 schema.addField('decCorners', type='ArrayD', size=4,
1531 doc='Declination of bounding box corners (degrees)')
1532 schema.addField('ra', type='D',
1533 doc='Right Ascension of bounding box center (degrees)')
1534 schema.addField('decl', type='D',
1535 doc='Declination of bounding box center (degrees)')
1536 schema.addField('zenithDistance', type='F',
1537 doc='Zenith distance of bounding box center (degrees)')
1538 schema.addField('zeroPoint', type='F',
1539 doc='Mean zeropoint in detector (mag)')
1540 schema.addField('skyBg', type='F',
1541 doc='Average sky background (ADU)')
1542 schema.addField('skyNoise', type='F',
1543 doc='Average sky noise (ADU)')
1544 schema.addField('meanVar', type='F',
1545 doc='Mean variance of the weight plane (ADU**2)')
1546 schema.addField('astromOffsetMean', type='F',
1547 doc='Mean offset of astrometric calibration matches (arcsec)')
1548 schema.addField('astromOffsetStd', type='F',
1549 doc='Standard deviation of offsets of astrometric calibration matches (arcsec)')
1550 schema.addField('nPsfStar', type='I', doc='Number of stars used for PSF model')
1551 schema.addField('psfStarDeltaE1Median', type='F',
1552 doc='Median E1 residual (starE1 - psfE1) for psf stars')
1553 schema.addField('psfStarDeltaE2Median', type='F',
1554 doc='Median E2 residual (starE2 - psfE2) for psf stars')
1555 schema.addField('psfStarDeltaE1Scatter', type='F',
1556 doc='Scatter (via MAD) of E1 residual (starE1 - psfE1) for psf stars')
1557 schema.addField('psfStarDeltaE2Scatter', type='F',
1558 doc='Scatter (via MAD) of E2 residual (starE2 - psfE2) for psf stars')
1559 schema.addField('psfStarDeltaSizeMedian', type='F',
1560 doc='Median size residual (starSize - psfSize) for psf stars (pixel)')
1561 schema.addField('psfStarDeltaSizeScatter', type='F',
1562 doc='Scatter (via MAD) of size residual (starSize - psfSize) for psf stars (pixel)')
1563 schema.addField('psfStarScaledDeltaSizeScatter', type='F',
1564 doc='Scatter (via MAD) of size residual scaled by median size squared')
1566 return schema
1569class VisitDataIdContainer(DataIdContainer):
1570 """DataIdContainer that groups sensor-level id's by visit
1571 """
1573 def makeDataRefList(self, namespace):
1574 """Make self.refList from self.idList
1576 Generate a list of data references grouped by visit.
1578 Parameters
1579 ----------
1580 namespace : `argparse.Namespace`
1581 Namespace used by `lsst.pipe.base.CmdLineTask` to parse command line arguments
1582 """
1583 # Group by visits
1584 visitRefs = defaultdict(list)
1585 for dataId in self.idList:
1586 if "visit" in dataId:
1587 visitId = dataId["visit"]
1588 # append all subsets to
1589 subset = namespace.butler.subset(self.datasetType, dataId=dataId)
1590 visitRefs[visitId].extend([dataRef for dataRef in subset])
1592 outputRefList = []
1593 for refList in visitRefs.values():
1594 existingRefs = [ref for ref in refList if ref.datasetExists()]
1595 if existingRefs:
1596 outputRefList.append(existingRefs)
1598 self.refList = outputRefList
1601class ConsolidateSourceTableConnections(pipeBase.PipelineTaskConnections,
1602 defaultTemplates={"catalogType": ""},
1603 dimensions=("instrument", "visit")):
1604 inputCatalogs = connectionTypes.Input(
1605 doc="Input per-detector Source Tables",
1606 name="{catalogType}sourceTable",
1607 storageClass="DataFrame",
1608 dimensions=("instrument", "visit", "detector"),
1609 multiple=True
1610 )
1611 outputCatalog = connectionTypes.Output(
1612 doc="Per-visit concatenation of Source Table",
1613 name="{catalogType}sourceTable_visit",
1614 storageClass="DataFrame",
1615 dimensions=("instrument", "visit")
1616 )
1619class ConsolidateSourceTableConfig(pipeBase.PipelineTaskConfig,
1620 pipelineConnections=ConsolidateSourceTableConnections):
1621 pass
1624class ConsolidateSourceTableTask(CmdLineTask, pipeBase.PipelineTask):
1625 """Concatenate `sourceTable` list into a per-visit `sourceTable_visit`
1626 """
1627 _DefaultName = 'consolidateSourceTable'
1628 ConfigClass = ConsolidateSourceTableConfig
1630 inputDataset = 'sourceTable'
1631 outputDataset = 'sourceTable_visit'
1633 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1634 from .makeCoaddTempExp import reorderRefs
1636 detectorOrder = [ref.dataId['detector'] for ref in inputRefs.inputCatalogs]
1637 detectorOrder.sort()
1638 inputRefs = reorderRefs(inputRefs, detectorOrder, dataIdKey='detector')
1639 inputs = butlerQC.get(inputRefs)
1640 self.log.info("Concatenating %s per-detector Source Tables",
1641 len(inputs['inputCatalogs']))
1642 df = pd.concat(inputs['inputCatalogs'])
1643 butlerQC.put(pipeBase.Struct(outputCatalog=df), outputRefs)
1645 def runDataRef(self, dataRefList):
1646 self.log.info("Concatenating %s per-detector Source Tables", len(dataRefList))
1647 df = pd.concat([dataRef.get().toDataFrame() for dataRef in dataRefList])
1648 dataRefList[0].put(ParquetTable(dataFrame=df), self.outputDataset)
1650 @classmethod
1651 def _makeArgumentParser(cls):
1652 parser = ArgumentParser(name=cls._DefaultName)
1654 parser.add_id_argument("--id", cls.inputDataset,
1655 help="data ID, e.g. --id visit=12345",
1656 ContainerClass=VisitDataIdContainer)
1657 return parser
1659 def writeMetadata(self, dataRef):
1660 """No metadata to write.
1661 """
1662 pass
1664 def writeConfig(self, butler, clobber=False, doBackup=True):
1665 """No config to write.
1666 """
1667 pass
1670class MakeCcdVisitTableConnections(pipeBase.PipelineTaskConnections,
1671 dimensions=("instrument",),
1672 defaultTemplates={"calexpType": ""}):
1673 visitSummaryRefs = connectionTypes.Input(
1674 doc="Data references for per-visit consolidated exposure metadata from ConsolidateVisitSummaryTask",
1675 name="{calexpType}visitSummary",
1676 storageClass="ExposureCatalog",
1677 dimensions=("instrument", "visit"),
1678 multiple=True,
1679 deferLoad=True,
1680 )
1681 outputCatalog = connectionTypes.Output(
1682 doc="CCD and Visit metadata table",
1683 name="ccdVisitTable",
1684 storageClass="DataFrame",
1685 dimensions=("instrument",)
1686 )
1689class MakeCcdVisitTableConfig(pipeBase.PipelineTaskConfig,
1690 pipelineConnections=MakeCcdVisitTableConnections):
1691 pass
1694class MakeCcdVisitTableTask(CmdLineTask, pipeBase.PipelineTask):
1695 """Produce a `ccdVisitTable` from the `visitSummary` exposure catalogs.
1696 """
1697 _DefaultName = 'makeCcdVisitTable'
1698 ConfigClass = MakeCcdVisitTableConfig
1700 def run(self, visitSummaryRefs):
1701 """ Make a table of ccd information from the `visitSummary` catalogs.
1702 Parameters
1703 ----------
1704 visitSummaryRefs : `list` of `lsst.daf.butler.DeferredDatasetHandle`
1705 List of DeferredDatasetHandles pointing to exposure catalogs with
1706 per-detector summary information.
1707 Returns
1708 -------
1709 result : `lsst.pipe.Base.Struct`
1710 Results struct with attribute:
1711 - `outputCatalog`
1712 Catalog of ccd and visit information.
1713 """
1714 ccdEntries = []
1715 for visitSummaryRef in visitSummaryRefs:
1716 visitSummary = visitSummaryRef.get()
1717 visitInfo = visitSummary[0].getVisitInfo()
1719 ccdEntry = {}
1720 summaryTable = visitSummary.asAstropy()
1721 selectColumns = ['id', 'visit', 'physical_filter', 'band', 'ra', 'decl', 'zenithDistance',
1722 'zeroPoint', 'psfSigma', 'skyBg', 'skyNoise']
1723 ccdEntry = summaryTable[selectColumns].to_pandas().set_index('id')
1724 # 'visit' is the human readible visit number
1725 # 'visitId' is the key to the visitId table. They are the same
1726 # Technically you should join to get the visit from the visit table
1727 ccdEntry = ccdEntry.rename(columns={"visit": "visitId"})
1728 dataIds = [DataCoordinate.standardize(visitSummaryRef.dataId, detector=id) for id in
1729 summaryTable['id']]
1730 packer = visitSummaryRef.dataId.universe.makePacker('visit_detector', visitSummaryRef.dataId)
1731 ccdVisitIds = [packer.pack(dataId) for dataId in dataIds]
1732 ccdEntry['ccdVisitId'] = ccdVisitIds
1733 ccdEntry['detector'] = summaryTable['id']
1734 pixToArcseconds = np.array([vR.getWcs().getPixelScale().asArcseconds() for vR in visitSummary])
1735 ccdEntry["seeing"] = visitSummary['psfSigma'] * np.sqrt(8 * np.log(2)) * pixToArcseconds
1737 ccdEntry["skyRotation"] = visitInfo.getBoresightRotAngle().asDegrees()
1738 ccdEntry["expMidpt"] = visitInfo.getDate().toPython()
1739 ccdEntry["expMidptMJD"] = visitInfo.getDate().get(dafBase.DateTime.MJD)
1740 expTime = visitInfo.getExposureTime()
1741 ccdEntry['expTime'] = expTime
1742 ccdEntry["obsStart"] = ccdEntry["expMidpt"] - 0.5 * pd.Timedelta(seconds=expTime)
1743 expTime_days = expTime / (60*60*24)
1744 ccdEntry["obsStartMJD"] = ccdEntry["expMidptMJD"] - 0.5 * expTime_days
1745 ccdEntry['darkTime'] = visitInfo.getDarkTime()
1746 ccdEntry['xSize'] = summaryTable['bbox_max_x'] - summaryTable['bbox_min_x']
1747 ccdEntry['ySize'] = summaryTable['bbox_max_y'] - summaryTable['bbox_min_y']
1748 ccdEntry['llcra'] = summaryTable['raCorners'][:, 0]
1749 ccdEntry['llcdec'] = summaryTable['decCorners'][:, 0]
1750 ccdEntry['ulcra'] = summaryTable['raCorners'][:, 1]
1751 ccdEntry['ulcdec'] = summaryTable['decCorners'][:, 1]
1752 ccdEntry['urcra'] = summaryTable['raCorners'][:, 2]
1753 ccdEntry['urcdec'] = summaryTable['decCorners'][:, 2]
1754 ccdEntry['lrcra'] = summaryTable['raCorners'][:, 3]
1755 ccdEntry['lrcdec'] = summaryTable['decCorners'][:, 3]
1756 # TODO: DM-30618, Add raftName, nExposures, ccdTemp, binX, binY, and flags,
1757 # and decide if WCS, and llcx, llcy, ulcx, ulcy, etc. values are actually wanted.
1758 ccdEntries.append(ccdEntry)
1760 outputCatalog = pd.concat(ccdEntries)
1761 outputCatalog.set_index('ccdVisitId', inplace=True, verify_integrity=True)
1762 return pipeBase.Struct(outputCatalog=outputCatalog)
1765class MakeVisitTableConnections(pipeBase.PipelineTaskConnections,
1766 dimensions=("instrument",),
1767 defaultTemplates={"calexpType": ""}):
1768 visitSummaries = connectionTypes.Input(
1769 doc="Per-visit consolidated exposure metadata from ConsolidateVisitSummaryTask",
1770 name="{calexpType}visitSummary",
1771 storageClass="ExposureCatalog",
1772 dimensions=("instrument", "visit",),
1773 multiple=True,
1774 deferLoad=True,
1775 )
1776 outputCatalog = connectionTypes.Output(
1777 doc="Visit metadata table",
1778 name="visitTable",
1779 storageClass="DataFrame",
1780 dimensions=("instrument",)
1781 )
1784class MakeVisitTableConfig(pipeBase.PipelineTaskConfig,
1785 pipelineConnections=MakeVisitTableConnections):
1786 pass
1789class MakeVisitTableTask(CmdLineTask, pipeBase.PipelineTask):
1790 """Produce a `visitTable` from the `visitSummary` exposure catalogs.
1791 """
1792 _DefaultName = 'makeVisitTable'
1793 ConfigClass = MakeVisitTableConfig
1795 def run(self, visitSummaries):
1796 """ Make a table of visit information from the `visitSummary` catalogs
1798 Parameters
1799 ----------
1800 visitSummaries : list of `lsst.afw.table.ExposureCatalog`
1801 List of exposure catalogs with per-detector summary information.
1802 Returns
1803 -------
1804 result : `lsst.pipe.Base.Struct`
1805 Results struct with attribute:
1806 ``outputCatalog``
1807 Catalog of visit information.
1808 """
1809 visitEntries = []
1810 for visitSummary in visitSummaries:
1811 visitSummary = visitSummary.get()
1812 visitRow = visitSummary[0]
1813 visitInfo = visitRow.getVisitInfo()
1815 visitEntry = {}
1816 visitEntry["visitId"] = visitRow['visit']
1817 visitEntry["visit"] = visitRow['visit']
1818 visitEntry["physical_filter"] = visitRow['physical_filter']
1819 visitEntry["band"] = visitRow['band']
1820 raDec = visitInfo.getBoresightRaDec()
1821 visitEntry["ra"] = raDec.getRa().asDegrees()
1822 visitEntry["decl"] = raDec.getDec().asDegrees()
1823 visitEntry["skyRotation"] = visitInfo.getBoresightRotAngle().asDegrees()
1824 azAlt = visitInfo.getBoresightAzAlt()
1825 visitEntry["azimuth"] = azAlt.getLongitude().asDegrees()
1826 visitEntry["altitude"] = azAlt.getLatitude().asDegrees()
1827 visitEntry["zenithDistance"] = 90 - azAlt.getLatitude().asDegrees()
1828 visitEntry["airmass"] = visitInfo.getBoresightAirmass()
1829 expTime = visitInfo.getExposureTime()
1830 visitEntry["expTime"] = expTime
1831 visitEntry["expMidpt"] = visitInfo.getDate().toPython()
1832 visitEntry["expMidptMJD"] = visitInfo.getDate().get(dafBase.DateTime.MJD)
1833 visitEntry["obsStart"] = visitEntry["expMidpt"] - 0.5 * pd.Timedelta(seconds=expTime)
1834 expTime_days = expTime / (60*60*24)
1835 visitEntry["obsStartMJD"] = visitEntry["expMidptMJD"] - 0.5 * expTime_days
1836 visitEntries.append(visitEntry)
1838 # TODO: DM-30623, Add programId, exposureType, cameraTemp, mirror1Temp, mirror2Temp,
1839 # mirror3Temp, domeTemp, externalTemp, dimmSeeing, pwvGPS, pwvMW, flags, nExposures
1841 outputCatalog = pd.DataFrame(data=visitEntries)
1842 outputCatalog.set_index('visitId', inplace=True, verify_integrity=True)
1843 return pipeBase.Struct(outputCatalog=outputCatalog)
1846class WriteForcedSourceTableConnections(pipeBase.PipelineTaskConnections,
1847 dimensions=("instrument", "visit", "detector", "skymap", "tract")):
1849 inputCatalog = connectionTypes.Input(
1850 doc="Primary per-detector, single-epoch forced-photometry catalog. "
1851 "By default, it is the output of ForcedPhotCcdTask on calexps",
1852 name="forced_src",
1853 storageClass="SourceCatalog",
1854 dimensions=("instrument", "visit", "detector", "skymap", "tract")
1855 )
1856 inputCatalogDiff = connectionTypes.Input(
1857 doc="Secondary multi-epoch, per-detector, forced photometry catalog. "
1858 "By default, it is the output of ForcedPhotCcdTask run on image differences.",
1859 name="forced_diff",
1860 storageClass="SourceCatalog",
1861 dimensions=("instrument", "visit", "detector", "skymap", "tract")
1862 )
1863 outputCatalog = connectionTypes.Output(
1864 doc="InputCatalogs horizonatally joined on `objectId` in Parquet format",
1865 name="mergedForcedSource",
1866 storageClass="DataFrame",
1867 dimensions=("instrument", "visit", "detector", "skymap", "tract")
1868 )
1871class WriteForcedSourceTableConfig(pipeBase.PipelineTaskConfig,
1872 pipelineConnections=WriteForcedSourceTableConnections):
1873 key = lsst.pex.config.Field(
1874 doc="Column on which to join the two input tables on and make the primary key of the output",
1875 dtype=str,
1876 default="objectId",
1877 )
1880class WriteForcedSourceTableTask(pipeBase.PipelineTask):
1881 """Merge and convert per-detector forced source catalogs to parquet
1883 Because the predecessor ForcedPhotCcdTask operates per-detector,
1884 per-tract, (i.e., it has tract in its dimensions), detectors
1885 on the tract boundary may have multiple forced source catalogs.
1887 The successor task TransformForcedSourceTable runs per-patch
1888 and temporally-aggregates overlapping mergedForcedSource catalogs from all
1889 available multiple epochs.
1890 """
1891 _DefaultName = "writeForcedSourceTable"
1892 ConfigClass = WriteForcedSourceTableConfig
1894 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1895 inputs = butlerQC.get(inputRefs)
1896 # Add ccdVisitId to allow joining with CcdVisitTable
1897 inputs['ccdVisitId'] = butlerQC.quantum.dataId.pack("visit_detector")
1898 inputs['band'] = butlerQC.quantum.dataId.full['band']
1899 outputs = self.run(**inputs)
1900 butlerQC.put(outputs, outputRefs)
1902 def run(self, inputCatalog, inputCatalogDiff, ccdVisitId=None, band=None):
1903 dfs = []
1904 for table, dataset, in zip((inputCatalog, inputCatalogDiff), ('calexp', 'diff')):
1905 df = table.asAstropy().to_pandas().set_index(self.config.key, drop=False)
1906 df = df.reindex(sorted(df.columns), axis=1)
1907 df['ccdVisitId'] = ccdVisitId if ccdVisitId else pd.NA
1908 df['band'] = band if band else pd.NA
1909 df.columns = pd.MultiIndex.from_tuples([(dataset, c) for c in df.columns],
1910 names=('dataset', 'column'))
1912 dfs.append(df)
1914 outputCatalog = functools.reduce(lambda d1, d2: d1.join(d2), dfs)
1915 return pipeBase.Struct(outputCatalog=outputCatalog)
1918class TransformForcedSourceTableConnections(pipeBase.PipelineTaskConnections,
1919 dimensions=("instrument", "skymap", "patch", "tract")):
1921 inputCatalogs = connectionTypes.Input(
1922 doc="Parquet table of merged ForcedSources produced by WriteForcedSourceTableTask",
1923 name="mergedForcedSource",
1924 storageClass="DataFrame",
1925 dimensions=("instrument", "visit", "detector", "skymap", "tract"),
1926 multiple=True,
1927 deferLoad=True
1928 )
1929 referenceCatalog = connectionTypes.Input(
1930 doc="Reference catalog which was used to seed the forcedPhot. Columns "
1931 "objectId, detect_isPrimary, detect_isTractInner, detect_isPatchInner "
1932 "are expected.",
1933 name="objectTable",
1934 storageClass="DataFrame",
1935 dimensions=("tract", "patch", "skymap"),
1936 deferLoad=True
1937 )
1938 outputCatalog = connectionTypes.Output(
1939 doc="Narrower, temporally-aggregated, per-patch ForcedSource Table transformed and converted per a "
1940 "specified set of functors",
1941 name="forcedSourceTable",
1942 storageClass="DataFrame",
1943 dimensions=("tract", "patch", "skymap")
1944 )
1947class TransformForcedSourceTableConfig(TransformCatalogBaseConfig,
1948 pipelineConnections=TransformForcedSourceTableConnections):
1949 referenceColumns = pexConfig.ListField(
1950 dtype=str,
1951 default=["detect_isPrimary", "detect_isTractInner", "detect_isPatchInner"],
1952 optional=True,
1953 doc="Columns to pull from reference catalog",
1954 )
1955 keyRef = lsst.pex.config.Field(
1956 doc="Column on which to join the two input tables on and make the primary key of the output",
1957 dtype=str,
1958 default="objectId",
1959 )
1960 key = lsst.pex.config.Field(
1961 doc="Rename the output DataFrame index to this name",
1962 dtype=str,
1963 default="forcedSourceId",
1964 )
1966 def setDefaults(self):
1967 super().setDefaults()
1968 self.functorFile = os.path.join('$PIPE_TASKS_DIR', 'schemas', 'ForcedSource.yaml')
1969 self.columnsFromDataId = ['tract', 'patch']
1972class TransformForcedSourceTableTask(TransformCatalogBaseTask):
1973 """Transform/standardize a ForcedSource catalog
1975 Transforms each wide, per-detector forcedSource parquet table per the
1976 specification file (per-camera defaults found in ForcedSource.yaml).
1977 All epochs that overlap the patch are aggregated into one per-patch
1978 narrow-parquet file.
1980 No de-duplication of rows is performed. Duplicate resolutions flags are
1981 pulled in from the referenceCatalog: `detect_isPrimary`,
1982 `detect_isTractInner`,`detect_isPatchInner`, so that user may de-duplicate
1983 for analysis or compare duplicates for QA.
1985 The resulting table includes multiple bands. Epochs (MJDs) and other useful
1986 per-visit rows can be retreived by joining with the CcdVisitTable on
1987 ccdVisitId.
1988 """
1989 _DefaultName = "transformForcedSourceTable"
1990 ConfigClass = TransformForcedSourceTableConfig
1992 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1993 inputs = butlerQC.get(inputRefs)
1994 if self.funcs is None:
1995 raise ValueError("config.functorFile is None. "
1996 "Must be a valid path to yaml in order to run Task as a PipelineTask.")
1997 outputs = self.run(inputs['inputCatalogs'], inputs['referenceCatalog'], funcs=self.funcs,
1998 dataId=outputRefs.outputCatalog.dataId.full)
2000 butlerQC.put(outputs, outputRefs)
2002 def run(self, inputCatalogs, referenceCatalog, funcs=None, dataId=None, band=None):
2003 dfs = []
2004 ref = referenceCatalog.get(parameters={"columns": self.config.referenceColumns})
2005 self.log.info("Aggregating %s input catalogs" % (len(inputCatalogs)))
2006 for handle in inputCatalogs:
2007 result = self.transform(None, handle, funcs, dataId)
2008 # Filter for only rows that were detected on (overlap) the patch
2009 dfs.append(result.df.join(ref, how='inner'))
2011 outputCatalog = pd.concat(dfs)
2013 # Now that we are done joining on config.keyRef
2014 # Change index to config.key by
2015 outputCatalog.index.rename(self.config.keyRef, inplace=True)
2016 # Add config.keyRef to the column list
2017 outputCatalog.reset_index(inplace=True)
2018 # set the forcedSourceId to the index. This is specified in the ForcedSource.yaml
2019 outputCatalog.set_index("forcedSourceId", inplace=True, verify_integrity=True)
2020 # Rename it to the config.key
2021 outputCatalog.index.rename(self.config.key, inplace=True)
2023 self.log.info("Made a table of %d columns and %d rows",
2024 len(outputCatalog.columns), len(outputCatalog))
2025 return pipeBase.Struct(outputCatalog=outputCatalog)
2028class ConsolidateTractConnections(pipeBase.PipelineTaskConnections,
2029 defaultTemplates={"catalogType": ""},
2030 dimensions=("instrument", "tract")):
2031 inputCatalogs = connectionTypes.Input(
2032 doc="Input per-patch DataFrame Tables to be concatenated",
2033 name="{catalogType}ForcedSourceTable",
2034 storageClass="DataFrame",
2035 dimensions=("tract", "patch", "skymap"),
2036 multiple=True,
2037 )
2039 outputCatalog = connectionTypes.Output(
2040 doc="Output per-tract concatenation of DataFrame Tables",
2041 name="{catalogType}ForcedSourceTable_tract",
2042 storageClass="DataFrame",
2043 dimensions=("tract", "skymap"),
2044 )
2047class ConsolidateTractConfig(pipeBase.PipelineTaskConfig,
2048 pipelineConnections=ConsolidateTractConnections):
2049 pass
2052class ConsolidateTractTask(CmdLineTask, pipeBase.PipelineTask):
2053 """Concatenate any per-patch, dataframe list into a single
2054 per-tract DataFrame
2055 """
2056 _DefaultName = 'ConsolidateTract'
2057 ConfigClass = ConsolidateTractConfig
2059 def runQuantum(self, butlerQC, inputRefs, outputRefs):
2060 inputs = butlerQC.get(inputRefs)
2061 # Not checking at least one inputCatalog exists because that'd be an empty QG
2062 self.log.info("Concatenating %s per-patch %s Tables",
2063 len(inputs['inputCatalogs']),
2064 inputRefs.inputCatalogs[0].datasetType.name)
2065 df = pd.concat(inputs['inputCatalogs'])
2066 butlerQC.put(pipeBase.Struct(outputCatalog=df), outputRefs)