lsst.pipe.tasks g93147bea04+fdc1f4635d
postprocess.py
Go to the documentation of this file.
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/>.
21
22import functools
23import pandas as pd
24from collections import defaultdict
25import logging
26import numpy as np
27import numbers
28import os
29
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
42
43from .parquetTable import ParquetTable
44from .multiBandUtils import makeMergeArgumentParser, MergeSourcesRunner
45from .functors import CompositeFunctor, Column
46
47log = logging.getLogger(__name__)
48
49
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)
63
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
70
71
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 )
103
104
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 )
117
118
119class WriteObjectTableTask(CmdLineTask, pipeBase.PipelineTask):
120 """Write filter-merged source tables to parquet
121 """
122 _DefaultName = "writeObjectTable"
123 ConfigClass = WriteObjectTableConfig
124 RunnerClass = MergeSourcesRunner
125
126 # Names of table datasets to be merged
127 inputDatasets = ('forced_src', 'meas', 'ref')
128
129 # Tag of output dataset written by `MergeSourcesTask.write`
130 outputDataset = 'obj'
131
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)
137
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))
148
149 def runQuantum(self, butlerQC, inputRefs, outputRefs):
150 inputs = butlerQC.get(inputRefs)
151
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'])}
156
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)
166
167 @classmethod
168 def _makeArgumentParser(cls):
169 """Create a suitable ArgumentParser.
170
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.
174
175 References first of self.inputDatasets, rather than
176 self.inputDataset
177 """
178 return makeMergeArgumentParser(cls._DefaultName, cls.inputDatasets[0])
179
180 def readCatalog(self, patchRef):
181 """Read input catalogs
182
183 Read all the input datasets given by the 'inputDatasets'
184 attribute.
185
186 Parameters
187 ----------
188 patchRef : `lsst.daf.persistence.ButlerDataRef`
189 Data reference for patch
190
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_filterLabel", 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
204
205 def run(self, catalogs, tract, patch):
206 """Merge multiple catalogs.
207
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
216
217 Returns
218 -------
219 catalog : `pandas.DataFrame`
220 Merged dataframe
221 """
222
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)
228
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
233
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)
238
239 catalog = functools.reduce(lambda d1, d2: d1.join(d2), dfs)
240 return catalog
241
242 def write(self, patchRef, catalog):
243 """Write the output.
244
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)
258
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
263
264
265class WriteSourceTableConnections(pipeBase.PipelineTaskConnections,
266 defaultTemplates={"catalogType": ""},
267 dimensions=("instrument", "visit", "detector")):
268
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 )
282
283
284class WriteSourceTableConfig(pipeBase.PipelineTaskConfig,
285 pipelineConnections=WriteSourceTableConnections):
286 pass
287
288
289class WriteSourceTableTask(CmdLineTask, pipeBase.PipelineTask):
290 """Write source table to parquet
291 """
292 _DefaultName = "writeSourceTable"
293 ConfigClass = WriteSourceTableConfig
294
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)
301
302 def run(self, catalog, ccdVisitId=None, **kwargs):
303 """Convert `src` catalog to parquet
304
305 Parameters
306 ----------
307 catalog: `afwTable.SourceCatalog`
308 catalog to be converted
309 ccdVisitId: `int`
310 ccdVisitId to be added as a column
311
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))
322
323
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 )
373
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")
394
395
396class WriteRecalibratedSourceTableConfig(WriteSourceTableConfig,
397 pipelineConnections=WriteRecalibratedSourceTableConnections):
398
399 doReevaluatePhotoCalib = pexConfig.Field(
400 dtype=bool,
401 default=False,
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=False,
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=False,
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=False,
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 )
438
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.")
447
448
449class WriteRecalibratedSourceTableTask(WriteSourceTableTask):
450 """Write source table to parquet
451 """
452 _DefaultName = "writeRecalibratedSourceTable"
453 ConfigClass = WriteRecalibratedSourceTableConfig
454
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")
459
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)
463
464 inputs['catalog'] = self.addCalibColumns(**inputs)
465
466 result = self.run(**inputs).table
467 outputs = pipeBase.Struct(outputCatalog=result.toDataFrame())
468 butlerQC.put(outputs, outputRefs)
469
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
474
475 When multiple tract-level calibrations overlap, select the one with the
476 center closest to detector.
477
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
492
493
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])
518
519 externalSkyWcsCatalog = externalSkyWcsTractCatalog[ind]
520
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])
540
541 externalPhotoCalibCatalog = externalPhotoCalibTractCatalog[ind]
542
543 return self.prepareCalibratedExposure(exposure, externalSkyWcsCatalog, externalPhotoCalibCatalog)
544
545 def getClosestTract(self, tracts, skyMap, bbox, wcs):
546 """Find the index of the tract closest to detector from list of tractIds
547
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
557 Detector Wcs object to map the detector center to SkyCoord
558
559 Returns
560 -------
561 index : `int`
562 """
563 if len(tracts) == 1:
564 return 0
565
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))
572
573 return np.argmin(sep)
574
575 def prepareCalibratedExposure(self, exposure, externalSkyWcsCatalog=None, externalPhotoCalibCatalog=None):
576 """Prepare a calibrated exposure and apply external calibrations
577 if so configured.
578
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.
591
592 Returns
593 -------
594 exposure : `lsst.afw.image.exposure.Exposure`
595 Exposure with adjusted calibrations.
596 """
597 detectorId = exposure.getInfo().getDetector().getId()
598
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)
611
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)
624
625 return exposure
626
627 def addCalibColumns(self, catalog, exposure, exposureIdInfo, **kwargs):
628 """Add replace columns with calibs evaluated at each centroid
629
630 Add or replace 'base_LocalWcs' `base_LocalPhotoCalib' columns in a
631 a source catalog, by rerunning the plugins.
632
633 Parameters
634 ----------
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`
641
642 Returns
643 -------
645 Source Catalog with requested local calib columns
646 """
647 measureConfig = SingleFrameMeasurementTask.ConfigClass()
648 measureConfig.doReplaceWithNoise = False
649
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)
658
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)
666
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)
672
673 measurement.run(measCat=newCat, exposure=exposure, exposureId=exposureIdInfo.expId)
674
675 return newCat
676
677
678class PostprocessAnalysis(object):
679 """Calculate columns from ParquetTable
680
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`).
687
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.
691
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.
697
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.
700
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).
705
706 Parameters
707 ----------
708 parq : `lsst.pipe.tasks.ParquetTable` (or list of such)
709 Source catalog(s) for computation
710
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.
717
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.
722
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.
726
727 refFlags : `list` (optional)
728 List of refFlags (only reference band) to include in output table.
729
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 = ()
738
739 def __init__(self, parq, functors, filt=None, flags=None, refFlags=None, forcedFlags=None):
740 self.parq = parq
741 self.functors = functors
742
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)
749
750 self._df = None
751
752 @property
753 def defaultFuncs(self):
754 funcs = dict(self._defaultFuncs)
755 return funcs
756
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})
763
764 if isinstance(self.functors, CompositeFunctor):
765 func = self.functors
766 else:
767 func = CompositeFunctor(self.functors)
768
769 func.funcDict.update(additionalFuncs)
770 func.filt = self.filt
771
772 return func
773
774 @property
775 def noDupCols(self):
776 return [name for name, func in self.func.funcDict.items() if func.noDup or func.dataset == 'ref']
777
778 @property
779 def df(self):
780 if self._df is None:
781 self.compute()
782 return self._df
783
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)
795
796 return self._df
797
798
799class TransformCatalogBaseConnections(pipeBase.PipelineTaskConnections,
800 dimensions=()):
801 """Expected Connections for subclasses of TransformCatalogBaseTask.
802
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 )
813
814
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
832
833class TransformCatalogBaseTask(CmdLineTask, pipeBase.PipelineTask):
834 """Base class for transforming/standardizing a catalog
835
836 by applying functors that convert units and apply calibrations.
837 The purpose of this task is to perform a set of computations on
838 an input `ParquetTable` dataset (such as `deepCoadd_obj`) and write the
839 results to a new dataset (which needs to be declared in an `outputDataset`
840 attribute).
841
842 The calculations to be performed are defined in a YAML file that specifies
843 a set of functors to be computed, provided as
844 a `--functorFile` config parameter. An example of such a YAML file
845 is the following:
846
847 funcs:
848 psfMag:
849 functor: Mag
850 args:
851 - base_PsfFlux
852 filt: HSC-G
853 dataset: meas
854 cmodel_magDiff:
855 functor: MagDiff
856 args:
857 - modelfit_CModel
858 - base_PsfFlux
859 filt: HSC-G
860 gauss_magDiff:
861 functor: MagDiff
862 args:
863 - base_GaussianFlux
864 - base_PsfFlux
865 filt: HSC-G
866 count:
867 functor: Column
868 args:
869 - base_InputCount_value
870 filt: HSC-G
871 deconvolved_moments:
872 functor: DeconvolvedMoments
873 filt: HSC-G
874 dataset: forced_src
875 refFlags:
876 - calib_psfUsed
877 - merge_measurement_i
878 - merge_measurement_r
879 - merge_measurement_z
880 - merge_measurement_y
881 - merge_measurement_g
882 - base_PixelFlags_flag_inexact_psfCenter
883 - detect_isPrimary
884
885 The names for each entry under "func" will become the names of columns in the
886 output dataset. All the functors referenced are defined in `lsst.pipe.tasks.functors`.
887 Positional arguments to be passed to each functor are in the `args` list,
888 and any additional entries for each column other than "functor" or "args" (e.g., `'filt'`,
889 `'dataset'`) are treated as keyword arguments to be passed to the functor initialization.
890
891 The "flags" entry is the default shortcut for `Column` functors.
892 All columns listed under "flags" will be copied to the output table
893 untransformed. They can be of any datatype.
894 In the special case of transforming a multi-level oject table with
895 band and dataset indices (deepCoadd_obj), these will be taked from the
896 `meas` dataset and exploded out per band.
897
898 There are two special shortcuts that only apply when transforming
899 multi-level Object (deepCoadd_obj) tables:
900 - The "refFlags" entry is shortcut for `Column` functor
901 taken from the `'ref'` dataset if transforming an ObjectTable.
902 - The "forcedFlags" entry is shortcut for `Column` functors.
903 taken from the ``forced_src`` dataset if transforming an ObjectTable.
904 These are expanded out per band.
905
906
907 This task uses the `lsst.pipe.tasks.postprocess.PostprocessAnalysis` object
908 to organize and excecute the calculations.
909
910 """
911 @property
912 def _DefaultName(self):
913 raise NotImplementedError('Subclass must define "_DefaultName" attribute')
914
915 @property
916 def outputDataset(self):
917 raise NotImplementedError('Subclass must define "outputDataset" attribute')
918
919 @property
920 def inputDataset(self):
921 raise NotImplementedError('Subclass must define "inputDataset" attribute')
922
923 @property
924 def ConfigClass(self):
925 raise NotImplementedError('Subclass must define "ConfigClass" attribute')
926
927 def __init__(self, *args, **kwargs):
928 super().__init__(*args, **kwargs)
929 if self.config.functorFile:
930 self.log.info('Loading tranform functor definitions from %s',
931 self.config.functorFile)
932 self.funcsfuncs = CompositeFunctor.from_file(self.config.functorFile)
933 self.funcsfuncs.update(dict(PostprocessAnalysis._defaultFuncs))
934 else:
935 self.funcsfuncs = None
936
937 def runQuantum(self, butlerQC, inputRefs, outputRefs):
938 inputs = butlerQC.get(inputRefs)
939 if self.funcsfuncs is None:
940 raise ValueError("config.functorFile is None. "
941 "Must be a valid path to yaml in order to run Task as a PipelineTask.")
942 result = self.runrun(parq=inputs['inputCatalog'], funcs=self.funcsfuncs,
943 dataId=outputRefs.outputCatalog.dataId.full)
944 outputs = pipeBase.Struct(outputCatalog=result)
945 butlerQC.put(outputs, outputRefs)
946
947 def runDataRef(self, dataRef):
948 parq = dataRef.get()
949 if self.funcsfuncs is None:
950 raise ValueError("config.functorFile is None. "
951 "Must be a valid path to yaml in order to run as a CommandlineTask.")
952 df = self.runrun(parq, funcs=self.funcsfuncs, dataId=dataRef.dataId)
953 self.writewrite(df, dataRef)
954 return df
955
956 def run(self, parq, funcs=None, dataId=None, band=None):
957 """Do postprocessing calculations
958
959 Takes a `ParquetTable` object and dataId,
960 returns a dataframe with results of postprocessing calculations.
961
962 Parameters
963 ----------
965 ParquetTable from which calculations are done.
966 funcs : `lsst.pipe.tasks.functors.Functors`
967 Functors to apply to the table's columns
968 dataId : dict, optional
969 Used to add a `patchId` column to the output dataframe.
970 band : `str`, optional
971 Filter band that is being processed.
972
973 Returns
974 ------
975 `pandas.DataFrame`
976
977 """
978 self.log.info("Transforming/standardizing the source table dataId: %s", dataId)
979
980 df = self.transformtransform(band, parq, funcs, dataId).df
981 self.log.info("Made a table of %d columns and %d rows", len(df.columns), len(df))
982 return df
983
984 def getFunctors(self):
985 return self.funcsfuncs
986
987 def getAnalysis(self, parq, funcs=None, band=None):
988 if funcs is None:
989 funcs = self.funcsfuncs
990 analysis = PostprocessAnalysis(parq, funcs, filt=band)
991 return analysis
992
993 def transform(self, band, parq, funcs, dataId):
994 analysis = self.getAnalysisgetAnalysis(parq, funcs=funcs, band=band)
995 df = analysis.df
996 if dataId is not None:
997 for key, value in dataId.items():
998 df[str(key)] = value
999
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)
1004
1005 return pipeBase.Struct(
1006 df=df,
1007 analysis=analysis
1008 )
1009
1010 def write(self, df, parqRef):
1011 parqRef.put(ParquetTable(dataFrame=df), self.outputDatasetoutputDataset)
1012
1013 def writeMetadata(self, dataRef):
1014 """No metadata to write.
1015 """
1016 pass
1017
1018
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 )
1037
1038
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 )
1093
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.goodFlags = ['calib_astrometry_used',
1099 'calib_photometry_reserved',
1100 'calib_photometry_used',
1101 'calib_psf_candidate',
1102 'calib_psf_reserved',
1103 'calib_psf_used']
1104
1105
1106class TransformObjectCatalogTask(TransformCatalogBaseTask):
1107 """Produce a flattened Object Table to match the format specified in
1108 sdm_schemas.
1109
1110 Do the same set of postprocessing calculations on all bands
1111
1112 This is identical to `TransformCatalogBaseTask`, except for that it does the
1113 specified functor calculations for all filters present in the
1114 input `deepCoadd_obj` table. Any specific `"filt"` keywords specified
1115 by the YAML file will be superceded.
1116 """
1117 _DefaultName = "transformObjectCatalog"
1118 ConfigClass = TransformObjectCatalogConfig
1119
1120 # Used by Gen 2 runDataRef only:
1121 inputDataset = 'deepCoadd_obj'
1122 outputDataset = 'objectTable'
1123
1124 @classmethod
1125 def _makeArgumentParser(cls):
1126 parser = ArgumentParser(name=cls._DefaultName)
1127 parser.add_id_argument("--id", cls.inputDataset,
1128 ContainerClass=CoaddDataIdContainer,
1129 help="data ID, e.g. --id tract=12345 patch=1,2")
1130 return parser
1131
1132 def run(self, parq, funcs=None, dataId=None, band=None):
1133 # NOTE: band kwarg is ignored here.
1134 dfDict = {}
1135 analysisDict = {}
1136 templateDf = pd.DataFrame()
1137
1138 if isinstance(parq, DeferredDatasetHandle):
1139 columns = parq.get(component='columns')
1140 inputBands = columns.unique(level=1).values
1141 else:
1142 inputBands = parq.columnLevelNames['band']
1143
1144 outputBands = self.config.outputBands if self.config.outputBands else inputBands
1145
1146 # Perform transform for data of filters that exist in parq.
1147 for inputBand in inputBands:
1148 if inputBand not in outputBands:
1149 self.log.info("Ignoring %s band data in the input", inputBand)
1150 continue
1151 self.log.info("Transforming the catalog of band %s", inputBand)
1152 result = self.transform(inputBand, parq, funcs, dataId)
1153 dfDict[inputBand] = result.df
1154 analysisDict[inputBand] = result.analysis
1155 if templateDf.empty:
1156 templateDf = result.df
1157
1158 # Put filler values in columns of other wanted bands
1159 for filt in outputBands:
1160 if filt not in dfDict:
1161 self.log.info("Adding empty columns for band %s", filt)
1162 dfTemp = templateDf.copy()
1163 for col in dfTemp.columns:
1164 testValue = dfTemp[col].values[0]
1165 if isinstance(testValue, (np.bool_, pd.BooleanDtype)):
1166 # Boolean flag type, check if it is a "good" flag
1167 if col in self.config.goodFlags:
1168 fillValue = False
1169 else:
1170 fillValue = True
1171 elif isinstance(testValue, numbers.Integral):
1172 # Checking numbers.Integral catches all flavors
1173 # of python, numpy, pandas, etc. integers.
1174 # We must ensure this is not an unsigned integer.
1175 if isinstance(testValue, np.unsignedinteger):
1176 raise ValueError("Parquet tables may not have unsigned integer columns.")
1177 else:
1178 fillValue = self.config.integerFillValue
1179 else:
1180 fillValue = self.config.floatFillValue
1181 dfTemp[col].values[:] = fillValue
1182 dfDict[filt] = dfTemp
1183
1184 # This makes a multilevel column index, with band as first level
1185 df = pd.concat(dfDict, axis=1, names=['band', 'column'])
1186
1187 if not self.config.multilevelOutput:
1188 noDupCols = list(set.union(*[set(v.noDupCols) for v in analysisDict.values()]))
1189 if self.config.primaryKey in noDupCols:
1190 noDupCols.remove(self.config.primaryKey)
1191 if dataId is not None:
1192 noDupCols += list(dataId.keys())
1193 df = flattenFilters(df, noDupCols=noDupCols, camelCase=self.config.camelCase,
1194 inputBands=inputBands)
1195
1196 self.log.info("Made a table of %d columns and %d rows", len(df.columns), len(df))
1197
1198 return df
1199
1200
1201class TractObjectDataIdContainer(CoaddDataIdContainer):
1202
1203 def makeDataRefList(self, namespace):
1204 """Make self.refList from self.idList
1205
1206 Generate a list of data references given tract and/or patch.
1207 This was adapted from `TractQADataIdContainer`, which was
1208 `TractDataIdContainer` modifie to not require "filter".
1209 Only existing dataRefs are returned.
1210 """
1211 def getPatchRefList(tract):
1212 return [namespace.butler.dataRef(datasetType=self.datasetType,
1213 tract=tract.getId(),
1214 patch="%d,%d" % patch.getIndex()) for patch in tract]
1215
1216 tractRefs = defaultdict(list) # Data references for each tract
1217 for dataId in self.idList:
1218 skymap = self.getSkymap(namespace)
1219
1220 if "tract" in dataId:
1221 tractId = dataId["tract"]
1222 if "patch" in dataId:
1223 tractRefs[tractId].append(namespace.butler.dataRef(datasetType=self.datasetType,
1224 tract=tractId,
1225 patch=dataId['patch']))
1226 else:
1227 tractRefs[tractId] += getPatchRefList(skymap[tractId])
1228 else:
1229 tractRefs = dict((tract.getId(), tractRefs.get(tract.getId(), []) + getPatchRefList(tract))
1230 for tract in skymap)
1231 outputRefList = []
1232 for tractRefList in tractRefs.values():
1233 existingRefs = [ref for ref in tractRefList if ref.datasetExists()]
1234 outputRefList.append(existingRefs)
1235
1236 self.refList = outputRefList
1237
1238
1239class ConsolidateObjectTableConnections(pipeBase.PipelineTaskConnections,
1240 dimensions=("tract", "skymap")):
1241 inputCatalogs = connectionTypes.Input(
1242 doc="Per-Patch objectTables conforming to the standard data model.",
1243 name="objectTable",
1244 storageClass="DataFrame",
1245 dimensions=("tract", "patch", "skymap"),
1246 multiple=True,
1247 )
1248 outputCatalog = connectionTypes.Output(
1249 doc="Pre-tract horizontal concatenation of the input objectTables",
1250 name="objectTable_tract",
1251 storageClass="DataFrame",
1252 dimensions=("tract", "skymap"),
1253 )
1254
1255
1256class ConsolidateObjectTableConfig(pipeBase.PipelineTaskConfig,
1257 pipelineConnections=ConsolidateObjectTableConnections):
1258 coaddName = pexConfig.Field(
1259 dtype=str,
1260 default="deep",
1261 doc="Name of coadd"
1262 )
1263
1264
1265class ConsolidateObjectTableTask(CmdLineTask, pipeBase.PipelineTask):
1266 """Write patch-merged source tables to a tract-level parquet file
1267
1268 Concatenates `objectTable` list into a per-visit `objectTable_tract`
1269 """
1270 _DefaultName = "consolidateObjectTable"
1271 ConfigClass = ConsolidateObjectTableConfig
1272
1273 inputDataset = 'objectTable'
1274 outputDataset = 'objectTable_tract'
1275
1276 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1277 inputs = butlerQC.get(inputRefs)
1278 self.log.info("Concatenating %s per-patch Object Tables",
1279 len(inputs['inputCatalogs']))
1280 df = pd.concat(inputs['inputCatalogs'])
1281 butlerQC.put(pipeBase.Struct(outputCatalog=df), outputRefs)
1282
1283 @classmethod
1284 def _makeArgumentParser(cls):
1285 parser = ArgumentParser(name=cls._DefaultName)
1286
1287 parser.add_id_argument("--id", cls.inputDataset,
1288 help="data ID, e.g. --id tract=12345",
1289 ContainerClass=TractObjectDataIdContainer)
1290 return parser
1291
1292 def runDataRef(self, patchRefList):
1293 df = pd.concat([patchRef.get().toDataFrame() for patchRef in patchRefList])
1294 patchRefList[0].put(ParquetTable(dataFrame=df), self.outputDataset)
1295
1296 def writeMetadata(self, dataRef):
1297 """No metadata to write.
1298 """
1299 pass
1300
1301
1302class TransformSourceTableConnections(pipeBase.PipelineTaskConnections,
1303 defaultTemplates={"catalogType": ""},
1304 dimensions=("instrument", "visit", "detector")):
1305
1306 inputCatalog = connectionTypes.Input(
1307 doc="Wide input catalog of sources produced by WriteSourceTableTask",
1308 name="{catalogType}source",
1309 storageClass="DataFrame",
1310 dimensions=("instrument", "visit", "detector"),
1311 deferLoad=True
1312 )
1313 outputCatalog = connectionTypes.Output(
1314 doc="Narrower, per-detector Source Table transformed and converted per a "
1315 "specified set of functors",
1316 name="{catalogType}sourceTable",
1317 storageClass="DataFrame",
1318 dimensions=("instrument", "visit", "detector")
1319 )
1320
1321
1322class TransformSourceTableConfig(TransformCatalogBaseConfig,
1323 pipelineConnections=TransformSourceTableConnections):
1324
1325 def setDefaults(self):
1326 super().setDefaults()
1327 self.functorFile = os.path.join('$PIPE_TASKS_DIR', 'schemas', 'Source.yaml')
1328 self.primaryKey = 'sourceId'
1329
1330
1331class TransformSourceTableTask(TransformCatalogBaseTask):
1332 """Transform/standardize a source catalog
1333 """
1334 _DefaultName = "transformSourceTable"
1335 ConfigClass = TransformSourceTableConfig
1336
1337 inputDataset = 'source'
1338 outputDataset = 'sourceTable'
1339
1340 @classmethod
1341 def _makeArgumentParser(cls):
1342 parser = ArgumentParser(name=cls._DefaultName)
1343 parser.add_id_argument("--id", datasetType=cls.inputDataset,
1344 level="sensor",
1345 help="data ID, e.g. --id visit=12345 ccd=0")
1346 return parser
1347
1348 def runDataRef(self, dataRef):
1349 """Override to specify band label to run()."""
1350 parq = dataRef.get()
1351 funcs = self.getFunctors()
1352 band = dataRef.get("calexp_filterLabel", immediate=True).bandLabel
1353 df = self.run(parq, funcs=funcs, dataId=dataRef.dataId, band=band)
1354 self.write(df, dataRef)
1355 return df
1356
1357
1358class ConsolidateVisitSummaryConnections(pipeBase.PipelineTaskConnections,
1359 dimensions=("instrument", "visit",),
1360 defaultTemplates={"calexpType": ""}):
1361 calexp = connectionTypes.Input(
1362 doc="Processed exposures used for metadata",
1363 name="{calexpType}calexp",
1364 storageClass="ExposureF",
1365 dimensions=("instrument", "visit", "detector"),
1366 deferLoad=True,
1367 multiple=True,
1368 )
1369 visitSummary = connectionTypes.Output(
1370 doc=("Per-visit consolidated exposure metadata. These catalogs use "
1371 "detector id for the id and are sorted for fast lookups of a "
1372 "detector."),
1373 name="{calexpType}visitSummary",
1374 storageClass="ExposureCatalog",
1375 dimensions=("instrument", "visit"),
1376 )
1377
1378
1379class ConsolidateVisitSummaryConfig(pipeBase.PipelineTaskConfig,
1380 pipelineConnections=ConsolidateVisitSummaryConnections):
1381 """Config for ConsolidateVisitSummaryTask"""
1382 pass
1383
1384
1385class ConsolidateVisitSummaryTask(pipeBase.PipelineTask, pipeBase.CmdLineTask):
1386 """Task to consolidate per-detector visit metadata.
1387
1388 This task aggregates the following metadata from all the detectors in a
1389 single visit into an exposure catalog:
1390 - The visitInfo.
1391 - The wcs.
1392 - The photoCalib.
1393 - The physical_filter and band (if available).
1394 - The psf size, shape, and effective area at the center of the detector.
1395 - The corners of the bounding box in right ascension/declination.
1396
1397 Other quantities such as Detector, Psf, ApCorrMap, and TransmissionCurve
1398 are not persisted here because of storage concerns, and because of their
1399 limited utility as summary statistics.
1400
1401 Tests for this task are performed in ci_hsc_gen3.
1402 """
1403 _DefaultName = "consolidateVisitSummary"
1404 ConfigClass = ConsolidateVisitSummaryConfig
1405
1406 @classmethod
1407 def _makeArgumentParser(cls):
1408 parser = ArgumentParser(name=cls._DefaultName)
1409
1410 parser.add_id_argument("--id", "calexp",
1411 help="data ID, e.g. --id visit=12345",
1412 ContainerClass=VisitDataIdContainer)
1413 return parser
1414
1415 def writeMetadata(self, dataRef):
1416 """No metadata to persist, so override to remove metadata persistance.
1417 """
1418 pass
1419
1420 def writeConfig(self, butler, clobber=False, doBackup=True):
1421 """No config to persist, so override to remove config persistance.
1422 """
1423 pass
1424
1425 def runDataRef(self, dataRefList):
1426 visit = dataRefList[0].dataId['visit']
1427
1428 self.log.debug("Concatenating metadata from %d per-detector calexps (visit %d)",
1429 len(dataRefList), visit)
1430
1431 expCatalog = self._combineExposureMetadata(visit, dataRefList, isGen3=False)
1432
1433 dataRefList[0].put(expCatalog, 'visitSummary', visit=visit)
1434
1435 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1436 dataRefs = butlerQC.get(inputRefs.calexp)
1437 visit = dataRefs[0].dataId.byName()['visit']
1438
1439 self.log.debug("Concatenating metadata from %d per-detector calexps (visit %d)",
1440 len(dataRefs), visit)
1441
1442 expCatalog = self._combineExposureMetadata(visit, dataRefs)
1443
1444 butlerQC.put(expCatalog, outputRefs.visitSummary)
1445
1446 def _combineExposureMetadata(self, visit, dataRefs, isGen3=True):
1447 """Make a combined exposure catalog from a list of dataRefs.
1448 These dataRefs must point to exposures with wcs, summaryStats,
1449 and other visit metadata.
1450
1451 Parameters
1452 ----------
1453 visit : `int`
1454 Visit identification number.
1455 dataRefs : `list`
1456 List of dataRefs in visit. May be list of
1457 `lsst.daf.persistence.ButlerDataRef` (Gen2) or
1458 `lsst.daf.butler.DeferredDatasetHandle` (Gen3).
1459 isGen3 : `bool`, optional
1460 Specifies if this is a Gen3 list of datarefs.
1461
1462 Returns
1463 -------
1464 visitSummary : `lsst.afw.table.ExposureCatalog`
1465 Exposure catalog with per-detector summary information.
1466 """
1467 schema = self._makeVisitSummarySchema()
1468 cat = afwTable.ExposureCatalog(schema)
1469 cat.resize(len(dataRefs))
1470
1471 cat['visit'] = visit
1472
1473 for i, dataRef in enumerate(dataRefs):
1474 if isGen3:
1475 visitInfo = dataRef.get(component='visitInfo')
1476 filterLabel = dataRef.get(component='filterLabel')
1477 summaryStats = dataRef.get(component='summaryStats')
1478 detector = dataRef.get(component='detector')
1479 wcs = dataRef.get(component='wcs')
1480 photoCalib = dataRef.get(component='photoCalib')
1481 detector = dataRef.get(component='detector')
1482 bbox = dataRef.get(component='bbox')
1483 validPolygon = dataRef.get(component='validPolygon')
1484 else:
1485 # Note that we need to read the calexp because there is
1486 # no magic access to the psf except through the exposure.
1487 gen2_read_bbox = lsst.geom.BoxI(lsst.geom.PointI(0, 0), lsst.geom.PointI(1, 1))
1488 exp = dataRef.get(datasetType='calexp_sub', bbox=gen2_read_bbox)
1489 visitInfo = exp.getInfo().getVisitInfo()
1490 filterLabel = dataRef.get("calexp_filterLabel")
1491 summaryStats = exp.getInfo().getSummaryStats()
1492 wcs = exp.getWcs()
1493 photoCalib = exp.getPhotoCalib()
1494 detector = exp.getDetector()
1495 bbox = dataRef.get(datasetType='calexp_bbox')
1496 validPolygon = exp.getInfo().getValidPolygon()
1497
1498 rec = cat[i]
1499 rec.setBBox(bbox)
1500 rec.setVisitInfo(visitInfo)
1501 rec.setWcs(wcs)
1502 rec.setPhotoCalib(photoCalib)
1503 rec.setValidPolygon(validPolygon)
1504
1505 rec['physical_filter'] = filterLabel.physicalLabel if filterLabel.hasPhysicalLabel() else ""
1506 rec['band'] = filterLabel.bandLabel if filterLabel.hasBandLabel() else ""
1507 rec.setId(detector.getId())
1508 rec['psfSigma'] = summaryStats.psfSigma
1509 rec['psfIxx'] = summaryStats.psfIxx
1510 rec['psfIyy'] = summaryStats.psfIyy
1511 rec['psfIxy'] = summaryStats.psfIxy
1512 rec['psfArea'] = summaryStats.psfArea
1513 rec['raCorners'][:] = summaryStats.raCorners
1514 rec['decCorners'][:] = summaryStats.decCorners
1515 rec['ra'] = summaryStats.ra
1516 rec['decl'] = summaryStats.decl
1517 rec['zenithDistance'] = summaryStats.zenithDistance
1518 rec['zeroPoint'] = summaryStats.zeroPoint
1519 rec['skyBg'] = summaryStats.skyBg
1520 rec['skyNoise'] = summaryStats.skyNoise
1521 rec['meanVar'] = summaryStats.meanVar
1522 rec['astromOffsetMean'] = summaryStats.astromOffsetMean
1523 rec['astromOffsetStd'] = summaryStats.astromOffsetStd
1524 rec['nPsfStar'] = summaryStats.nPsfStar
1525 rec['psfStarDeltaE1Median'] = summaryStats.psfStarDeltaE1Median
1526 rec['psfStarDeltaE2Median'] = summaryStats.psfStarDeltaE2Median
1527 rec['psfStarDeltaE1Scatter'] = summaryStats.psfStarDeltaE1Scatter
1528 rec['psfStarDeltaE2Scatter'] = summaryStats.psfStarDeltaE2Scatter
1529 rec['psfStarDeltaSizeMedian'] = summaryStats.psfStarDeltaSizeMedian
1530 rec['psfStarDeltaSizeScatter'] = summaryStats.psfStarDeltaSizeScatter
1531 rec['psfStarScaledDeltaSizeScatter'] = summaryStats.psfStarScaledDeltaSizeScatter
1532
1533 metadata = dafBase.PropertyList()
1534 metadata.add("COMMENT", "Catalog id is detector id, sorted.")
1535 # We are looping over existing datarefs, so the following is true
1536 metadata.add("COMMENT", "Only detectors with data have entries.")
1537 cat.setMetadata(metadata)
1538
1539 cat.sort()
1540 return cat
1541
1542 def _makeVisitSummarySchema(self):
1543 """Make the schema for the visitSummary catalog."""
1544 schema = afwTable.ExposureTable.makeMinimalSchema()
1545 schema.addField('visit', type='L', doc='Visit number')
1546 schema.addField('physical_filter', type='String', size=32, doc='Physical filter')
1547 schema.addField('band', type='String', size=32, doc='Name of band')
1548 schema.addField('psfSigma', type='F',
1549 doc='PSF model second-moments determinant radius (center of chip) (pixel)')
1550 schema.addField('psfArea', type='F',
1551 doc='PSF model effective area (center of chip) (pixel**2)')
1552 schema.addField('psfIxx', type='F',
1553 doc='PSF model Ixx (center of chip) (pixel**2)')
1554 schema.addField('psfIyy', type='F',
1555 doc='PSF model Iyy (center of chip) (pixel**2)')
1556 schema.addField('psfIxy', type='F',
1557 doc='PSF model Ixy (center of chip) (pixel**2)')
1558 schema.addField('raCorners', type='ArrayD', size=4,
1559 doc='Right Ascension of bounding box corners (degrees)')
1560 schema.addField('decCorners', type='ArrayD', size=4,
1561 doc='Declination of bounding box corners (degrees)')
1562 schema.addField('ra', type='D',
1563 doc='Right Ascension of bounding box center (degrees)')
1564 schema.addField('decl', type='D',
1565 doc='Declination of bounding box center (degrees)')
1566 schema.addField('zenithDistance', type='F',
1567 doc='Zenith distance of bounding box center (degrees)')
1568 schema.addField('zeroPoint', type='F',
1569 doc='Mean zeropoint in detector (mag)')
1570 schema.addField('skyBg', type='F',
1571 doc='Average sky background (ADU)')
1572 schema.addField('skyNoise', type='F',
1573 doc='Average sky noise (ADU)')
1574 schema.addField('meanVar', type='F',
1575 doc='Mean variance of the weight plane (ADU**2)')
1576 schema.addField('astromOffsetMean', type='F',
1577 doc='Mean offset of astrometric calibration matches (arcsec)')
1578 schema.addField('astromOffsetStd', type='F',
1579 doc='Standard deviation of offsets of astrometric calibration matches (arcsec)')
1580 schema.addField('nPsfStar', type='I', doc='Number of stars used for PSF model')
1581 schema.addField('psfStarDeltaE1Median', type='F',
1582 doc='Median E1 residual (starE1 - psfE1) for psf stars')
1583 schema.addField('psfStarDeltaE2Median', type='F',
1584 doc='Median E2 residual (starE2 - psfE2) for psf stars')
1585 schema.addField('psfStarDeltaE1Scatter', type='F',
1586 doc='Scatter (via MAD) of E1 residual (starE1 - psfE1) for psf stars')
1587 schema.addField('psfStarDeltaE2Scatter', type='F',
1588 doc='Scatter (via MAD) of E2 residual (starE2 - psfE2) for psf stars')
1589 schema.addField('psfStarDeltaSizeMedian', type='F',
1590 doc='Median size residual (starSize - psfSize) for psf stars (pixel)')
1591 schema.addField('psfStarDeltaSizeScatter', type='F',
1592 doc='Scatter (via MAD) of size residual (starSize - psfSize) for psf stars (pixel)')
1593 schema.addField('psfStarScaledDeltaSizeScatter', type='F',
1594 doc='Scatter (via MAD) of size residual scaled by median size squared')
1595
1596 return schema
1597
1598
1599class VisitDataIdContainer(DataIdContainer):
1600 """DataIdContainer that groups sensor-level id's by visit
1601 """
1602
1603 def makeDataRefList(self, namespace):
1604 """Make self.refList from self.idList
1605
1606 Generate a list of data references grouped by visit.
1607
1608 Parameters
1609 ----------
1610 namespace : `argparse.Namespace`
1611 Namespace used by `lsst.pipe.base.CmdLineTask` to parse command line arguments
1612 """
1613 # Group by visits
1614 visitRefs = defaultdict(list)
1615 for dataId in self.idList:
1616 if "visit" in dataId:
1617 visitId = dataId["visit"]
1618 # append all subsets to
1619 subset = namespace.butler.subset(self.datasetType, dataId=dataId)
1620 visitRefs[visitId].extend([dataRef for dataRef in subset])
1621
1622 outputRefList = []
1623 for refList in visitRefs.values():
1624 existingRefs = [ref for ref in refList if ref.datasetExists()]
1625 if existingRefs:
1626 outputRefList.append(existingRefs)
1627
1628 self.refList = outputRefList
1629
1630
1631class ConsolidateSourceTableConnections(pipeBase.PipelineTaskConnections,
1632 defaultTemplates={"catalogType": ""},
1633 dimensions=("instrument", "visit")):
1634 inputCatalogs = connectionTypes.Input(
1635 doc="Input per-detector Source Tables",
1636 name="{catalogType}sourceTable",
1637 storageClass="DataFrame",
1638 dimensions=("instrument", "visit", "detector"),
1639 multiple=True
1640 )
1641 outputCatalog = connectionTypes.Output(
1642 doc="Per-visit concatenation of Source Table",
1643 name="{catalogType}sourceTable_visit",
1644 storageClass="DataFrame",
1645 dimensions=("instrument", "visit")
1646 )
1647
1648
1649class ConsolidateSourceTableConfig(pipeBase.PipelineTaskConfig,
1650 pipelineConnections=ConsolidateSourceTableConnections):
1651 pass
1652
1653
1654class ConsolidateSourceTableTask(CmdLineTask, pipeBase.PipelineTask):
1655 """Concatenate `sourceTable` list into a per-visit `sourceTable_visit`
1656 """
1657 _DefaultName = 'consolidateSourceTable'
1658 ConfigClass = ConsolidateSourceTableConfig
1659
1660 inputDataset = 'sourceTable'
1661 outputDataset = 'sourceTable_visit'
1662
1663 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1664 from .makeCoaddTempExp import reorderRefs
1665
1666 detectorOrder = [ref.dataId['detector'] for ref in inputRefs.inputCatalogs]
1667 detectorOrder.sort()
1668 inputRefs = reorderRefs(inputRefs, detectorOrder, dataIdKey='detector')
1669 inputs = butlerQC.get(inputRefs)
1670 self.log.info("Concatenating %s per-detector Source Tables",
1671 len(inputs['inputCatalogs']))
1672 df = pd.concat(inputs['inputCatalogs'])
1673 butlerQC.put(pipeBase.Struct(outputCatalog=df), outputRefs)
1674
1675 def runDataRef(self, dataRefList):
1676 self.log.info("Concatenating %s per-detector Source Tables", len(dataRefList))
1677 df = pd.concat([dataRef.get().toDataFrame() for dataRef in dataRefList])
1678 dataRefList[0].put(ParquetTable(dataFrame=df), self.outputDataset)
1679
1680 @classmethod
1681 def _makeArgumentParser(cls):
1682 parser = ArgumentParser(name=cls._DefaultName)
1683
1684 parser.add_id_argument("--id", cls.inputDataset,
1685 help="data ID, e.g. --id visit=12345",
1686 ContainerClass=VisitDataIdContainer)
1687 return parser
1688
1689 def writeMetadata(self, dataRef):
1690 """No metadata to write.
1691 """
1692 pass
1693
1694 def writeConfig(self, butler, clobber=False, doBackup=True):
1695 """No config to write.
1696 """
1697 pass
1698
1699
1700class MakeCcdVisitTableConnections(pipeBase.PipelineTaskConnections,
1701 dimensions=("instrument",),
1702 defaultTemplates={"calexpType": ""}):
1703 visitSummaryRefs = connectionTypes.Input(
1704 doc="Data references for per-visit consolidated exposure metadata from ConsolidateVisitSummaryTask",
1705 name="{calexpType}visitSummary",
1706 storageClass="ExposureCatalog",
1707 dimensions=("instrument", "visit"),
1708 multiple=True,
1709 deferLoad=True,
1710 )
1711 outputCatalog = connectionTypes.Output(
1712 doc="CCD and Visit metadata table",
1713 name="ccdVisitTable",
1714 storageClass="DataFrame",
1715 dimensions=("instrument",)
1716 )
1717
1718
1719class MakeCcdVisitTableConfig(pipeBase.PipelineTaskConfig,
1720 pipelineConnections=MakeCcdVisitTableConnections):
1721 pass
1722
1723
1724class MakeCcdVisitTableTask(CmdLineTask, pipeBase.PipelineTask):
1725 """Produce a `ccdVisitTable` from the `visitSummary` exposure catalogs.
1726 """
1727 _DefaultName = 'makeCcdVisitTable'
1728 ConfigClass = MakeCcdVisitTableConfig
1729
1730 def run(self, visitSummaryRefs):
1731 """ Make a table of ccd information from the `visitSummary` catalogs.
1732 Parameters
1733 ----------
1734 visitSummaryRefs : `list` of `lsst.daf.butler.DeferredDatasetHandle`
1735 List of DeferredDatasetHandles pointing to exposure catalogs with
1736 per-detector summary information.
1737 Returns
1738 -------
1739 result : `lsst.pipe.Base.Struct`
1740 Results struct with attribute:
1741 - `outputCatalog`
1742 Catalog of ccd and visit information.
1743 """
1744 ccdEntries = []
1745 for visitSummaryRef in visitSummaryRefs:
1746 visitSummary = visitSummaryRef.get()
1747 visitInfo = visitSummary[0].getVisitInfo()
1748
1749 ccdEntry = {}
1750 summaryTable = visitSummary.asAstropy()
1751 selectColumns = ['id', 'visit', 'physical_filter', 'band', 'ra', 'decl', 'zenithDistance',
1752 'zeroPoint', 'psfSigma', 'skyBg', 'skyNoise']
1753 ccdEntry = summaryTable[selectColumns].to_pandas().set_index('id')
1754 # 'visit' is the human readible visit number
1755 # 'visitId' is the key to the visitId table. They are the same
1756 # Technically you should join to get the visit from the visit table
1757 ccdEntry = ccdEntry.rename(columns={"visit": "visitId"})
1758 dataIds = [DataCoordinate.standardize(visitSummaryRef.dataId, detector=id) for id in
1759 summaryTable['id']]
1760 packer = visitSummaryRef.dataId.universe.makePacker('visit_detector', visitSummaryRef.dataId)
1761 ccdVisitIds = [packer.pack(dataId) for dataId in dataIds]
1762 ccdEntry['ccdVisitId'] = ccdVisitIds
1763 ccdEntry['detector'] = summaryTable['id']
1764 pixToArcseconds = np.array([vR.getWcs().getPixelScale().asArcseconds() for vR in visitSummary])
1765 ccdEntry["seeing"] = visitSummary['psfSigma'] * np.sqrt(8 * np.log(2)) * pixToArcseconds
1766
1767 ccdEntry["skyRotation"] = visitInfo.getBoresightRotAngle().asDegrees()
1768 ccdEntry["expMidpt"] = visitInfo.getDate().toPython()
1769 ccdEntry["expMidptMJD"] = visitInfo.getDate().get(dafBase.DateTime.MJD)
1770 expTime = visitInfo.getExposureTime()
1771 ccdEntry['expTime'] = expTime
1772 ccdEntry["obsStart"] = ccdEntry["expMidpt"] - 0.5 * pd.Timedelta(seconds=expTime)
1773 expTime_days = expTime / (60*60*24)
1774 ccdEntry["obsStartMJD"] = ccdEntry["expMidptMJD"] - 0.5 * expTime_days
1775 ccdEntry['darkTime'] = visitInfo.getDarkTime()
1776 ccdEntry['xSize'] = summaryTable['bbox_max_x'] - summaryTable['bbox_min_x']
1777 ccdEntry['ySize'] = summaryTable['bbox_max_y'] - summaryTable['bbox_min_y']
1778 ccdEntry['llcra'] = summaryTable['raCorners'][:, 0]
1779 ccdEntry['llcdec'] = summaryTable['decCorners'][:, 0]
1780 ccdEntry['ulcra'] = summaryTable['raCorners'][:, 1]
1781 ccdEntry['ulcdec'] = summaryTable['decCorners'][:, 1]
1782 ccdEntry['urcra'] = summaryTable['raCorners'][:, 2]
1783 ccdEntry['urcdec'] = summaryTable['decCorners'][:, 2]
1784 ccdEntry['lrcra'] = summaryTable['raCorners'][:, 3]
1785 ccdEntry['lrcdec'] = summaryTable['decCorners'][:, 3]
1786 # TODO: DM-30618, Add raftName, nExposures, ccdTemp, binX, binY, and flags,
1787 # and decide if WCS, and llcx, llcy, ulcx, ulcy, etc. values are actually wanted.
1788 ccdEntries.append(ccdEntry)
1789
1790 outputCatalog = pd.concat(ccdEntries)
1791 outputCatalog.set_index('ccdVisitId', inplace=True, verify_integrity=True)
1792 return pipeBase.Struct(outputCatalog=outputCatalog)
1793
1794
1795class MakeVisitTableConnections(pipeBase.PipelineTaskConnections,
1796 dimensions=("instrument",),
1797 defaultTemplates={"calexpType": ""}):
1798 visitSummaries = connectionTypes.Input(
1799 doc="Per-visit consolidated exposure metadata from ConsolidateVisitSummaryTask",
1800 name="{calexpType}visitSummary",
1801 storageClass="ExposureCatalog",
1802 dimensions=("instrument", "visit",),
1803 multiple=True,
1804 deferLoad=True,
1805 )
1806 outputCatalog = connectionTypes.Output(
1807 doc="Visit metadata table",
1808 name="visitTable",
1809 storageClass="DataFrame",
1810 dimensions=("instrument",)
1811 )
1812
1813
1814class MakeVisitTableConfig(pipeBase.PipelineTaskConfig,
1815 pipelineConnections=MakeVisitTableConnections):
1816 pass
1817
1818
1819class MakeVisitTableTask(CmdLineTask, pipeBase.PipelineTask):
1820 """Produce a `visitTable` from the `visitSummary` exposure catalogs.
1821 """
1822 _DefaultName = 'makeVisitTable'
1823 ConfigClass = MakeVisitTableConfig
1824
1825 def run(self, visitSummaries):
1826 """ Make a table of visit information from the `visitSummary` catalogs
1827
1828 Parameters
1829 ----------
1830 visitSummaries : list of `lsst.afw.table.ExposureCatalog`
1831 List of exposure catalogs with per-detector summary information.
1832 Returns
1833 -------
1834 result : `lsst.pipe.Base.Struct`
1835 Results struct with attribute:
1836 ``outputCatalog``
1837 Catalog of visit information.
1838 """
1839 visitEntries = []
1840 for visitSummary in visitSummaries:
1841 visitSummary = visitSummary.get()
1842 visitRow = visitSummary[0]
1843 visitInfo = visitRow.getVisitInfo()
1844
1845 visitEntry = {}
1846 visitEntry["visitId"] = visitRow['visit']
1847 visitEntry["visit"] = visitRow['visit']
1848 visitEntry["physical_filter"] = visitRow['physical_filter']
1849 visitEntry["band"] = visitRow['band']
1850 raDec = visitInfo.getBoresightRaDec()
1851 visitEntry["ra"] = raDec.getRa().asDegrees()
1852 visitEntry["decl"] = raDec.getDec().asDegrees()
1853 visitEntry["skyRotation"] = visitInfo.getBoresightRotAngle().asDegrees()
1854 azAlt = visitInfo.getBoresightAzAlt()
1855 visitEntry["azimuth"] = azAlt.getLongitude().asDegrees()
1856 visitEntry["altitude"] = azAlt.getLatitude().asDegrees()
1857 visitEntry["zenithDistance"] = 90 - azAlt.getLatitude().asDegrees()
1858 visitEntry["airmass"] = visitInfo.getBoresightAirmass()
1859 expTime = visitInfo.getExposureTime()
1860 visitEntry["expTime"] = expTime
1861 visitEntry["expMidpt"] = visitInfo.getDate().toPython()
1862 visitEntry["expMidptMJD"] = visitInfo.getDate().get(dafBase.DateTime.MJD)
1863 visitEntry["obsStart"] = visitEntry["expMidpt"] - 0.5 * pd.Timedelta(seconds=expTime)
1864 expTime_days = expTime / (60*60*24)
1865 visitEntry["obsStartMJD"] = visitEntry["expMidptMJD"] - 0.5 * expTime_days
1866 visitEntries.append(visitEntry)
1867
1868 # TODO: DM-30623, Add programId, exposureType, cameraTemp, mirror1Temp, mirror2Temp,
1869 # mirror3Temp, domeTemp, externalTemp, dimmSeeing, pwvGPS, pwvMW, flags, nExposures
1870
1871 outputCatalog = pd.DataFrame(data=visitEntries)
1872 outputCatalog.set_index('visitId', inplace=True, verify_integrity=True)
1873 return pipeBase.Struct(outputCatalog=outputCatalog)
1874
1875
1876class WriteForcedSourceTableConnections(pipeBase.PipelineTaskConnections,
1877 dimensions=("instrument", "visit", "detector", "skymap", "tract")):
1878
1879 inputCatalog = connectionTypes.Input(
1880 doc="Primary per-detector, single-epoch forced-photometry catalog. "
1881 "By default, it is the output of ForcedPhotCcdTask on calexps",
1882 name="forced_src",
1883 storageClass="SourceCatalog",
1884 dimensions=("instrument", "visit", "detector", "skymap", "tract")
1885 )
1886 inputCatalogDiff = connectionTypes.Input(
1887 doc="Secondary multi-epoch, per-detector, forced photometry catalog. "
1888 "By default, it is the output of ForcedPhotCcdTask run on image differences.",
1889 name="forced_diff",
1890 storageClass="SourceCatalog",
1891 dimensions=("instrument", "visit", "detector", "skymap", "tract")
1892 )
1893 outputCatalog = connectionTypes.Output(
1894 doc="InputCatalogs horizonatally joined on `objectId` in Parquet format",
1895 name="mergedForcedSource",
1896 storageClass="DataFrame",
1897 dimensions=("instrument", "visit", "detector", "skymap", "tract")
1898 )
1899
1900
1901class WriteForcedSourceTableConfig(pipeBase.PipelineTaskConfig,
1902 pipelineConnections=WriteForcedSourceTableConnections):
1903 key = lsst.pex.config.Field(
1904 doc="Column on which to join the two input tables on and make the primary key of the output",
1905 dtype=str,
1906 default="objectId",
1907 )
1908
1909
1910class WriteForcedSourceTableTask(pipeBase.PipelineTask):
1911 """Merge and convert per-detector forced source catalogs to parquet
1912
1913 Because the predecessor ForcedPhotCcdTask operates per-detector,
1914 per-tract, (i.e., it has tract in its dimensions), detectors
1915 on the tract boundary may have multiple forced source catalogs.
1916
1917 The successor task TransformForcedSourceTable runs per-patch
1918 and temporally-aggregates overlapping mergedForcedSource catalogs from all
1919 available multiple epochs.
1920 """
1921 _DefaultName = "writeForcedSourceTable"
1922 ConfigClass = WriteForcedSourceTableConfig
1923
1924 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1925 inputs = butlerQC.get(inputRefs)
1926 # Add ccdVisitId to allow joining with CcdVisitTable
1927 inputs['ccdVisitId'] = butlerQC.quantum.dataId.pack("visit_detector")
1928 inputs['band'] = butlerQC.quantum.dataId.full['band']
1929 outputs = self.run(**inputs)
1930 butlerQC.put(outputs, outputRefs)
1931
1932 def run(self, inputCatalog, inputCatalogDiff, ccdVisitId=None, band=None):
1933 dfs = []
1934 for table, dataset, in zip((inputCatalog, inputCatalogDiff), ('calexp', 'diff')):
1935 df = table.asAstropy().to_pandas().set_index(self.config.key, drop=False)
1936 df = df.reindex(sorted(df.columns), axis=1)
1937 df['ccdVisitId'] = ccdVisitId if ccdVisitId else pd.NA
1938 df['band'] = band if band else pd.NA
1939 df.columns = pd.MultiIndex.from_tuples([(dataset, c) for c in df.columns],
1940 names=('dataset', 'column'))
1941
1942 dfs.append(df)
1943
1944 outputCatalog = functools.reduce(lambda d1, d2: d1.join(d2), dfs)
1945 return pipeBase.Struct(outputCatalog=outputCatalog)
1946
1947
1948class TransformForcedSourceTableConnections(pipeBase.PipelineTaskConnections,
1949 dimensions=("instrument", "skymap", "patch", "tract")):
1950
1951 inputCatalogs = connectionTypes.Input(
1952 doc="Parquet table of merged ForcedSources produced by WriteForcedSourceTableTask",
1953 name="mergedForcedSource",
1954 storageClass="DataFrame",
1955 dimensions=("instrument", "visit", "detector", "skymap", "tract"),
1956 multiple=True,
1957 deferLoad=True
1958 )
1959 referenceCatalog = connectionTypes.Input(
1960 doc="Reference catalog which was used to seed the forcedPhot. Columns "
1961 "objectId, detect_isPrimary, detect_isTractInner, detect_isPatchInner "
1962 "are expected.",
1963 name="objectTable",
1964 storageClass="DataFrame",
1965 dimensions=("tract", "patch", "skymap"),
1966 deferLoad=True
1967 )
1968 outputCatalog = connectionTypes.Output(
1969 doc="Narrower, temporally-aggregated, per-patch ForcedSource Table transformed and converted per a "
1970 "specified set of functors",
1971 name="forcedSourceTable",
1972 storageClass="DataFrame",
1973 dimensions=("tract", "patch", "skymap")
1974 )
1975
1976
1977class TransformForcedSourceTableConfig(TransformCatalogBaseConfig,
1978 pipelineConnections=TransformForcedSourceTableConnections):
1979 referenceColumns = pexConfig.ListField(
1980 dtype=str,
1981 default=["detect_isPrimary", "detect_isTractInner", "detect_isPatchInner"],
1982 optional=True,
1983 doc="Columns to pull from reference catalog",
1984 )
1985 keyRef = lsst.pex.config.Field(
1986 doc="Column on which to join the two input tables on and make the primary key of the output",
1987 dtype=str,
1988 default="objectId",
1989 )
1990 key = lsst.pex.config.Field(
1991 doc="Rename the output DataFrame index to this name",
1992 dtype=str,
1993 default="forcedSourceId",
1994 )
1995
1996 def setDefaults(self):
1997 super().setDefaults()
1998 self.functorFile = os.path.join('$PIPE_TASKS_DIR', 'schemas', 'ForcedSource.yaml')
1999
2000
2001class TransformForcedSourceTableTask(TransformCatalogBaseTask):
2002 """Transform/standardize a ForcedSource catalog
2003
2004 Transforms each wide, per-detector forcedSource parquet table per the
2005 specification file (per-camera defaults found in ForcedSource.yaml).
2006 All epochs that overlap the patch are aggregated into one per-patch
2007 narrow-parquet file.
2008
2009 No de-duplication of rows is performed. Duplicate resolutions flags are
2010 pulled in from the referenceCatalog: `detect_isPrimary`,
2011 `detect_isTractInner`,`detect_isPatchInner`, so that user may de-duplicate
2012 for analysis or compare duplicates for QA.
2013
2014 The resulting table includes multiple bands. Epochs (MJDs) and other useful
2015 per-visit rows can be retreived by joining with the CcdVisitTable on
2016 ccdVisitId.
2017 """
2018 _DefaultName = "transformForcedSourceTable"
2019 ConfigClass = TransformForcedSourceTableConfig
2020
2021 def runQuantum(self, butlerQC, inputRefs, outputRefs):
2022 inputs = butlerQC.get(inputRefs)
2023 if self.funcs is None:
2024 raise ValueError("config.functorFile is None. "
2025 "Must be a valid path to yaml in order to run Task as a PipelineTask.")
2026 outputs = self.run(inputs['inputCatalogs'], inputs['referenceCatalog'], funcs=self.funcs,
2027 dataId=outputRefs.outputCatalog.dataId.full)
2028
2029 butlerQC.put(outputs, outputRefs)
2030
2031 def run(self, inputCatalogs, referenceCatalog, funcs=None, dataId=None, band=None):
2032 dfs = []
2033 ref = referenceCatalog.get(parameters={"columns": self.config.referenceColumns})
2034 self.log.info("Aggregating %s input catalogs" % (len(inputCatalogs)))
2035 for handle in inputCatalogs:
2036 result = self.transform(None, handle, funcs, dataId)
2037 # Filter for only rows that were detected on (overlap) the patch
2038 dfs.append(result.df.join(ref, how='inner'))
2039
2040 outputCatalog = pd.concat(dfs)
2041
2042 # Now that we are done joining on config.keyRef
2043 # Change index to config.key by
2044 outputCatalog.index.rename(self.config.keyRef, inplace=True)
2045 # Add config.keyRef to the column list
2046 outputCatalog.reset_index(inplace=True)
2047 # set the forcedSourceId to the index. This is specified in the ForcedSource.yaml
2048 outputCatalog.set_index("forcedSourceId", inplace=True, verify_integrity=True)
2049 # Rename it to the config.key
2050 outputCatalog.index.rename(self.config.key, inplace=True)
2051
2052 self.log.info("Made a table of %d columns and %d rows",
2053 len(outputCatalog.columns), len(outputCatalog))
2054 return pipeBase.Struct(outputCatalog=outputCatalog)
2055
2056
2057class ConsolidateTractConnections(pipeBase.PipelineTaskConnections,
2058 defaultTemplates={"catalogType": ""},
2059 dimensions=("instrument", "tract")):
2060 inputCatalogs = connectionTypes.Input(
2061 doc="Input per-patch DataFrame Tables to be concatenated",
2062 name="{catalogType}ForcedSourceTable",
2063 storageClass="DataFrame",
2064 dimensions=("tract", "patch", "skymap"),
2065 multiple=True,
2066 )
2067
2068 outputCatalog = connectionTypes.Output(
2069 doc="Output per-tract concatenation of DataFrame Tables",
2070 name="{catalogType}ForcedSourceTable_tract",
2071 storageClass="DataFrame",
2072 dimensions=("tract", "skymap"),
2073 )
2074
2075
2076class ConsolidateTractConfig(pipeBase.PipelineTaskConfig,
2077 pipelineConnections=ConsolidateTractConnections):
2078 pass
2079
2080
2081class ConsolidateTractTask(CmdLineTask, pipeBase.PipelineTask):
2082 """Concatenate any per-patch, dataframe list into a single
2083 per-tract DataFrame
2084 """
2085 _DefaultName = 'ConsolidateTract'
2086 ConfigClass = ConsolidateTractConfig
2087
2088 def runQuantum(self, butlerQC, inputRefs, outputRefs):
2089 inputs = butlerQC.get(inputRefs)
2090 # Not checking at least one inputCatalog exists because that'd be an empty QG
2091 self.log.info("Concatenating %s per-patch %s Tables",
2092 len(inputs['inputCatalogs']),
2093 inputRefs.inputCatalogs[0].datasetType.name)
2094 df = pd.concat(inputs['inputCatalogs'])
2095 butlerQC.put(pipeBase.Struct(outputCatalog=df), outputRefs)
def getAnalysis(self, parq, funcs=None, band=None)
Definition: postprocess.py:987
def transform(self, band, parq, funcs, dataId)
Definition: postprocess.py:993
def run(self, parq, funcs=None, dataId=None, band=None)
Definition: postprocess.py:956
def runQuantum(self, butlerQC, inputRefs, outputRefs)
Definition: postprocess.py:937
def writeMetadata(self, dataRefList)
No metadata to write, and not sure how to write it for a list of dataRefs.
def makeMergeArgumentParser(name, dataset)
Create a suitable ArgumentParser.
def readCatalog(task, patchRef)
Read input catalog.
def flattenFilters(df, noDupCols=['coord_ra', 'coord_dec'], camelCase=False, inputBands=None)
Definition: postprocess.py:50