lsst.pipe.tasks gafc8b9801c+4217b6d8cf
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
24import logging
25import numpy as np
26import numbers
27import os
28
29import lsst.geom
30import lsst.pex.config as pexConfig
31import lsst.pipe.base as pipeBase
32import lsst.daf.base as dafBase
33from lsst.obs.base import ExposureIdInfo
34from lsst.pipe.base import connectionTypes
35import lsst.afw.table as afwTable
36from lsst.meas.base import SingleFrameMeasurementTask
37from lsst.daf.butler import DeferredDatasetHandle, DataCoordinate
38from lsst.skymap import BaseSkyMap
39
40from .parquetTable import ParquetTable
41from .functors import CompositeFunctor, Column
42
43log = logging.getLogger(__name__)
44
45
46def flattenFilters(df, noDupCols=['coord_ra', 'coord_dec'], camelCase=False, inputBands=None):
47 """Flattens a dataframe with multilevel column index.
48 """
49 newDf = pd.DataFrame()
50 # band is the level 0 index
51 dfBands = df.columns.unique(level=0).values
52 for band in dfBands:
53 subdf = df[band]
54 columnFormat = '{0}{1}' if camelCase else '{0}_{1}'
55 newColumns = {c: columnFormat.format(band, c)
56 for c in subdf.columns if c not in noDupCols}
57 cols = list(newColumns.keys())
58 newDf = pd.concat([newDf, subdf[cols].rename(columns=newColumns)], axis=1)
59
60 # Band must be present in the input and output or else column is all NaN:
61 presentBands = dfBands if inputBands is None else list(set(inputBands).intersection(dfBands))
62 # Get the unexploded columns from any present band's partition
63 noDupDf = df[presentBands[0]][noDupCols]
64 newDf = pd.concat([noDupDf, newDf], axis=1)
65 return newDf
66
67
68class WriteObjectTableConnections(pipeBase.PipelineTaskConnections,
69 defaultTemplates={"coaddName": "deep"},
70 dimensions=("tract", "patch", "skymap")):
71 inputCatalogMeas = connectionTypes.Input(
72 doc="Catalog of source measurements on the deepCoadd.",
73 dimensions=("tract", "patch", "band", "skymap"),
74 storageClass="SourceCatalog",
75 name="{coaddName}Coadd_meas",
76 multiple=True
77 )
78 inputCatalogForcedSrc = connectionTypes.Input(
79 doc="Catalog of forced measurements (shape and position parameters held fixed) on the deepCoadd.",
80 dimensions=("tract", "patch", "band", "skymap"),
81 storageClass="SourceCatalog",
82 name="{coaddName}Coadd_forced_src",
83 multiple=True
84 )
85 inputCatalogRef = connectionTypes.Input(
86 doc="Catalog marking the primary detection (which band provides a good shape and position)"
87 "for each detection in deepCoadd_mergeDet.",
88 dimensions=("tract", "patch", "skymap"),
89 storageClass="SourceCatalog",
90 name="{coaddName}Coadd_ref"
91 )
92 outputCatalog = connectionTypes.Output(
93 doc="A vertical concatenation of the deepCoadd_{ref|meas|forced_src} catalogs, "
94 "stored as a DataFrame with a multi-level column index per-patch.",
95 dimensions=("tract", "patch", "skymap"),
96 storageClass="DataFrame",
97 name="{coaddName}Coadd_obj"
98 )
99
100
101class WriteObjectTableConfig(pipeBase.PipelineTaskConfig,
102 pipelineConnections=WriteObjectTableConnections):
103 engine = pexConfig.Field(
104 dtype=str,
105 default="pyarrow",
106 doc="Parquet engine for writing (pyarrow or fastparquet)"
107 )
108 coaddName = pexConfig.Field(
109 dtype=str,
110 default="deep",
111 doc="Name of coadd"
112 )
113
114
115class WriteObjectTableTask(pipeBase.PipelineTask):
116 """Write filter-merged source tables to parquet
117 """
118 _DefaultName = "writeObjectTable"
119 ConfigClass = WriteObjectTableConfig
120
121 # Names of table datasets to be merged
122 inputDatasets = ('forced_src', 'meas', 'ref')
123
124 # Tag of output dataset written by `MergeSourcesTask.write`
125 outputDataset = 'obj'
126
127 def runQuantum(self, butlerQC, inputRefs, outputRefs):
128 inputs = butlerQC.get(inputRefs)
129
130 measDict = {ref.dataId['band']: {'meas': cat} for ref, cat in
131 zip(inputRefs.inputCatalogMeas, inputs['inputCatalogMeas'])}
132 forcedSourceDict = {ref.dataId['band']: {'forced_src': cat} for ref, cat in
133 zip(inputRefs.inputCatalogForcedSrc, inputs['inputCatalogForcedSrc'])}
134
135 catalogs = {}
136 for band in measDict.keys():
137 catalogs[band] = {'meas': measDict[band]['meas'],
138 'forced_src': forcedSourceDict[band]['forced_src'],
139 'ref': inputs['inputCatalogRef']}
140 dataId = butlerQC.quantum.dataId
141 df = self.run(catalogs=catalogs, tract=dataId['tract'], patch=dataId['patch'])
142 outputs = pipeBase.Struct(outputCatalog=df)
143 butlerQC.put(outputs, outputRefs)
144
145 def run(self, catalogs, tract, patch):
146 """Merge multiple catalogs.
147
148 Parameters
149 ----------
150 catalogs : `dict`
151 Mapping from filter names to dict of catalogs.
152 tract : int
153 tractId to use for the tractId column.
154 patch : str
155 patchId to use for the patchId column.
156
157 Returns
158 -------
159 catalog : `pandas.DataFrame`
160 Merged dataframe.
161 """
162
163 dfs = []
164 for filt, tableDict in catalogs.items():
165 for dataset, table in tableDict.items():
166 # Convert afwTable to pandas DataFrame
167 df = table.asAstropy().to_pandas().set_index('id', drop=True)
168
169 # Sort columns by name, to ensure matching schema among patches
170 df = df.reindex(sorted(df.columns), axis=1)
171 df['tractId'] = tract
172 df['patchId'] = patch
173
174 # Make columns a 3-level MultiIndex
175 df.columns = pd.MultiIndex.from_tuples([(dataset, filt, c) for c in df.columns],
176 names=('dataset', 'band', 'column'))
177 dfs.append(df)
178
179 catalog = functools.reduce(lambda d1, d2: d1.join(d2), dfs)
180 return catalog
181
182
183class WriteSourceTableConnections(pipeBase.PipelineTaskConnections,
184 defaultTemplates={"catalogType": ""},
185 dimensions=("instrument", "visit", "detector")):
186
187 catalog = connectionTypes.Input(
188 doc="Input full-depth catalog of sources produced by CalibrateTask",
189 name="{catalogType}src",
190 storageClass="SourceCatalog",
191 dimensions=("instrument", "visit", "detector")
192 )
193 outputCatalog = connectionTypes.Output(
194 doc="Catalog of sources, `src` in Parquet format. The 'id' column is "
195 "replaced with an index; all other columns are unchanged.",
196 name="{catalogType}source",
197 storageClass="DataFrame",
198 dimensions=("instrument", "visit", "detector")
199 )
200
201
202class WriteSourceTableConfig(pipeBase.PipelineTaskConfig,
203 pipelineConnections=WriteSourceTableConnections):
204 pass
205
206
207class WriteSourceTableTask(pipeBase.PipelineTask):
208 """Write source table to parquet.
209 """
210 _DefaultName = "writeSourceTable"
211 ConfigClass = WriteSourceTableConfig
212
213 def runQuantum(self, butlerQC, inputRefs, outputRefs):
214 inputs = butlerQC.get(inputRefs)
215 inputs['ccdVisitId'] = butlerQC.quantum.dataId.pack("visit_detector")
216 result = self.run(**inputs).table
217 outputs = pipeBase.Struct(outputCatalog=result.toDataFrame())
218 butlerQC.put(outputs, outputRefs)
219
220 def run(self, catalog, ccdVisitId=None, **kwargs):
221 """Convert `src` catalog to parquet
222
223 Parameters
224 ----------
225 catalog: `afwTable.SourceCatalog`
226 catalog to be converted
227 ccdVisitId: `int`
228 ccdVisitId to be added as a column
229
230 Returns
231 -------
232 result : `lsst.pipe.base.Struct`
233 ``table``
234 `ParquetTable` version of the input catalog
235 """
236 self.log.info("Generating parquet table from src catalog ccdVisitId=%s", ccdVisitId)
237 df = catalog.asAstropy().to_pandas().set_index('id', drop=True)
238 df['ccdVisitId'] = ccdVisitId
239 return pipeBase.Struct(table=ParquetTable(dataFrame=df))
240
241
242class WriteRecalibratedSourceTableConnections(WriteSourceTableConnections,
243 defaultTemplates={"catalogType": "",
244 "skyWcsName": "jointcal",
245 "photoCalibName": "fgcm"},
246 dimensions=("instrument", "visit", "detector", "skymap")):
247 skyMap = connectionTypes.Input(
248 doc="skyMap needed to choose which tract-level calibrations to use when multiple available",
249 name=BaseSkyMap.SKYMAP_DATASET_TYPE_NAME,
250 storageClass="SkyMap",
251 dimensions=("skymap",),
252 )
253 exposure = connectionTypes.Input(
254 doc="Input exposure to perform photometry on.",
255 name="calexp",
256 storageClass="ExposureF",
257 dimensions=["instrument", "visit", "detector"],
258 )
259 externalSkyWcsTractCatalog = connectionTypes.Input(
260 doc=("Per-tract, per-visit wcs calibrations. These catalogs use the detector "
261 "id for the catalog id, sorted on id for fast lookup."),
262 name="{skyWcsName}SkyWcsCatalog",
263 storageClass="ExposureCatalog",
264 dimensions=["instrument", "visit", "tract"],
265 multiple=True
266 )
267 externalSkyWcsGlobalCatalog = connectionTypes.Input(
268 doc=("Per-visit wcs calibrations computed globally (with no tract information). "
269 "These catalogs use the detector id for the catalog id, sorted on id for "
270 "fast lookup."),
271 name="{skyWcsName}SkyWcsCatalog",
272 storageClass="ExposureCatalog",
273 dimensions=["instrument", "visit"],
274 )
275 externalPhotoCalibTractCatalog = connectionTypes.Input(
276 doc=("Per-tract, per-visit photometric calibrations. These catalogs use the "
277 "detector id for the catalog id, sorted on id for fast lookup."),
278 name="{photoCalibName}PhotoCalibCatalog",
279 storageClass="ExposureCatalog",
280 dimensions=["instrument", "visit", "tract"],
281 multiple=True
282 )
283 externalPhotoCalibGlobalCatalog = connectionTypes.Input(
284 doc=("Per-visit photometric calibrations computed globally (with no tract "
285 "information). These catalogs use the detector id for the catalog id, "
286 "sorted on id for fast lookup."),
287 name="{photoCalibName}PhotoCalibCatalog",
288 storageClass="ExposureCatalog",
289 dimensions=["instrument", "visit"],
290 )
291
292 def __init__(self, *, config=None):
293 super().__init__(config=config)
294 # Same connection boilerplate as all other applications of
295 # Global/Tract calibrations
296 if config.doApplyExternalSkyWcs and config.doReevaluateSkyWcs:
297 if config.useGlobalExternalSkyWcs:
298 self.inputs.remove("externalSkyWcsTractCatalog")
299 else:
300 self.inputs.remove("externalSkyWcsGlobalCatalog")
301 else:
302 self.inputs.remove("externalSkyWcsTractCatalog")
303 self.inputs.remove("externalSkyWcsGlobalCatalog")
304 if config.doApplyExternalPhotoCalib and config.doReevaluatePhotoCalib:
305 if config.useGlobalExternalPhotoCalib:
306 self.inputs.remove("externalPhotoCalibTractCatalog")
307 else:
308 self.inputs.remove("externalPhotoCalibGlobalCatalog")
309 else:
310 self.inputs.remove("externalPhotoCalibTractCatalog")
311 self.inputs.remove("externalPhotoCalibGlobalCatalog")
312
313
314class WriteRecalibratedSourceTableConfig(WriteSourceTableConfig,
315 pipelineConnections=WriteRecalibratedSourceTableConnections):
316
317 doReevaluatePhotoCalib = pexConfig.Field(
318 dtype=bool,
319 default=True,
320 doc=("Add or replace local photoCalib columns")
321 )
322 doReevaluateSkyWcs = pexConfig.Field(
323 dtype=bool,
324 default=True,
325 doc=("Add or replace local WCS columns and update the coord columns, coord_ra and coord_dec")
326 )
327 doApplyExternalPhotoCalib = pexConfig.Field(
328 dtype=bool,
329 default=True,
330 doc=("If and only if doReevaluatePhotoCalib, apply the photometric calibrations from an external ",
331 "algorithm such as FGCM or jointcal, else use the photoCalib already attached to the exposure."),
332 )
333 doApplyExternalSkyWcs = pexConfig.Field(
334 dtype=bool,
335 default=True,
336 doc=("if and only if doReevaluateSkyWcs, apply the WCS from an external algorithm such as jointcal, ",
337 "else use the wcs already attached to the exposure."),
338 )
339 useGlobalExternalPhotoCalib = pexConfig.Field(
340 dtype=bool,
341 default=True,
342 doc=("When using doApplyExternalPhotoCalib, use 'global' calibrations "
343 "that are not run per-tract. When False, use per-tract photometric "
344 "calibration files.")
345 )
346 useGlobalExternalSkyWcs = pexConfig.Field(
347 dtype=bool,
348 default=False,
349 doc=("When using doApplyExternalSkyWcs, use 'global' calibrations "
350 "that are not run per-tract. When False, use per-tract wcs "
351 "files.")
352 )
353
354 def validate(self):
355 super().validate()
356 if self.doApplyExternalSkyWcs and not self.doReevaluateSkyWcs:
357 log.warning("doApplyExternalSkyWcs=True but doReevaluateSkyWcs=False"
358 "External SkyWcs will not be read or evaluated.")
359 if self.doApplyExternalPhotoCalib and not self.doReevaluatePhotoCalib:
360 log.warning("doApplyExternalPhotoCalib=True but doReevaluatePhotoCalib=False."
361 "External PhotoCalib will not be read or evaluated.")
362
363
364class WriteRecalibratedSourceTableTask(WriteSourceTableTask):
365 """Write source table to parquet
366 """
367 _DefaultName = "writeRecalibratedSourceTable"
368 ConfigClass = WriteRecalibratedSourceTableConfig
369
370 def runQuantum(self, butlerQC, inputRefs, outputRefs):
371 inputs = butlerQC.get(inputRefs)
372 inputs['ccdVisitId'] = butlerQC.quantum.dataId.pack("visit_detector")
373 inputs['exposureIdInfo'] = ExposureIdInfo.fromDataId(butlerQC.quantum.dataId, "visit_detector")
374
375 if self.config.doReevaluatePhotoCalib or self.config.doReevaluateSkyWcs:
376 if self.config.doApplyExternalPhotoCalib or self.config.doApplyExternalSkyWcs:
377 inputs['exposure'] = self.attachCalibs(inputRefs, **inputs)
378
379 inputs['catalog'] = self.addCalibColumns(**inputs)
380
381 result = self.run(**inputs).table
382 outputs = pipeBase.Struct(outputCatalog=result.toDataFrame())
383 butlerQC.put(outputs, outputRefs)
384
385 def attachCalibs(self, inputRefs, skyMap, exposure, externalSkyWcsGlobalCatalog=None,
386 externalSkyWcsTractCatalog=None, externalPhotoCalibGlobalCatalog=None,
387 externalPhotoCalibTractCatalog=None, **kwargs):
388 """Apply external calibrations to exposure per configuration
389
390 When multiple tract-level calibrations overlap, select the one with the
391 center closest to detector.
392
393 Parameters
394 ----------
395 inputRefs : `lsst.pipe.base.InputQuantizedConnection`, for dataIds of
396 tract-level calibs.
397 skyMap : `lsst.skymap.SkyMap`
398 exposure : `lsst.afw.image.exposure.Exposure`
399 Input exposure to adjust calibrations.
400 externalSkyWcsGlobalCatalog : `lsst.afw.table.ExposureCatalog`, optional
401 Exposure catalog with external skyWcs to be applied per config
402 externalSkyWcsTractCatalog : `lsst.afw.table.ExposureCatalog`, optional
403 Exposure catalog with external skyWcs to be applied per config
404 externalPhotoCalibGlobalCatalog : `lsst.afw.table.ExposureCatalog`, optional
405 Exposure catalog with external photoCalib to be applied per config
406 externalPhotoCalibTractCatalog : `lsst.afw.table.ExposureCatalog`, optional
407
408
409 Returns
410 -------
411 exposure : `lsst.afw.image.exposure.Exposure`
412 Exposure with adjusted calibrations.
413 """
414 if not self.config.doApplyExternalSkyWcs:
415 # Do not modify the exposure's SkyWcs
416 externalSkyWcsCatalog = None
417 elif self.config.useGlobalExternalSkyWcs:
418 # Use the global external SkyWcs
419 externalSkyWcsCatalog = externalSkyWcsGlobalCatalog
420 self.log.info('Applying global SkyWcs')
421 else:
422 # use tract-level external SkyWcs from the closest overlapping tract
423 inputRef = getattr(inputRefs, 'externalSkyWcsTractCatalog')
424 tracts = [ref.dataId['tract'] for ref in inputRef]
425 if len(tracts) == 1:
426 ind = 0
427 self.log.info('Applying tract-level SkyWcs from tract %s', tracts[ind])
428 else:
429 ind = self.getClosestTract(tracts, skyMap,
430 exposure.getBBox(), exposure.getWcs())
431 self.log.info('Multiple overlapping externalSkyWcsTractCatalogs found (%s). '
432 'Applying closest to detector center: tract=%s', str(tracts), tracts[ind])
433
434 externalSkyWcsCatalog = externalSkyWcsTractCatalog[ind]
435
436 if not self.config.doApplyExternalPhotoCalib:
437 # Do not modify the exposure's PhotoCalib
438 externalPhotoCalibCatalog = None
439 elif self.config.useGlobalExternalPhotoCalib:
440 # Use the global external PhotoCalib
441 externalPhotoCalibCatalog = externalPhotoCalibGlobalCatalog
442 self.log.info('Applying global PhotoCalib')
443 else:
444 # use tract-level external PhotoCalib from the closest overlapping tract
445 inputRef = getattr(inputRefs, 'externalPhotoCalibTractCatalog')
446 tracts = [ref.dataId['tract'] for ref in inputRef]
447 if len(tracts) == 1:
448 ind = 0
449 self.log.info('Applying tract-level PhotoCalib from tract %s', tracts[ind])
450 else:
451 ind = self.getClosestTract(tracts, skyMap,
452 exposure.getBBox(), exposure.getWcs())
453 self.log.info('Multiple overlapping externalPhotoCalibTractCatalogs found (%s). '
454 'Applying closest to detector center: tract=%s', str(tracts), tracts[ind])
455
456 externalPhotoCalibCatalog = externalPhotoCalibTractCatalog[ind]
457
458 return self.prepareCalibratedExposure(exposure, externalSkyWcsCatalog, externalPhotoCalibCatalog)
459
460 def getClosestTract(self, tracts, skyMap, bbox, wcs):
461 """Find the index of the tract closest to detector from list of tractIds
462
463 Parameters
464 ----------
465 tracts: `list` [`int`]
466 Iterable of integer tractIds
467 skyMap : `lsst.skymap.SkyMap`
468 skyMap to lookup tract geometry and wcs
469 bbox : `lsst.geom.Box2I`
470 Detector bbox, center of which will compared to tract centers
472 Detector Wcs object to map the detector center to SkyCoord
473
474 Returns
475 -------
476 index : `int`
477 """
478 if len(tracts) == 1:
479 return 0
480
481 center = wcs.pixelToSky(bbox.getCenter())
482 sep = []
483 for tractId in tracts:
484 tract = skyMap[tractId]
485 tractCenter = tract.getWcs().pixelToSky(tract.getBBox().getCenter())
486 sep.append(center.separation(tractCenter))
487
488 return np.argmin(sep)
489
490 def prepareCalibratedExposure(self, exposure, externalSkyWcsCatalog=None, externalPhotoCalibCatalog=None):
491 """Prepare a calibrated exposure and apply external calibrations
492 if so configured.
493
494 Parameters
495 ----------
496 exposure : `lsst.afw.image.exposure.Exposure`
497 Input exposure to adjust calibrations.
498 externalSkyWcsCatalog : `lsst.afw.table.ExposureCatalog`, optional
499 Exposure catalog with external skyWcs to be applied
500 if config.doApplyExternalSkyWcs=True. Catalog uses the detector id
501 for the catalog id, sorted on id for fast lookup.
502 externalPhotoCalibCatalog : `lsst.afw.table.ExposureCatalog`, optional
503 Exposure catalog with external photoCalib to be applied
504 if config.doApplyExternalPhotoCalib=True. Catalog uses the detector
505 id for the catalog id, sorted on id for fast lookup.
506
507 Returns
508 -------
509 exposure : `lsst.afw.image.exposure.Exposure`
510 Exposure with adjusted calibrations.
511 """
512 detectorId = exposure.getInfo().getDetector().getId()
513
514 if externalPhotoCalibCatalog is not None:
515 row = externalPhotoCalibCatalog.find(detectorId)
516 if row is None:
517 self.log.warning("Detector id %s not found in externalPhotoCalibCatalog; "
518 "Using original photoCalib.", detectorId)
519 else:
520 photoCalib = row.getPhotoCalib()
521 if photoCalib is None:
522 self.log.warning("Detector id %s has None for photoCalib in externalPhotoCalibCatalog; "
523 "Using original photoCalib.", detectorId)
524 else:
525 exposure.setPhotoCalib(photoCalib)
526
527 if externalSkyWcsCatalog is not None:
528 row = externalSkyWcsCatalog.find(detectorId)
529 if row is None:
530 self.log.warning("Detector id %s not found in externalSkyWcsCatalog; "
531 "Using original skyWcs.", detectorId)
532 else:
533 skyWcs = row.getWcs()
534 if skyWcs is None:
535 self.log.warning("Detector id %s has None for skyWcs in externalSkyWcsCatalog; "
536 "Using original skyWcs.", detectorId)
537 else:
538 exposure.setWcs(skyWcs)
539
540 return exposure
541
542 def addCalibColumns(self, catalog, exposure, exposureIdInfo, **kwargs):
543 """Add replace columns with calibs evaluated at each centroid
544
545 Add or replace 'base_LocalWcs' `base_LocalPhotoCalib' columns in a
546 a source catalog, by rerunning the plugins.
547
548 Parameters
549 ----------
551 catalog to which calib columns will be added
552 exposure : `lsst.afw.image.exposure.Exposure`
553 Exposure with attached PhotoCalibs and SkyWcs attributes to be
554 reevaluated at local centroids. Pixels are not required.
555 exposureIdInfo : `lsst.obs.base.ExposureIdInfo`
556
557 Returns
558 -------
560 Source Catalog with requested local calib columns
561 """
562 measureConfig = SingleFrameMeasurementTask.ConfigClass()
563 measureConfig.doReplaceWithNoise = False
564
565 # Clear all slots, because we aren't running the relevant plugins.
566 for slot in measureConfig.slots:
567 setattr(measureConfig.slots, slot, None)
568
569 measureConfig.plugins.names = []
570 if self.config.doReevaluateSkyWcs:
571 measureConfig.plugins.names.add('base_LocalWcs')
572 self.log.info("Re-evaluating base_LocalWcs plugin")
573 if self.config.doReevaluatePhotoCalib:
574 measureConfig.plugins.names.add('base_LocalPhotoCalib')
575 self.log.info("Re-evaluating base_LocalPhotoCalib plugin")
576 pluginsNotToCopy = tuple(measureConfig.plugins.names)
577
578 # Create a new schema and catalog
579 # Copy all columns from original except for the ones to reevaluate
580 aliasMap = catalog.schema.getAliasMap()
581 mapper = afwTable.SchemaMapper(catalog.schema)
582 for item in catalog.schema:
583 if not item.field.getName().startswith(pluginsNotToCopy):
584 mapper.addMapping(item.key)
585
586 schema = mapper.getOutputSchema()
587 measurement = SingleFrameMeasurementTask(config=measureConfig, schema=schema)
588 schema.setAliasMap(aliasMap)
589 newCat = afwTable.SourceCatalog(schema)
590 newCat.extend(catalog, mapper=mapper)
591
592 # Fluxes in sourceCatalogs are in counts, so there are no fluxes to
593 # update here. LocalPhotoCalibs are applied during transform tasks.
594 # Update coord_ra/coord_dec, which are expected to be positions on the
595 # sky and are used as such in sdm tables without transform
596 if self.config.doReevaluateSkyWcs:
597 afwTable.updateSourceCoords(exposure.wcs, newCat)
598
599 measurement.run(measCat=newCat, exposure=exposure, exposureId=exposureIdInfo.expId)
600
601 return newCat
602
603
605 """Calculate columns from ParquetTable.
606
607 This object manages and organizes an arbitrary set of computations
608 on a catalog. The catalog is defined by a
609 `lsst.pipe.tasks.parquetTable.ParquetTable` object (or list thereof), such
610 as a `deepCoadd_obj` dataset, and the computations are defined by a
611 collection of `lsst.pipe.tasks.functor.Functor` objects (or, equivalently,
612 a `CompositeFunctor`).
613
614 After the object is initialized, accessing the `.df` attribute (which
615 holds the `pandas.DataFrame` containing the results of the calculations)
616 triggers computation of said dataframe.
617
618 One of the conveniences of using this object is the ability to define a
619 desired common filter for all functors. This enables the same functor
620 collection to be passed to several different `PostprocessAnalysis` objects
621 without having to change the original functor collection, since the `filt`
622 keyword argument of this object triggers an overwrite of the `filt`
623 property for all functors in the collection.
624
625 This object also allows a list of refFlags to be passed, and defines a set
626 of default refFlags that are always included even if not requested.
627
628 If a list of `ParquetTable` object is passed, rather than a single one,
629 then the calculations will be mapped over all the input catalogs. In
630 principle, it should be straightforward to parallelize this activity, but
631 initial tests have failed (see TODO in code comments).
632
633 Parameters
634 ----------
635 parq : `lsst.pipe.tasks.ParquetTable` (or list of such)
636 Source catalog(s) for computation.
637
638 functors : `list`, `dict`, or `lsst.pipe.tasks.functors.CompositeFunctor`
639 Computations to do (functors that act on `parq`).
640 If a dict, the output
641 DataFrame will have columns keyed accordingly.
642 If a list, the column keys will come from the
643 `.shortname` attribute of each functor.
644
645 filt : `str`, optional
646 Filter in which to calculate. If provided,
647 this will overwrite any existing `.filt` attribute
648 of the provided functors.
649
650 flags : `list`, optional
651 List of flags (per-band) to include in output table.
652 Taken from the `meas` dataset if applied to a multilevel Object Table.
653
654 refFlags : `list`, optional
655 List of refFlags (only reference band) to include in output table.
656
657 forcedFlags : `list`, optional
658 List of flags (per-band) to include in output table.
659 Taken from the ``forced_src`` dataset if applied to a
660 multilevel Object Table. Intended for flags from measurement plugins
661 only run during multi-band forced-photometry.
662 """
663 _defaultRefFlags = []
664 _defaultFuncs = ()
665
666 def __init__(self, parq, functors, filt=None, flags=None, refFlags=None, forcedFlags=None):
667 self.parq = parq
668 self.functors = functors
669
670 self.filt = filt
671 self.flags = list(flags) if flags is not None else []
672 self.forcedFlags = list(forcedFlags) if forcedFlags is not None else []
673 self.refFlags = list(self._defaultRefFlags)
674 if refFlags is not None:
675 self.refFlags += list(refFlags)
676
677 self._df = None
678
679 @property
680 def defaultFuncs(self):
681 funcs = dict(self._defaultFuncs)
682 return funcs
683
684 @property
685 def func(self):
686 additionalFuncs = self.defaultFuncs
687 additionalFuncs.update({flag: Column(flag, dataset='forced_src') for flag in self.forcedFlags})
688 additionalFuncs.update({flag: Column(flag, dataset='ref') for flag in self.refFlags})
689 additionalFuncs.update({flag: Column(flag, dataset='meas') for flag in self.flags})
690
691 if isinstance(self.functors, CompositeFunctor):
692 func = self.functors
693 else:
694 func = CompositeFunctor(self.functors)
695
696 func.funcDict.update(additionalFuncs)
697 func.filt = self.filt
698
699 return func
700
701 @property
702 def noDupCols(self):
703 return [name for name, func in self.func.funcDict.items() if func.noDup or func.dataset == 'ref']
704
705 @property
706 def df(self):
707 if self._df is None:
708 self.compute()
709 return self._df
710
711 def compute(self, dropna=False, pool=None):
712 # map over multiple parquet tables
713 if type(self.parq) in (list, tuple):
714 if pool is None:
715 dflist = [self.func(parq, dropna=dropna) for parq in self.parq]
716 else:
717 # TODO: Figure out why this doesn't work (pyarrow pickling
718 # issues?)
719 dflist = pool.map(functools.partial(self.func, dropna=dropna), self.parq)
720 self._df = pd.concat(dflist)
721 else:
722 self._df = self.func(self.parq, dropna=dropna)
723
724 return self._df
725
726
727class TransformCatalogBaseConnections(pipeBase.PipelineTaskConnections,
728 dimensions=()):
729 """Expected Connections for subclasses of TransformCatalogBaseTask.
730
731 Must be subclassed.
732 """
733 inputCatalog = connectionTypes.Input(
734 name="",
735 storageClass="DataFrame",
736 )
737 outputCatalog = connectionTypes.Output(
738 name="",
739 storageClass="DataFrame",
740 )
741
742
743class TransformCatalogBaseConfig(pipeBase.PipelineTaskConfig,
744 pipelineConnections=TransformCatalogBaseConnections):
745 functorFile = pexConfig.Field(
746 dtype=str,
747 doc="Path to YAML file specifying Science Data Model functors to use "
748 "when copying columns and computing calibrated values.",
749 default=None,
750 optional=True
751 )
752 primaryKey = pexConfig.Field(
753 dtype=str,
754 doc="Name of column to be set as the DataFrame index. If None, the index"
755 "will be named `id`",
756 default=None,
757 optional=True
758 )
759 columnsFromDataId = pexConfig.ListField(
760 dtype=str,
761 default=None,
762 optional=True,
763 doc="Columns to extract from the dataId",
764 )
765
766
767class TransformCatalogBaseTask(pipeBase.PipelineTask):
768 """Base class for transforming/standardizing a catalog
769
770 by applying functors that convert units and apply calibrations.
771 The purpose of this task is to perform a set of computations on
772 an input `ParquetTable` dataset (such as `deepCoadd_obj`) and write the
773 results to a new dataset (which needs to be declared in an `outputDataset`
774 attribute).
775
776 The calculations to be performed are defined in a YAML file that specifies
777 a set of functors to be computed, provided as
778 a `--functorFile` config parameter. An example of such a YAML file
779 is the following:
780
781 funcs:
782 psfMag:
783 functor: Mag
784 args:
785 - base_PsfFlux
786 filt: HSC-G
787 dataset: meas
788 cmodel_magDiff:
789 functor: MagDiff
790 args:
791 - modelfit_CModel
792 - base_PsfFlux
793 filt: HSC-G
794 gauss_magDiff:
795 functor: MagDiff
796 args:
797 - base_GaussianFlux
798 - base_PsfFlux
799 filt: HSC-G
800 count:
801 functor: Column
802 args:
803 - base_InputCount_value
804 filt: HSC-G
805 deconvolved_moments:
806 functor: DeconvolvedMoments
807 filt: HSC-G
808 dataset: forced_src
809 refFlags:
810 - calib_psfUsed
811 - merge_measurement_i
812 - merge_measurement_r
813 - merge_measurement_z
814 - merge_measurement_y
815 - merge_measurement_g
816 - base_PixelFlags_flag_inexact_psfCenter
817 - detect_isPrimary
818
819 The names for each entry under "func" will become the names of columns in
820 the output dataset. All the functors referenced are defined in
821 `lsst.pipe.tasks.functors`. Positional arguments to be passed to each
822 functor are in the `args` list, and any additional entries for each column
823 other than "functor" or "args" (e.g., `'filt'`, `'dataset'`) are treated as
824 keyword arguments to be passed to the functor initialization.
825
826 The "flags" entry is the default shortcut for `Column` functors.
827 All columns listed under "flags" will be copied to the output table
828 untransformed. They can be of any datatype.
829 In the special case of transforming a multi-level oject table with
830 band and dataset indices (deepCoadd_obj), these will be taked from the
831 `meas` dataset and exploded out per band.
832
833 There are two special shortcuts that only apply when transforming
834 multi-level Object (deepCoadd_obj) tables:
835 - The "refFlags" entry is shortcut for `Column` functor
836 taken from the `'ref'` dataset if transforming an ObjectTable.
837 - The "forcedFlags" entry is shortcut for `Column` functors.
838 taken from the ``forced_src`` dataset if transforming an ObjectTable.
839 These are expanded out per band.
840
841
842 This task uses the `lsst.pipe.tasks.postprocess.PostprocessAnalysis` object
843 to organize and excecute the calculations.
844 """
845 @property
846 def _DefaultName(self):
847 raise NotImplementedError('Subclass must define "_DefaultName" attribute')
848
849 @property
850 def outputDataset(self):
851 raise NotImplementedError('Subclass must define "outputDataset" attribute')
852
853 @property
854 def inputDataset(self):
855 raise NotImplementedError('Subclass must define "inputDataset" attribute')
856
857 @property
858 def ConfigClass(self):
859 raise NotImplementedError('Subclass must define "ConfigClass" attribute')
860
861 def __init__(self, *args, **kwargs):
862 super().__init__(*args, **kwargs)
863 if self.config.functorFile:
864 self.log.info('Loading tranform functor definitions from %s',
865 self.config.functorFile)
866 self.funcs = CompositeFunctor.from_file(self.config.functorFile)
867 self.funcs.update(dict(PostprocessAnalysis._defaultFuncs))
868 else:
869 self.funcs = None
870
871 def runQuantum(self, butlerQC, inputRefs, outputRefs):
872 inputs = butlerQC.get(inputRefs)
873 if self.funcs is None:
874 raise ValueError("config.functorFile is None. "
875 "Must be a valid path to yaml in order to run Task as a PipelineTask.")
876 result = self.run(parq=inputs['inputCatalog'], funcs=self.funcs,
877 dataId=outputRefs.outputCatalog.dataId.full)
878 outputs = pipeBase.Struct(outputCatalog=result)
879 butlerQC.put(outputs, outputRefs)
880
881 def run(self, parq, funcs=None, dataId=None, band=None):
882 """Do postprocessing calculations
883
884 Takes a `ParquetTable` object and dataId,
885 returns a dataframe with results of postprocessing calculations.
886
887 Parameters
888 ----------
890 ParquetTable from which calculations are done.
891 funcs : `lsst.pipe.tasks.functors.Functors`
892 Functors to apply to the table's columns
893 dataId : dict, optional
894 Used to add a `patchId` column to the output dataframe.
895 band : `str`, optional
896 Filter band that is being processed.
897
898 Returns
899 ------
900 df : `pandas.DataFrame`
901 """
902 self.log.info("Transforming/standardizing the source table dataId: %s", dataId)
903
904 df = self.transform(band, parq, funcs, dataId).df
905 self.log.info("Made a table of %d columns and %d rows", len(df.columns), len(df))
906 return df
907
908 def getFunctors(self):
909 return self.funcs
910
911 def getAnalysis(self, parq, funcs=None, band=None):
912 if funcs is None:
913 funcs = self.funcs
914 analysis = PostprocessAnalysis(parq, funcs, filt=band)
915 return analysis
916
917 def transform(self, band, parq, funcs, dataId):
918 analysis = self.getAnalysis(parq, funcs=funcs, band=band)
919 df = analysis.df
920 if dataId and self.config.columnsFromDataId:
921 for key in self.config.columnsFromDataId:
922 if key in dataId:
923 df[str(key)] = dataId[key]
924 else:
925 raise ValueError(f"'{key}' in config.columnsFromDataId not found in dataId: {dataId}")
926
927 if self.config.primaryKey:
928 if df.index.name != self.config.primaryKey and self.config.primaryKey in df:
929 df.reset_index(inplace=True, drop=True)
930 df.set_index(self.config.primaryKey, inplace=True)
931
932 return pipeBase.Struct(
933 df=df,
934 analysis=analysis
935 )
936
937
938class TransformObjectCatalogConnections(pipeBase.PipelineTaskConnections,
939 defaultTemplates={"coaddName": "deep"},
940 dimensions=("tract", "patch", "skymap")):
941 inputCatalog = connectionTypes.Input(
942 doc="The vertical concatenation of the deepCoadd_{ref|meas|forced_src} catalogs, "
943 "stored as a DataFrame with a multi-level column index per-patch.",
944 dimensions=("tract", "patch", "skymap"),
945 storageClass="DataFrame",
946 name="{coaddName}Coadd_obj",
947 deferLoad=True,
948 )
949 outputCatalog = connectionTypes.Output(
950 doc="Per-Patch Object Table of columns transformed from the deepCoadd_obj table per the standard "
951 "data model.",
952 dimensions=("tract", "patch", "skymap"),
953 storageClass="DataFrame",
954 name="objectTable"
955 )
956
957
958class TransformObjectCatalogConfig(TransformCatalogBaseConfig,
959 pipelineConnections=TransformObjectCatalogConnections):
960 coaddName = pexConfig.Field(
961 dtype=str,
962 default="deep",
963 doc="Name of coadd"
964 )
965 # TODO: remove in DM-27177
966 filterMap = pexConfig.DictField(
967 keytype=str,
968 itemtype=str,
969 default={},
970 doc=("Dictionary mapping full filter name to short one for column name munging."
971 "These filters determine the output columns no matter what filters the "
972 "input data actually contain."),
973 deprecated=("Coadds are now identified by the band, so this transform is unused."
974 "Will be removed after v22.")
975 )
976 outputBands = pexConfig.ListField(
977 dtype=str,
978 default=None,
979 optional=True,
980 doc=("These bands and only these bands will appear in the output,"
981 " NaN-filled if the input does not include them."
982 " If None, then use all bands found in the input.")
983 )
984 camelCase = pexConfig.Field(
985 dtype=bool,
986 default=False,
987 doc=("Write per-band columns names with camelCase, else underscore "
988 "For example: gPsFlux instead of g_PsFlux.")
989 )
990 multilevelOutput = pexConfig.Field(
991 dtype=bool,
992 default=False,
993 doc=("Whether results dataframe should have a multilevel column index (True) or be flat "
994 "and name-munged (False).")
995 )
996 goodFlags = pexConfig.ListField(
997 dtype=str,
998 default=[],
999 doc=("List of 'good' flags that should be set False when populating empty tables. "
1000 "All other flags are considered to be 'bad' flags and will be set to True.")
1001 )
1002 floatFillValue = pexConfig.Field(
1003 dtype=float,
1004 default=np.nan,
1005 doc="Fill value for float fields when populating empty tables."
1006 )
1007 integerFillValue = pexConfig.Field(
1008 dtype=int,
1009 default=-1,
1010 doc="Fill value for integer fields when populating empty tables."
1011 )
1012
1013 def setDefaults(self):
1014 super().setDefaults()
1015 self.functorFile = os.path.join('$PIPE_TASKS_DIR', 'schemas', 'Object.yaml')
1016 self.primaryKey = 'objectId'
1017 self.columnsFromDataId = ['tract', 'patch']
1018 self.goodFlags = ['calib_astrometry_used',
1019 'calib_photometry_reserved',
1020 'calib_photometry_used',
1021 'calib_psf_candidate',
1022 'calib_psf_reserved',
1023 'calib_psf_used']
1024
1025
1026class TransformObjectCatalogTask(TransformCatalogBaseTask):
1027 """Produce a flattened Object Table to match the format specified in
1028 sdm_schemas.
1029
1030 Do the same set of postprocessing calculations on all bands.
1031
1032 This is identical to `TransformCatalogBaseTask`, except for that it does
1033 the specified functor calculations for all filters present in the
1034 input `deepCoadd_obj` table. Any specific `"filt"` keywords specified
1035 by the YAML file will be superceded.
1036 """
1037 _DefaultName = "transformObjectCatalog"
1038 ConfigClass = TransformObjectCatalogConfig
1039
1040 def run(self, parq, funcs=None, dataId=None, band=None):
1041 # NOTE: band kwarg is ignored here.
1042 dfDict = {}
1043 analysisDict = {}
1044 templateDf = pd.DataFrame()
1045
1046 if isinstance(parq, DeferredDatasetHandle):
1047 columns = parq.get(component='columns')
1048 inputBands = columns.unique(level=1).values
1049 else:
1050 inputBands = parq.columnLevelNames['band']
1051
1052 outputBands = self.config.outputBands if self.config.outputBands else inputBands
1053
1054 # Perform transform for data of filters that exist in parq.
1055 for inputBand in inputBands:
1056 if inputBand not in outputBands:
1057 self.log.info("Ignoring %s band data in the input", inputBand)
1058 continue
1059 self.log.info("Transforming the catalog of band %s", inputBand)
1060 result = self.transform(inputBand, parq, funcs, dataId)
1061 dfDict[inputBand] = result.df
1062 analysisDict[inputBand] = result.analysis
1063 if templateDf.empty:
1064 templateDf = result.df
1065
1066 # Put filler values in columns of other wanted bands
1067 for filt in outputBands:
1068 if filt not in dfDict:
1069 self.log.info("Adding empty columns for band %s", filt)
1070 dfTemp = templateDf.copy()
1071 for col in dfTemp.columns:
1072 testValue = dfTemp[col].values[0]
1073 if isinstance(testValue, (np.bool_, pd.BooleanDtype)):
1074 # Boolean flag type, check if it is a "good" flag
1075 if col in self.config.goodFlags:
1076 fillValue = False
1077 else:
1078 fillValue = True
1079 elif isinstance(testValue, numbers.Integral):
1080 # Checking numbers.Integral catches all flavors
1081 # of python, numpy, pandas, etc. integers.
1082 # We must ensure this is not an unsigned integer.
1083 if isinstance(testValue, np.unsignedinteger):
1084 raise ValueError("Parquet tables may not have unsigned integer columns.")
1085 else:
1086 fillValue = self.config.integerFillValue
1087 else:
1088 fillValue = self.config.floatFillValue
1089 dfTemp[col].values[:] = fillValue
1090 dfDict[filt] = dfTemp
1091
1092 # This makes a multilevel column index, with band as first level
1093 df = pd.concat(dfDict, axis=1, names=['band', 'column'])
1094
1095 if not self.config.multilevelOutput:
1096 noDupCols = list(set.union(*[set(v.noDupCols) for v in analysisDict.values()]))
1097 if self.config.primaryKey in noDupCols:
1098 noDupCols.remove(self.config.primaryKey)
1099 if dataId and self.config.columnsFromDataId:
1100 noDupCols += self.config.columnsFromDataId
1101 df = flattenFilters(df, noDupCols=noDupCols, camelCase=self.config.camelCase,
1102 inputBands=inputBands)
1103
1104 self.log.info("Made a table of %d columns and %d rows", len(df.columns), len(df))
1105
1106 return df
1107
1108
1109class ConsolidateObjectTableConnections(pipeBase.PipelineTaskConnections,
1110 dimensions=("tract", "skymap")):
1111 inputCatalogs = connectionTypes.Input(
1112 doc="Per-Patch objectTables conforming to the standard data model.",
1113 name="objectTable",
1114 storageClass="DataFrame",
1115 dimensions=("tract", "patch", "skymap"),
1116 multiple=True,
1117 )
1118 outputCatalog = connectionTypes.Output(
1119 doc="Pre-tract horizontal concatenation of the input objectTables",
1120 name="objectTable_tract",
1121 storageClass="DataFrame",
1122 dimensions=("tract", "skymap"),
1123 )
1124
1125
1126class ConsolidateObjectTableConfig(pipeBase.PipelineTaskConfig,
1127 pipelineConnections=ConsolidateObjectTableConnections):
1128 coaddName = pexConfig.Field(
1129 dtype=str,
1130 default="deep",
1131 doc="Name of coadd"
1132 )
1133
1134
1135class ConsolidateObjectTableTask(pipeBase.PipelineTask):
1136 """Write patch-merged source tables to a tract-level parquet file.
1137
1138 Concatenates `objectTable` list into a per-visit `objectTable_tract`.
1139 """
1140 _DefaultName = "consolidateObjectTable"
1141 ConfigClass = ConsolidateObjectTableConfig
1142
1143 inputDataset = 'objectTable'
1144 outputDataset = 'objectTable_tract'
1145
1146 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1147 inputs = butlerQC.get(inputRefs)
1148 self.log.info("Concatenating %s per-patch Object Tables",
1149 len(inputs['inputCatalogs']))
1150 df = pd.concat(inputs['inputCatalogs'])
1151 butlerQC.put(pipeBase.Struct(outputCatalog=df), outputRefs)
1152
1153
1154class TransformSourceTableConnections(pipeBase.PipelineTaskConnections,
1155 defaultTemplates={"catalogType": ""},
1156 dimensions=("instrument", "visit", "detector")):
1157
1158 inputCatalog = connectionTypes.Input(
1159 doc="Wide input catalog of sources produced by WriteSourceTableTask",
1160 name="{catalogType}source",
1161 storageClass="DataFrame",
1162 dimensions=("instrument", "visit", "detector"),
1163 deferLoad=True
1164 )
1165 outputCatalog = connectionTypes.Output(
1166 doc="Narrower, per-detector Source Table transformed and converted per a "
1167 "specified set of functors",
1168 name="{catalogType}sourceTable",
1169 storageClass="DataFrame",
1170 dimensions=("instrument", "visit", "detector")
1171 )
1172
1173
1174class TransformSourceTableConfig(TransformCatalogBaseConfig,
1175 pipelineConnections=TransformSourceTableConnections):
1176
1177 def setDefaults(self):
1178 super().setDefaults()
1179 self.functorFile = os.path.join('$PIPE_TASKS_DIR', 'schemas', 'Source.yaml')
1180 self.primaryKey = 'sourceId'
1181 self.columnsFromDataId = ['visit', 'detector', 'band', 'physical_filter']
1182
1183
1184class TransformSourceTableTask(TransformCatalogBaseTask):
1185 """Transform/standardize a source catalog
1186 """
1187 _DefaultName = "transformSourceTable"
1188 ConfigClass = TransformSourceTableConfig
1189
1190
1191class ConsolidateVisitSummaryConnections(pipeBase.PipelineTaskConnections,
1192 dimensions=("instrument", "visit",),
1193 defaultTemplates={"calexpType": ""}):
1194 calexp = connectionTypes.Input(
1195 doc="Processed exposures used for metadata",
1196 name="{calexpType}calexp",
1197 storageClass="ExposureF",
1198 dimensions=("instrument", "visit", "detector"),
1199 deferLoad=True,
1200 multiple=True,
1201 )
1202 visitSummary = connectionTypes.Output(
1203 doc=("Per-visit consolidated exposure metadata. These catalogs use "
1204 "detector id for the id and are sorted for fast lookups of a "
1205 "detector."),
1206 name="{calexpType}visitSummary",
1207 storageClass="ExposureCatalog",
1208 dimensions=("instrument", "visit"),
1209 )
1210
1211
1212class ConsolidateVisitSummaryConfig(pipeBase.PipelineTaskConfig,
1213 pipelineConnections=ConsolidateVisitSummaryConnections):
1214 """Config for ConsolidateVisitSummaryTask"""
1215 pass
1216
1217
1218class ConsolidateVisitSummaryTask(pipeBase.PipelineTask):
1219 """Task to consolidate per-detector visit metadata.
1220
1221 This task aggregates the following metadata from all the detectors in a
1222 single visit into an exposure catalog:
1223 - The visitInfo.
1224 - The wcs.
1225 - The photoCalib.
1226 - The physical_filter and band (if available).
1227 - The psf size, shape, and effective area at the center of the detector.
1228 - The corners of the bounding box in right ascension/declination.
1229
1230 Other quantities such as Detector, Psf, ApCorrMap, and TransmissionCurve
1231 are not persisted here because of storage concerns, and because of their
1232 limited utility as summary statistics.
1233
1234 Tests for this task are performed in ci_hsc_gen3.
1235 """
1236 _DefaultName = "consolidateVisitSummary"
1237 ConfigClass = ConsolidateVisitSummaryConfig
1238
1239 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1240 dataRefs = butlerQC.get(inputRefs.calexp)
1241 visit = dataRefs[0].dataId.byName()['visit']
1242
1243 self.log.debug("Concatenating metadata from %d per-detector calexps (visit %d)",
1244 len(dataRefs), visit)
1245
1246 expCatalog = self._combineExposureMetadata(visit, dataRefs)
1247
1248 butlerQC.put(expCatalog, outputRefs.visitSummary)
1249
1250 def _combineExposureMetadata(self, visit, dataRefs):
1251 """Make a combined exposure catalog from a list of dataRefs.
1252 These dataRefs must point to exposures with wcs, summaryStats,
1253 and other visit metadata.
1254
1255 Parameters
1256 ----------
1257 visit : `int`
1258 Visit identification number.
1259 dataRefs : `list` of `lsst.daf.butler.DeferredDatasetHandle`
1260 List of dataRefs in visit.
1261
1262 Returns
1263 -------
1264 visitSummary : `lsst.afw.table.ExposureCatalog`
1265 Exposure catalog with per-detector summary information.
1266 """
1267 schema = self._makeVisitSummarySchema()
1268 cat = afwTable.ExposureCatalog(schema)
1269 cat.resize(len(dataRefs))
1270
1271 cat['visit'] = visit
1272
1273 for i, dataRef in enumerate(dataRefs):
1274 visitInfo = dataRef.get(component='visitInfo')
1275 filterLabel = dataRef.get(component='filter')
1276 summaryStats = dataRef.get(component='summaryStats')
1277 detector = dataRef.get(component='detector')
1278 wcs = dataRef.get(component='wcs')
1279 photoCalib = dataRef.get(component='photoCalib')
1280 detector = dataRef.get(component='detector')
1281 bbox = dataRef.get(component='bbox')
1282 validPolygon = dataRef.get(component='validPolygon')
1283
1284 rec = cat[i]
1285 rec.setBBox(bbox)
1286 rec.setVisitInfo(visitInfo)
1287 rec.setWcs(wcs)
1288 rec.setPhotoCalib(photoCalib)
1289 rec.setValidPolygon(validPolygon)
1290
1291 rec['physical_filter'] = filterLabel.physicalLabel if filterLabel.hasPhysicalLabel() else ""
1292 rec['band'] = filterLabel.bandLabel if filterLabel.hasBandLabel() else ""
1293 rec.setId(detector.getId())
1294 rec['psfSigma'] = summaryStats.psfSigma
1295 rec['psfIxx'] = summaryStats.psfIxx
1296 rec['psfIyy'] = summaryStats.psfIyy
1297 rec['psfIxy'] = summaryStats.psfIxy
1298 rec['psfArea'] = summaryStats.psfArea
1299 rec['raCorners'][:] = summaryStats.raCorners
1300 rec['decCorners'][:] = summaryStats.decCorners
1301 rec['ra'] = summaryStats.ra
1302 rec['decl'] = summaryStats.decl
1303 rec['zenithDistance'] = summaryStats.zenithDistance
1304 rec['zeroPoint'] = summaryStats.zeroPoint
1305 rec['skyBg'] = summaryStats.skyBg
1306 rec['skyNoise'] = summaryStats.skyNoise
1307 rec['meanVar'] = summaryStats.meanVar
1308 rec['astromOffsetMean'] = summaryStats.astromOffsetMean
1309 rec['astromOffsetStd'] = summaryStats.astromOffsetStd
1310 rec['nPsfStar'] = summaryStats.nPsfStar
1311 rec['psfStarDeltaE1Median'] = summaryStats.psfStarDeltaE1Median
1312 rec['psfStarDeltaE2Median'] = summaryStats.psfStarDeltaE2Median
1313 rec['psfStarDeltaE1Scatter'] = summaryStats.psfStarDeltaE1Scatter
1314 rec['psfStarDeltaE2Scatter'] = summaryStats.psfStarDeltaE2Scatter
1315 rec['psfStarDeltaSizeMedian'] = summaryStats.psfStarDeltaSizeMedian
1316 rec['psfStarDeltaSizeScatter'] = summaryStats.psfStarDeltaSizeScatter
1317 rec['psfStarScaledDeltaSizeScatter'] = summaryStats.psfStarScaledDeltaSizeScatter
1318
1319 metadata = dafBase.PropertyList()
1320 metadata.add("COMMENT", "Catalog id is detector id, sorted.")
1321 # We are looping over existing datarefs, so the following is true
1322 metadata.add("COMMENT", "Only detectors with data have entries.")
1323 cat.setMetadata(metadata)
1324
1325 cat.sort()
1326 return cat
1327
1328 def _makeVisitSummarySchema(self):
1329 """Make the schema for the visitSummary catalog."""
1330 schema = afwTable.ExposureTable.makeMinimalSchema()
1331 schema.addField('visit', type='L', doc='Visit number')
1332 schema.addField('physical_filter', type='String', size=32, doc='Physical filter')
1333 schema.addField('band', type='String', size=32, doc='Name of band')
1334 schema.addField('psfSigma', type='F',
1335 doc='PSF model second-moments determinant radius (center of chip) (pixel)')
1336 schema.addField('psfArea', type='F',
1337 doc='PSF model effective area (center of chip) (pixel**2)')
1338 schema.addField('psfIxx', type='F',
1339 doc='PSF model Ixx (center of chip) (pixel**2)')
1340 schema.addField('psfIyy', type='F',
1341 doc='PSF model Iyy (center of chip) (pixel**2)')
1342 schema.addField('psfIxy', type='F',
1343 doc='PSF model Ixy (center of chip) (pixel**2)')
1344 schema.addField('raCorners', type='ArrayD', size=4,
1345 doc='Right Ascension of bounding box corners (degrees)')
1346 schema.addField('decCorners', type='ArrayD', size=4,
1347 doc='Declination of bounding box corners (degrees)')
1348 schema.addField('ra', type='D',
1349 doc='Right Ascension of bounding box center (degrees)')
1350 schema.addField('decl', type='D',
1351 doc='Declination of bounding box center (degrees)')
1352 schema.addField('zenithDistance', type='F',
1353 doc='Zenith distance of bounding box center (degrees)')
1354 schema.addField('zeroPoint', type='F',
1355 doc='Mean zeropoint in detector (mag)')
1356 schema.addField('skyBg', type='F',
1357 doc='Average sky background (ADU)')
1358 schema.addField('skyNoise', type='F',
1359 doc='Average sky noise (ADU)')
1360 schema.addField('meanVar', type='F',
1361 doc='Mean variance of the weight plane (ADU**2)')
1362 schema.addField('astromOffsetMean', type='F',
1363 doc='Mean offset of astrometric calibration matches (arcsec)')
1364 schema.addField('astromOffsetStd', type='F',
1365 doc='Standard deviation of offsets of astrometric calibration matches (arcsec)')
1366 schema.addField('nPsfStar', type='I', doc='Number of stars used for PSF model')
1367 schema.addField('psfStarDeltaE1Median', type='F',
1368 doc='Median E1 residual (starE1 - psfE1) for psf stars')
1369 schema.addField('psfStarDeltaE2Median', type='F',
1370 doc='Median E2 residual (starE2 - psfE2) for psf stars')
1371 schema.addField('psfStarDeltaE1Scatter', type='F',
1372 doc='Scatter (via MAD) of E1 residual (starE1 - psfE1) for psf stars')
1373 schema.addField('psfStarDeltaE2Scatter', type='F',
1374 doc='Scatter (via MAD) of E2 residual (starE2 - psfE2) for psf stars')
1375 schema.addField('psfStarDeltaSizeMedian', type='F',
1376 doc='Median size residual (starSize - psfSize) for psf stars (pixel)')
1377 schema.addField('psfStarDeltaSizeScatter', type='F',
1378 doc='Scatter (via MAD) of size residual (starSize - psfSize) for psf stars (pixel)')
1379 schema.addField('psfStarScaledDeltaSizeScatter', type='F',
1380 doc='Scatter (via MAD) of size residual scaled by median size squared')
1381
1382 return schema
1383
1384
1385class ConsolidateSourceTableConnections(pipeBase.PipelineTaskConnections,
1386 defaultTemplates={"catalogType": ""},
1387 dimensions=("instrument", "visit")):
1388 inputCatalogs = connectionTypes.Input(
1389 doc="Input per-detector Source Tables",
1390 name="{catalogType}sourceTable",
1391 storageClass="DataFrame",
1392 dimensions=("instrument", "visit", "detector"),
1393 multiple=True
1394 )
1395 outputCatalog = connectionTypes.Output(
1396 doc="Per-visit concatenation of Source Table",
1397 name="{catalogType}sourceTable_visit",
1398 storageClass="DataFrame",
1399 dimensions=("instrument", "visit")
1400 )
1401
1402
1403class ConsolidateSourceTableConfig(pipeBase.PipelineTaskConfig,
1404 pipelineConnections=ConsolidateSourceTableConnections):
1405 pass
1406
1407
1408class ConsolidateSourceTableTask(pipeBase.PipelineTask):
1409 """Concatenate `sourceTable` list into a per-visit `sourceTable_visit`
1410 """
1411 _DefaultName = 'consolidateSourceTable'
1412 ConfigClass = ConsolidateSourceTableConfig
1413
1414 inputDataset = 'sourceTable'
1415 outputDataset = 'sourceTable_visit'
1416
1417 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1418 from .makeWarp import reorderRefs
1419
1420 detectorOrder = [ref.dataId['detector'] for ref in inputRefs.inputCatalogs]
1421 detectorOrder.sort()
1422 inputRefs = reorderRefs(inputRefs, detectorOrder, dataIdKey='detector')
1423 inputs = butlerQC.get(inputRefs)
1424 self.log.info("Concatenating %s per-detector Source Tables",
1425 len(inputs['inputCatalogs']))
1426 df = pd.concat(inputs['inputCatalogs'])
1427 butlerQC.put(pipeBase.Struct(outputCatalog=df), outputRefs)
1428
1429
1430class MakeCcdVisitTableConnections(pipeBase.PipelineTaskConnections,
1431 dimensions=("instrument",),
1432 defaultTemplates={"calexpType": ""}):
1433 visitSummaryRefs = connectionTypes.Input(
1434 doc="Data references for per-visit consolidated exposure metadata from ConsolidateVisitSummaryTask",
1435 name="{calexpType}visitSummary",
1436 storageClass="ExposureCatalog",
1437 dimensions=("instrument", "visit"),
1438 multiple=True,
1439 deferLoad=True,
1440 )
1441 outputCatalog = connectionTypes.Output(
1442 doc="CCD and Visit metadata table",
1443 name="ccdVisitTable",
1444 storageClass="DataFrame",
1445 dimensions=("instrument",)
1446 )
1447
1448
1449class MakeCcdVisitTableConfig(pipeBase.PipelineTaskConfig,
1450 pipelineConnections=MakeCcdVisitTableConnections):
1451 pass
1452
1453
1454class MakeCcdVisitTableTask(pipeBase.PipelineTask):
1455 """Produce a `ccdVisitTable` from the `visitSummary` exposure catalogs.
1456 """
1457 _DefaultName = 'makeCcdVisitTable'
1458 ConfigClass = MakeCcdVisitTableConfig
1459
1460 def run(self, visitSummaryRefs):
1461 """Make a table of ccd information from the `visitSummary` catalogs.
1462
1463 Parameters
1464 ----------
1465 visitSummaryRefs : `list` of `lsst.daf.butler.DeferredDatasetHandle`
1466 List of DeferredDatasetHandles pointing to exposure catalogs with
1467 per-detector summary information.
1468
1469 Returns
1470 -------
1471 result : `lsst.pipe.Base.Struct`
1472 Results struct with attribute:
1473
1474 ``outputCatalog``
1475 Catalog of ccd and visit information.
1476 """
1477 ccdEntries = []
1478 for visitSummaryRef in visitSummaryRefs:
1479 visitSummary = visitSummaryRef.get()
1480 visitInfo = visitSummary[0].getVisitInfo()
1481
1482 ccdEntry = {}
1483 summaryTable = visitSummary.asAstropy()
1484 selectColumns = ['id', 'visit', 'physical_filter', 'band', 'ra', 'decl', 'zenithDistance',
1485 'zeroPoint', 'psfSigma', 'skyBg', 'skyNoise',
1486 'astromOffsetMean', 'astromOffsetStd', 'nPsfStar',
1487 'psfStarDeltaE1Median', 'psfStarDeltaE2Median',
1488 'psfStarDeltaE1Scatter', 'psfStarDeltaE2Scatter',
1489 'psfStarDeltaSizeMedian', 'psfStarDeltaSizeScatter',
1490 'psfStarScaledDeltaSizeScatter']
1491 ccdEntry = summaryTable[selectColumns].to_pandas().set_index('id')
1492 # 'visit' is the human readable visit number.
1493 # 'visitId' is the key to the visitId table. They are the same.
1494 # Technically you should join to get the visit from the visit
1495 # table.
1496 ccdEntry = ccdEntry.rename(columns={"visit": "visitId"})
1497 dataIds = [DataCoordinate.standardize(visitSummaryRef.dataId, detector=id) for id in
1498 summaryTable['id']]
1499 packer = visitSummaryRef.dataId.universe.makePacker('visit_detector', visitSummaryRef.dataId)
1500 ccdVisitIds = [packer.pack(dataId) for dataId in dataIds]
1501 ccdEntry['ccdVisitId'] = ccdVisitIds
1502 ccdEntry['detector'] = summaryTable['id']
1503 pixToArcseconds = np.array([vR.getWcs().getPixelScale().asArcseconds() for vR in visitSummary])
1504 ccdEntry["seeing"] = visitSummary['psfSigma'] * np.sqrt(8 * np.log(2)) * pixToArcseconds
1505
1506 ccdEntry["skyRotation"] = visitInfo.getBoresightRotAngle().asDegrees()
1507 ccdEntry["expMidpt"] = visitInfo.getDate().toPython()
1508 ccdEntry["expMidptMJD"] = visitInfo.getDate().get(dafBase.DateTime.MJD)
1509 expTime = visitInfo.getExposureTime()
1510 ccdEntry['expTime'] = expTime
1511 ccdEntry["obsStart"] = ccdEntry["expMidpt"] - 0.5 * pd.Timedelta(seconds=expTime)
1512 expTime_days = expTime / (60*60*24)
1513 ccdEntry["obsStartMJD"] = ccdEntry["expMidptMJD"] - 0.5 * expTime_days
1514 ccdEntry['darkTime'] = visitInfo.getDarkTime()
1515 ccdEntry['xSize'] = summaryTable['bbox_max_x'] - summaryTable['bbox_min_x']
1516 ccdEntry['ySize'] = summaryTable['bbox_max_y'] - summaryTable['bbox_min_y']
1517 ccdEntry['llcra'] = summaryTable['raCorners'][:, 0]
1518 ccdEntry['llcdec'] = summaryTable['decCorners'][:, 0]
1519 ccdEntry['ulcra'] = summaryTable['raCorners'][:, 1]
1520 ccdEntry['ulcdec'] = summaryTable['decCorners'][:, 1]
1521 ccdEntry['urcra'] = summaryTable['raCorners'][:, 2]
1522 ccdEntry['urcdec'] = summaryTable['decCorners'][:, 2]
1523 ccdEntry['lrcra'] = summaryTable['raCorners'][:, 3]
1524 ccdEntry['lrcdec'] = summaryTable['decCorners'][:, 3]
1525 # TODO: DM-30618, Add raftName, nExposures, ccdTemp, binX, binY,
1526 # and flags, and decide if WCS, and llcx, llcy, ulcx, ulcy, etc.
1527 # values are actually wanted.
1528 ccdEntries.append(ccdEntry)
1529
1530 outputCatalog = pd.concat(ccdEntries)
1531 outputCatalog.set_index('ccdVisitId', inplace=True, verify_integrity=True)
1532 return pipeBase.Struct(outputCatalog=outputCatalog)
1533
1534
1535class MakeVisitTableConnections(pipeBase.PipelineTaskConnections,
1536 dimensions=("instrument",),
1537 defaultTemplates={"calexpType": ""}):
1538 visitSummaries = connectionTypes.Input(
1539 doc="Per-visit consolidated exposure metadata from ConsolidateVisitSummaryTask",
1540 name="{calexpType}visitSummary",
1541 storageClass="ExposureCatalog",
1542 dimensions=("instrument", "visit",),
1543 multiple=True,
1544 deferLoad=True,
1545 )
1546 outputCatalog = connectionTypes.Output(
1547 doc="Visit metadata table",
1548 name="visitTable",
1549 storageClass="DataFrame",
1550 dimensions=("instrument",)
1551 )
1552
1553
1554class MakeVisitTableConfig(pipeBase.PipelineTaskConfig,
1555 pipelineConnections=MakeVisitTableConnections):
1556 pass
1557
1558
1559class MakeVisitTableTask(pipeBase.PipelineTask):
1560 """Produce a `visitTable` from the `visitSummary` exposure catalogs.
1561 """
1562 _DefaultName = 'makeVisitTable'
1563 ConfigClass = MakeVisitTableConfig
1564
1565 def run(self, visitSummaries):
1566 """Make a table of visit information from the `visitSummary` catalogs.
1567
1568 Parameters
1569 ----------
1570 visitSummaries : `list` of `lsst.afw.table.ExposureCatalog`
1571 List of exposure catalogs with per-detector summary information.
1572 Returns
1573 -------
1574 result : `lsst.pipe.Base.Struct`
1575 Results struct with attribute:
1576
1577 ``outputCatalog``
1578 Catalog of visit information.
1579 """
1580 visitEntries = []
1581 for visitSummary in visitSummaries:
1582 visitSummary = visitSummary.get()
1583 visitRow = visitSummary[0]
1584 visitInfo = visitRow.getVisitInfo()
1585
1586 visitEntry = {}
1587 visitEntry["visitId"] = visitRow['visit']
1588 visitEntry["visit"] = visitRow['visit']
1589 visitEntry["physical_filter"] = visitRow['physical_filter']
1590 visitEntry["band"] = visitRow['band']
1591 raDec = visitInfo.getBoresightRaDec()
1592 visitEntry["ra"] = raDec.getRa().asDegrees()
1593 visitEntry["decl"] = raDec.getDec().asDegrees()
1594 visitEntry["skyRotation"] = visitInfo.getBoresightRotAngle().asDegrees()
1595 azAlt = visitInfo.getBoresightAzAlt()
1596 visitEntry["azimuth"] = azAlt.getLongitude().asDegrees()
1597 visitEntry["altitude"] = azAlt.getLatitude().asDegrees()
1598 visitEntry["zenithDistance"] = 90 - azAlt.getLatitude().asDegrees()
1599 visitEntry["airmass"] = visitInfo.getBoresightAirmass()
1600 expTime = visitInfo.getExposureTime()
1601 visitEntry["expTime"] = expTime
1602 visitEntry["expMidpt"] = visitInfo.getDate().toPython()
1603 visitEntry["expMidptMJD"] = visitInfo.getDate().get(dafBase.DateTime.MJD)
1604 visitEntry["obsStart"] = visitEntry["expMidpt"] - 0.5 * pd.Timedelta(seconds=expTime)
1605 expTime_days = expTime / (60*60*24)
1606 visitEntry["obsStartMJD"] = visitEntry["expMidptMJD"] - 0.5 * expTime_days
1607 visitEntries.append(visitEntry)
1608
1609 # TODO: DM-30623, Add programId, exposureType, cameraTemp,
1610 # mirror1Temp, mirror2Temp, mirror3Temp, domeTemp, externalTemp,
1611 # dimmSeeing, pwvGPS, pwvMW, flags, nExposures.
1612
1613 outputCatalog = pd.DataFrame(data=visitEntries)
1614 outputCatalog.set_index('visitId', inplace=True, verify_integrity=True)
1615 return pipeBase.Struct(outputCatalog=outputCatalog)
1616
1617
1618class WriteForcedSourceTableConnections(pipeBase.PipelineTaskConnections,
1619 dimensions=("instrument", "visit", "detector", "skymap", "tract")):
1620
1621 inputCatalog = connectionTypes.Input(
1622 doc="Primary per-detector, single-epoch forced-photometry catalog. "
1623 "By default, it is the output of ForcedPhotCcdTask on calexps",
1624 name="forced_src",
1625 storageClass="SourceCatalog",
1626 dimensions=("instrument", "visit", "detector", "skymap", "tract")
1627 )
1628 inputCatalogDiff = connectionTypes.Input(
1629 doc="Secondary multi-epoch, per-detector, forced photometry catalog. "
1630 "By default, it is the output of ForcedPhotCcdTask run on image differences.",
1631 name="forced_diff",
1632 storageClass="SourceCatalog",
1633 dimensions=("instrument", "visit", "detector", "skymap", "tract")
1634 )
1635 outputCatalog = connectionTypes.Output(
1636 doc="InputCatalogs horizonatally joined on `objectId` in Parquet format",
1637 name="mergedForcedSource",
1638 storageClass="DataFrame",
1639 dimensions=("instrument", "visit", "detector", "skymap", "tract")
1640 )
1641
1642
1643class WriteForcedSourceTableConfig(pipeBase.PipelineTaskConfig,
1644 pipelineConnections=WriteForcedSourceTableConnections):
1645 key = lsst.pex.config.Field(
1646 doc="Column on which to join the two input tables on and make the primary key of the output",
1647 dtype=str,
1648 default="objectId",
1649 )
1650
1651
1652class WriteForcedSourceTableTask(pipeBase.PipelineTask):
1653 """Merge and convert per-detector forced source catalogs to parquet.
1654
1655 Because the predecessor ForcedPhotCcdTask operates per-detector,
1656 per-tract, (i.e., it has tract in its dimensions), detectors
1657 on the tract boundary may have multiple forced source catalogs.
1658
1659 The successor task TransformForcedSourceTable runs per-patch
1660 and temporally-aggregates overlapping mergedForcedSource catalogs from all
1661 available multiple epochs.
1662 """
1663 _DefaultName = "writeForcedSourceTable"
1664 ConfigClass = WriteForcedSourceTableConfig
1665
1666 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1667 inputs = butlerQC.get(inputRefs)
1668 # Add ccdVisitId to allow joining with CcdVisitTable
1669 inputs['ccdVisitId'] = butlerQC.quantum.dataId.pack("visit_detector")
1670 inputs['band'] = butlerQC.quantum.dataId.full['band']
1671 outputs = self.run(**inputs)
1672 butlerQC.put(outputs, outputRefs)
1673
1674 def run(self, inputCatalog, inputCatalogDiff, ccdVisitId=None, band=None):
1675 dfs = []
1676 for table, dataset, in zip((inputCatalog, inputCatalogDiff), ('calexp', 'diff')):
1677 df = table.asAstropy().to_pandas().set_index(self.config.key, drop=False)
1678 df = df.reindex(sorted(df.columns), axis=1)
1679 df['ccdVisitId'] = ccdVisitId if ccdVisitId else pd.NA
1680 df['band'] = band if band else pd.NA
1681 df.columns = pd.MultiIndex.from_tuples([(dataset, c) for c in df.columns],
1682 names=('dataset', 'column'))
1683
1684 dfs.append(df)
1685
1686 outputCatalog = functools.reduce(lambda d1, d2: d1.join(d2), dfs)
1687 return pipeBase.Struct(outputCatalog=outputCatalog)
1688
1689
1690class TransformForcedSourceTableConnections(pipeBase.PipelineTaskConnections,
1691 dimensions=("instrument", "skymap", "patch", "tract")):
1692
1693 inputCatalogs = connectionTypes.Input(
1694 doc="Parquet table of merged ForcedSources produced by WriteForcedSourceTableTask",
1695 name="mergedForcedSource",
1696 storageClass="DataFrame",
1697 dimensions=("instrument", "visit", "detector", "skymap", "tract"),
1698 multiple=True,
1699 deferLoad=True
1700 )
1701 referenceCatalog = connectionTypes.Input(
1702 doc="Reference catalog which was used to seed the forcedPhot. Columns "
1703 "objectId, detect_isPrimary, detect_isTractInner, detect_isPatchInner "
1704 "are expected.",
1705 name="objectTable",
1706 storageClass="DataFrame",
1707 dimensions=("tract", "patch", "skymap"),
1708 deferLoad=True
1709 )
1710 outputCatalog = connectionTypes.Output(
1711 doc="Narrower, temporally-aggregated, per-patch ForcedSource Table transformed and converted per a "
1712 "specified set of functors",
1713 name="forcedSourceTable",
1714 storageClass="DataFrame",
1715 dimensions=("tract", "patch", "skymap")
1716 )
1717
1718
1719class TransformForcedSourceTableConfig(TransformCatalogBaseConfig,
1720 pipelineConnections=TransformForcedSourceTableConnections):
1721 referenceColumns = pexConfig.ListField(
1722 dtype=str,
1723 default=["detect_isPrimary", "detect_isTractInner", "detect_isPatchInner"],
1724 optional=True,
1725 doc="Columns to pull from reference catalog",
1726 )
1727 keyRef = lsst.pex.config.Field(
1728 doc="Column on which to join the two input tables on and make the primary key of the output",
1729 dtype=str,
1730 default="objectId",
1731 )
1732 key = lsst.pex.config.Field(
1733 doc="Rename the output DataFrame index to this name",
1734 dtype=str,
1735 default="forcedSourceId",
1736 )
1737
1738 def setDefaults(self):
1739 super().setDefaults()
1740 self.functorFile = os.path.join('$PIPE_TASKS_DIR', 'schemas', 'ForcedSource.yaml')
1741 self.columnsFromDataId = ['tract', 'patch']
1742
1743
1744class TransformForcedSourceTableTask(TransformCatalogBaseTask):
1745 """Transform/standardize a ForcedSource catalog
1746
1747 Transforms each wide, per-detector forcedSource parquet table per the
1748 specification file (per-camera defaults found in ForcedSource.yaml).
1749 All epochs that overlap the patch are aggregated into one per-patch
1750 narrow-parquet file.
1751
1752 No de-duplication of rows is performed. Duplicate resolutions flags are
1753 pulled in from the referenceCatalog: `detect_isPrimary`,
1754 `detect_isTractInner`,`detect_isPatchInner`, so that user may de-duplicate
1755 for analysis or compare duplicates for QA.
1756
1757 The resulting table includes multiple bands. Epochs (MJDs) and other useful
1758 per-visit rows can be retreived by joining with the CcdVisitTable on
1759 ccdVisitId.
1760 """
1761 _DefaultName = "transformForcedSourceTable"
1762 ConfigClass = TransformForcedSourceTableConfig
1763
1764 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1765 inputs = butlerQC.get(inputRefs)
1766 if self.funcs is None:
1767 raise ValueError("config.functorFile is None. "
1768 "Must be a valid path to yaml in order to run Task as a PipelineTask.")
1769 outputs = self.run(inputs['inputCatalogs'], inputs['referenceCatalog'], funcs=self.funcs,
1770 dataId=outputRefs.outputCatalog.dataId.full)
1771
1772 butlerQC.put(outputs, outputRefs)
1773
1774 def run(self, inputCatalogs, referenceCatalog, funcs=None, dataId=None, band=None):
1775 dfs = []
1776 ref = referenceCatalog.get(parameters={"columns": self.config.referenceColumns})
1777 self.log.info("Aggregating %s input catalogs" % (len(inputCatalogs)))
1778 for handle in inputCatalogs:
1779 result = self.transform(None, handle, funcs, dataId)
1780 # Filter for only rows that were detected on (overlap) the patch
1781 dfs.append(result.df.join(ref, how='inner'))
1782
1783 outputCatalog = pd.concat(dfs)
1784
1785 # Now that we are done joining on config.keyRef
1786 # Change index to config.key by
1787 outputCatalog.index.rename(self.config.keyRef, inplace=True)
1788 # Add config.keyRef to the column list
1789 outputCatalog.reset_index(inplace=True)
1790 # Set the forcedSourceId to the index. This is specified in the
1791 # ForcedSource.yaml
1792 outputCatalog.set_index("forcedSourceId", inplace=True, verify_integrity=True)
1793 # Rename it to the config.key
1794 outputCatalog.index.rename(self.config.key, inplace=True)
1795
1796 self.log.info("Made a table of %d columns and %d rows",
1797 len(outputCatalog.columns), len(outputCatalog))
1798 return pipeBase.Struct(outputCatalog=outputCatalog)
1799
1800
1801class ConsolidateTractConnections(pipeBase.PipelineTaskConnections,
1802 defaultTemplates={"catalogType": ""},
1803 dimensions=("instrument", "tract")):
1804 inputCatalogs = connectionTypes.Input(
1805 doc="Input per-patch DataFrame Tables to be concatenated",
1806 name="{catalogType}ForcedSourceTable",
1807 storageClass="DataFrame",
1808 dimensions=("tract", "patch", "skymap"),
1809 multiple=True,
1810 )
1811
1812 outputCatalog = connectionTypes.Output(
1813 doc="Output per-tract concatenation of DataFrame Tables",
1814 name="{catalogType}ForcedSourceTable_tract",
1815 storageClass="DataFrame",
1816 dimensions=("tract", "skymap"),
1817 )
1818
1819
1820class ConsolidateTractConfig(pipeBase.PipelineTaskConfig,
1821 pipelineConnections=ConsolidateTractConnections):
1822 pass
1823
1824
1825class ConsolidateTractTask(pipeBase.PipelineTask):
1826 """Concatenate any per-patch, dataframe list into a single
1827 per-tract DataFrame.
1828 """
1829 _DefaultName = 'ConsolidateTract'
1830 ConfigClass = ConsolidateTractConfig
1831
1832 def runQuantum(self, butlerQC, inputRefs, outputRefs):
1833 inputs = butlerQC.get(inputRefs)
1834 # Not checking at least one inputCatalog exists because that'd be an
1835 # empty QG.
1836 self.log.info("Concatenating %s per-patch %s Tables",
1837 len(inputs['inputCatalogs']),
1838 inputRefs.inputCatalogs[0].datasetType.name)
1839 df = pd.concat(inputs['inputCatalogs'])
1840 butlerQC.put(pipeBase.Struct(outputCatalog=df), outputRefs)
def compute(self, dropna=False, pool=None)
Definition: postprocess.py:711
def __init__(self, parq, functors, filt=None, flags=None, refFlags=None, forcedFlags=None)
Definition: postprocess.py:666
def getAnalysis(self, parq, funcs=None, band=None)
Definition: postprocess.py:911
def transform(self, band, parq, funcs, dataId)
Definition: postprocess.py:917
def run(self, parq, funcs=None, dataId=None, band=None)
Definition: postprocess.py:881
def runQuantum(self, butlerQC, inputRefs, outputRefs)
Definition: postprocess.py:871
def flattenFilters(df, noDupCols=['coord_ra', 'coord_dec'], camelCase=False, inputBands=None)
Definition: postprocess.py:46