Hide keyboard shortcuts

Hot-keys on this page

r m x p   toggle line displays

j k   next/prev highlighted chunk

0   (zero) top of page

1   (one) first highlighted chunk

1# This file is part of pipe_tasks. 

2# 

3# LSST Data Management System 

4# This product includes software developed by the 

5# LSST Project (http://www.lsst.org/). 

6# See COPYRIGHT file at the top of the source tree. 

7# 

8# This program is free software: you can redistribute it and/or modify 

9# it under the terms of the GNU General Public License as published by 

10# the Free Software Foundation, either version 3 of the License, or 

11# (at your option) any later version. 

12# 

13# This program is distributed in the hope that it will be useful, 

14# but WITHOUT ANY WARRANTY; without even the implied warranty of 

15# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 

16# GNU General Public License for more details. 

17# 

18# You should have received a copy of the LSST License Statement and 

19# the GNU General Public License along with this program. If not, 

20# see <https://www.lsstcorp.org/LegalNotices/>. 

21# 

22import yaml 

23import re 

24from itertools import product 

25import os.path 

26 

27import pandas as pd 

28import numpy as np 

29import astropy.units as u 

30 

31from lsst.daf.persistence import doImport 

32from lsst.daf.butler import DeferredDatasetHandle 

33import lsst.geom as geom 

34import lsst.sphgeom as sphgeom 

35 

36from .parquetTable import ParquetTable, MultilevelParquetTable 

37 

38 

39def init_fromDict(initDict, basePath='lsst.pipe.tasks.functors', 

40 typeKey='functor', name=None): 

41 """Initialize an object defined in a dictionary 

42 

43 The object needs to be importable as 

44 f'{basePath}.{initDict[typeKey]}' 

45 The positional and keyword arguments (if any) are contained in 

46 "args" and "kwargs" entries in the dictionary, respectively. 

47 This is used in `functors.CompositeFunctor.from_yaml` to initialize 

48 a composite functor from a specification in a YAML file. 

49 

50 Parameters 

51 ---------- 

52 initDict : dictionary 

53 Dictionary describing object's initialization. Must contain 

54 an entry keyed by ``typeKey`` that is the name of the object, 

55 relative to ``basePath``. 

56 basePath : str 

57 Path relative to module in which ``initDict[typeKey]`` is defined. 

58 typeKey : str 

59 Key of ``initDict`` that is the name of the object 

60 (relative to `basePath`). 

61 """ 

62 initDict = initDict.copy() 

63 # TO DO: DM-21956 We should be able to define functors outside this module 

64 pythonType = doImport(f'{basePath}.{initDict.pop(typeKey)}') 

65 args = [] 

66 if 'args' in initDict: 

67 args = initDict.pop('args') 

68 if isinstance(args, str): 

69 args = [args] 

70 try: 

71 element = pythonType(*args, **initDict) 

72 except Exception as e: 

73 message = f'Error in constructing functor "{name}" of type {pythonType.__name__} with args: {args}' 

74 raise type(e)(message, e.args) 

75 return element 

76 

77 

78class Functor(object): 

79 """Define and execute a calculation on a ParquetTable 

80 

81 The `__call__` method accepts either a `ParquetTable` object or a 

82 `DeferredDatasetHandle`, and returns the 

83 result of the calculation as a single column. Each functor defines what 

84 columns are needed for the calculation, and only these columns are read 

85 from the `ParquetTable`. 

86 

87 The action of `__call__` consists of two steps: first, loading the 

88 necessary columns from disk into memory as a `pandas.DataFrame` object; 

89 and second, performing the computation on this dataframe and returning the 

90 result. 

91 

92 

93 To define a new `Functor`, a subclass must define a `_func` method, 

94 that takes a `pandas.DataFrame` and returns result in a `pandas.Series`. 

95 In addition, it must define the following attributes 

96 

97 * `_columns`: The columns necessary to perform the calculation 

98 * `name`: A name appropriate for a figure axis label 

99 * `shortname`: A name appropriate for use as a dictionary key 

100 

101 On initialization, a `Functor` should declare what band (`filt` kwarg) 

102 and dataset (e.g. `'ref'`, `'meas'`, `'forced_src'`) it is intended to be 

103 applied to. This enables the `_get_data` method to extract the proper 

104 columns from the parquet file. If not specified, the dataset will fall back 

105 on the `_defaultDataset`attribute. If band is not specified and `dataset` 

106 is anything other than `'ref'`, then an error will be raised when trying to 

107 perform the calculation. 

108 

109 Originally, `Functor` was set up to expect 

110 datasets formatted like the `deepCoadd_obj` dataset; that is, a 

111 dataframe with a multi-level column index, with the levels of the 

112 column index being `band`, `dataset`, and `column`. 

113 It has since been generalized to apply to dataframes without mutli-level 

114 indices and multi-level indices with just `dataset` and `column` levels. 

115 In addition, the `_get_data` method that reads 

116 the dataframe from the `ParquetTable` will return a dataframe with column 

117 index levels defined by the `_dfLevels` attribute; by default, this is 

118 `column`. 

119 

120 The `_dfLevels` attributes should generally not need to 

121 be changed, unless `_func` needs columns from multiple filters or datasets 

122 to do the calculation. 

123 An example of this is the `lsst.pipe.tasks.functors.Color` functor, for 

124 which `_dfLevels = ('band', 'column')`, and `_func` expects the dataframe 

125 it gets to have those levels in the column index. 

126 

127 Parameters 

128 ---------- 

129 filt : str 

130 Filter upon which to do the calculation 

131 

132 dataset : str 

133 Dataset upon which to do the calculation 

134 (e.g., 'ref', 'meas', 'forced_src'). 

135 

136 """ 

137 

138 _defaultDataset = 'ref' 

139 _dfLevels = ('column',) 

140 _defaultNoDup = False 

141 

142 def __init__(self, filt=None, dataset=None, noDup=None): 

143 self.filt = filt 

144 self.dataset = dataset if dataset is not None else self._defaultDataset 

145 self._noDup = noDup 

146 

147 @property 

148 def noDup(self): 

149 if self._noDup is not None: 

150 return self._noDup 

151 else: 

152 return self._defaultNoDup 

153 

154 @property 

155 def columns(self): 

156 """Columns required to perform calculation 

157 """ 

158 if not hasattr(self, '_columns'): 

159 raise NotImplementedError('Must define columns property or _columns attribute') 

160 return self._columns 

161 

162 def _get_data_columnLevels(self, data, columnIndex=None): 

163 """Gets the names of the column index levels 

164 

165 This should only be called in the context of a multilevel table. 

166 The logic here is to enable this to work both with the gen2 `MultilevelParquetTable` 

167 and with the gen3 `DeferredDatasetHandle`. 

168 

169 Parameters 

170 ---------- 

171 data : `MultilevelParquetTable` or `DeferredDatasetHandle` 

172 

173 columnnIndex (optional): pandas `Index` object 

174 if not passed, then it is read from the `DeferredDatasetHandle` 

175 """ 

176 if isinstance(data, DeferredDatasetHandle): 

177 if columnIndex is None: 

178 columnIndex = data.get(component="columns") 

179 if columnIndex is not None: 

180 return columnIndex.names 

181 if isinstance(data, MultilevelParquetTable): 

182 return data.columnLevels 

183 else: 

184 raise TypeError(f"Unknown type for data: {type(data)}!") 

185 

186 def _get_data_columnLevelNames(self, data, columnIndex=None): 

187 """Gets the content of each of the column levels for a multilevel table 

188 

189 Similar to `_get_data_columnLevels`, this enables backward compatibility with gen2. 

190 

191 Mirrors original gen2 implementation within `pipe.tasks.parquetTable.MultilevelParquetTable` 

192 """ 

193 if isinstance(data, DeferredDatasetHandle): 

194 if columnIndex is None: 

195 columnIndex = data.get(component="columns") 

196 if columnIndex is not None: 

197 columnLevels = columnIndex.names 

198 columnLevelNames = { 

199 level: list(np.unique(np.array([c for c in columnIndex])[:, i])) 

200 for i, level in enumerate(columnLevels) 

201 } 

202 return columnLevelNames 

203 if isinstance(data, MultilevelParquetTable): 

204 return data.columnLevelNames 

205 else: 

206 raise TypeError(f"Unknown type for data: {type(data)}!") 

207 

208 def _colsFromDict(self, colDict, columnIndex=None): 

209 """Converts dictionary column specficiation to a list of columns 

210 

211 This mirrors the original gen2 implementation within `pipe.tasks.parquetTable.MultilevelParquetTable` 

212 """ 

213 new_colDict = {} 

214 columnLevels = self._get_data_columnLevels(None, columnIndex=columnIndex) 

215 

216 for i, lev in enumerate(columnLevels): 

217 if lev in colDict: 

218 if isinstance(colDict[lev], str): 

219 new_colDict[lev] = [colDict[lev]] 

220 else: 

221 new_colDict[lev] = colDict[lev] 

222 else: 

223 new_colDict[lev] = columnIndex.levels[i] 

224 

225 levelCols = [new_colDict[lev] for lev in columnLevels] 

226 cols = product(*levelCols) 

227 return list(cols) 

228 

229 def multilevelColumns(self, data, columnIndex=None, returnTuple=False): 

230 """Returns columns needed by functor from multilevel dataset 

231 

232 To access tables with multilevel column structure, the `MultilevelParquetTable` 

233 or `DeferredDatasetHandle` need to be passed either a list of tuples or a 

234 dictionary. 

235 

236 Parameters 

237 ---------- 

238 data : `MultilevelParquetTable` or `DeferredDatasetHandle` 

239 

240 columnIndex (optional): pandas `Index` object 

241 either passed or read in from `DeferredDatasetHandle`. 

242 

243 `returnTuple` : bool 

244 If true, then return a list of tuples rather than the column dictionary 

245 specification. This is set to `True` by `CompositeFunctor` in order to be able to 

246 combine columns from the various component functors. 

247 

248 """ 

249 if isinstance(data, DeferredDatasetHandle) and columnIndex is None: 

250 columnIndex = data.get(component="columns") 

251 

252 # Confirm that the dataset has the column levels the functor is expecting it to have. 

253 columnLevels = self._get_data_columnLevels(data, columnIndex) 

254 

255 columnDict = {'column': self.columns, 

256 'dataset': self.dataset} 

257 if self.filt is None: 

258 columnLevelNames = self._get_data_columnLevelNames(data, columnIndex) 

259 if "band" in columnLevels: 

260 if self.dataset == "ref": 

261 columnDict["band"] = columnLevelNames["band"][0] 

262 else: 

263 raise ValueError(f"'filt' not set for functor {self.name}" 

264 f"(dataset {self.dataset}) " 

265 "and ParquetTable " 

266 "contains multiple filters in column index. " 

267 "Set 'filt' or set 'dataset' to 'ref'.") 

268 else: 

269 columnDict['band'] = self.filt 

270 

271 if isinstance(data, MultilevelParquetTable): 

272 return data._colsFromDict(columnDict) 

273 elif isinstance(data, DeferredDatasetHandle): 

274 if returnTuple: 

275 return self._colsFromDict(columnDict, columnIndex=columnIndex) 

276 else: 

277 return columnDict 

278 

279 def _func(self, df, dropna=True): 

280 raise NotImplementedError('Must define calculation on dataframe') 

281 

282 def _get_columnIndex(self, data): 

283 """Return columnIndex 

284 """ 

285 

286 if isinstance(data, DeferredDatasetHandle): 

287 return data.get(component="columns") 

288 else: 

289 return None 

290 

291 def _get_data(self, data): 

292 """Retrieve dataframe necessary for calculation. 

293 

294 The data argument can be a DataFrame, a ParquetTable instance, or a gen3 DeferredDatasetHandle 

295 

296 Returns dataframe upon which `self._func` can act. 

297 

298 N.B. while passing a raw pandas `DataFrame` *should* work here, it has not been tested. 

299 """ 

300 if isinstance(data, pd.DataFrame): 

301 return data 

302 

303 # First thing to do: check to see if the data source has a multilevel column index or not. 

304 columnIndex = self._get_columnIndex(data) 

305 is_multiLevel = isinstance(data, MultilevelParquetTable) or isinstance(columnIndex, pd.MultiIndex) 

306 

307 # Simple single-level parquet table, gen2 

308 if isinstance(data, ParquetTable) and not is_multiLevel: 

309 columns = self.columns 

310 df = data.toDataFrame(columns=columns) 

311 return df 

312 

313 # Get proper columns specification for this functor 

314 if is_multiLevel: 

315 columns = self.multilevelColumns(data, columnIndex=columnIndex) 

316 else: 

317 columns = self.columns 

318 

319 if isinstance(data, MultilevelParquetTable): 

320 # Load in-memory dataframe with appropriate columns the gen2 way 

321 df = data.toDataFrame(columns=columns, droplevels=False) 

322 elif isinstance(data, DeferredDatasetHandle): 

323 # Load in-memory dataframe with appropriate columns the gen3 way 

324 df = data.get(parameters={"columns": columns}) 

325 

326 # Drop unnecessary column levels 

327 if is_multiLevel: 

328 df = self._setLevels(df) 

329 

330 return df 

331 

332 def _setLevels(self, df): 

333 levelsToDrop = [n for n in df.columns.names if n not in self._dfLevels] 

334 df.columns = df.columns.droplevel(levelsToDrop) 

335 return df 

336 

337 def _dropna(self, vals): 

338 return vals.dropna() 

339 

340 def __call__(self, data, dropna=False): 

341 try: 

342 df = self._get_data(data) 

343 vals = self._func(df) 

344 except Exception: 

345 vals = self.fail(df) 

346 if dropna: 

347 vals = self._dropna(vals) 

348 

349 return vals 

350 

351 def difference(self, data1, data2, **kwargs): 

352 """Computes difference between functor called on two different ParquetTable objects 

353 """ 

354 return self(data1, **kwargs) - self(data2, **kwargs) 

355 

356 def fail(self, df): 

357 return pd.Series(np.full(len(df), np.nan), index=df.index) 

358 

359 @property 

360 def name(self): 

361 """Full name of functor (suitable for figure labels) 

362 """ 

363 return NotImplementedError 

364 

365 @property 

366 def shortname(self): 

367 """Short name of functor (suitable for column name/dict key) 

368 """ 

369 return self.name 

370 

371 

372class CompositeFunctor(Functor): 

373 """Perform multiple calculations at once on a catalog 

374 

375 The role of a `CompositeFunctor` is to group together computations from 

376 multiple functors. Instead of returning `pandas.Series` a 

377 `CompositeFunctor` returns a `pandas.Dataframe`, with the column names 

378 being the keys of `funcDict`. 

379 

380 The `columns` attribute of a `CompositeFunctor` is the union of all columns 

381 in all the component functors. 

382 

383 A `CompositeFunctor` does not use a `_func` method itself; rather, 

384 when a `CompositeFunctor` is called, all its columns are loaded 

385 at once, and the resulting dataframe is passed to the `_func` method of each component 

386 functor. This has the advantage of only doing I/O (reading from parquet file) once, 

387 and works because each individual `_func` method of each component functor does not 

388 care if there are *extra* columns in the dataframe being passed; only that it must contain 

389 *at least* the `columns` it expects. 

390 

391 An important and useful class method is `from_yaml`, which takes as argument the path to a YAML 

392 file specifying a collection of functors. 

393 

394 Parameters 

395 ---------- 

396 funcs : `dict` or `list` 

397 Dictionary or list of functors. If a list, then it will be converted 

398 into a dictonary according to the `.shortname` attribute of each functor. 

399 

400 """ 

401 dataset = None 

402 

403 def __init__(self, funcs, **kwargs): 

404 

405 if type(funcs) == dict: 

406 self.funcDict = funcs 

407 else: 

408 self.funcDict = {f.shortname: f for f in funcs} 

409 

410 self._filt = None 

411 

412 super().__init__(**kwargs) 

413 

414 @property 

415 def filt(self): 

416 return self._filt 

417 

418 @filt.setter 

419 def filt(self, filt): 

420 if filt is not None: 

421 for _, f in self.funcDict.items(): 

422 f.filt = filt 

423 self._filt = filt 

424 

425 def update(self, new): 

426 if isinstance(new, dict): 

427 self.funcDict.update(new) 

428 elif isinstance(new, CompositeFunctor): 

429 self.funcDict.update(new.funcDict) 

430 else: 

431 raise TypeError('Can only update with dictionary or CompositeFunctor.') 

432 

433 # Make sure new functors have the same 'filt' set 

434 if self.filt is not None: 

435 self.filt = self.filt 

436 

437 @property 

438 def columns(self): 

439 return list(set([x for y in [f.columns for f in self.funcDict.values()] for x in y])) 

440 

441 def multilevelColumns(self, data, **kwargs): 

442 # Get the union of columns for all component functors. Note the need to have `returnTuple=True` here. 

443 return list( 

444 set( 

445 [ 

446 x 

447 for y in [ 

448 f.multilevelColumns(data, returnTuple=True, **kwargs) for f in self.funcDict.values() 

449 ] 

450 for x in y 

451 ] 

452 ) 

453 ) 

454 

455 def __call__(self, data, **kwargs): 

456 """Apply the functor to the data table 

457 

458 Parameters 

459 ---------- 

460 data : `lsst.daf.butler.DeferredDatasetHandle`, 

461 `lsst.pipe.tasks.parquetTable.MultilevelParquetTable`, 

462 `lsst.pipe.tasks.parquetTable.ParquetTable`, 

463 or `pandas.DataFrame`. 

464 The table or a pointer to a table on disk from which columns can 

465 be accessed 

466 """ 

467 columnIndex = self._get_columnIndex(data) 

468 

469 # First, determine whether data has a multilevel index (either gen2 or gen3) 

470 is_multiLevel = isinstance(data, MultilevelParquetTable) or isinstance(columnIndex, pd.MultiIndex) 

471 

472 # Multilevel index, gen2 or gen3 

473 if is_multiLevel: 

474 columns = self.multilevelColumns(data, columnIndex=columnIndex) 

475 

476 if isinstance(data, MultilevelParquetTable): 

477 # Read data into memory the gen2 way 

478 df = data.toDataFrame(columns=columns, droplevels=False) 

479 elif isinstance(data, DeferredDatasetHandle): 

480 # Read data into memory the gen3 way 

481 df = data.get(parameters={"columns": columns}) 

482 

483 valDict = {} 

484 for k, f in self.funcDict.items(): 

485 try: 

486 subdf = f._setLevels( 

487 df[f.multilevelColumns(data, returnTuple=True, columnIndex=columnIndex)] 

488 ) 

489 valDict[k] = f._func(subdf) 

490 except Exception: 

491 valDict[k] = f.fail(subdf) 

492 

493 else: 

494 if isinstance(data, DeferredDatasetHandle): 

495 # input if Gen3 deferLoad=True 

496 df = data.get(parameters={"columns": self.columns}) 

497 elif isinstance(data, pd.DataFrame): 

498 # input if Gen3 deferLoad=False 

499 df = data 

500 else: 

501 # Original Gen2 input is type ParquetTable and the fallback 

502 df = data.toDataFrame(columns=self.columns) 

503 

504 valDict = {k: f._func(df) for k, f in self.funcDict.items()} 

505 

506 try: 

507 valDf = pd.concat(valDict, axis=1) 

508 except TypeError: 

509 print([(k, type(v)) for k, v in valDict.items()]) 

510 raise 

511 

512 if kwargs.get('dropna', False): 

513 valDf = valDf.dropna(how='any') 

514 

515 return valDf 

516 

517 @classmethod 

518 def renameCol(cls, col, renameRules): 

519 if renameRules is None: 

520 return col 

521 for old, new in renameRules: 

522 if col.startswith(old): 

523 col = col.replace(old, new) 

524 return col 

525 

526 @classmethod 

527 def from_file(cls, filename, **kwargs): 

528 # Allow environment variables in the filename. 

529 filename = os.path.expandvars(filename) 

530 with open(filename) as f: 

531 translationDefinition = yaml.safe_load(f) 

532 

533 return cls.from_yaml(translationDefinition, **kwargs) 

534 

535 @classmethod 

536 def from_yaml(cls, translationDefinition, **kwargs): 

537 funcs = {} 

538 for func, val in translationDefinition['funcs'].items(): 

539 funcs[func] = init_fromDict(val, name=func) 

540 

541 if 'flag_rename_rules' in translationDefinition: 

542 renameRules = translationDefinition['flag_rename_rules'] 

543 else: 

544 renameRules = None 

545 

546 if 'calexpFlags' in translationDefinition: 

547 for flag in translationDefinition['calexpFlags']: 

548 funcs[cls.renameCol(flag, renameRules)] = Column(flag, dataset='calexp') 

549 

550 if 'refFlags' in translationDefinition: 

551 for flag in translationDefinition['refFlags']: 

552 funcs[cls.renameCol(flag, renameRules)] = Column(flag, dataset='ref') 

553 

554 if 'forcedFlags' in translationDefinition: 

555 for flag in translationDefinition['forcedFlags']: 

556 funcs[cls.renameCol(flag, renameRules)] = Column(flag, dataset='forced_src') 

557 

558 if 'flags' in translationDefinition: 

559 for flag in translationDefinition['flags']: 

560 funcs[cls.renameCol(flag, renameRules)] = Column(flag, dataset='meas') 

561 

562 return cls(funcs, **kwargs) 

563 

564 

565def mag_aware_eval(df, expr): 

566 """Evaluate an expression on a DataFrame, knowing what the 'mag' function means 

567 

568 Builds on `pandas.DataFrame.eval`, which parses and executes math on dataframes. 

569 

570 Parameters 

571 ---------- 

572 df : pandas.DataFrame 

573 Dataframe on which to evaluate expression. 

574 

575 expr : str 

576 Expression. 

577 """ 

578 try: 

579 expr_new = re.sub(r'mag\((\w+)\)', r'-2.5*log(\g<1>)/log(10)', expr) 

580 val = df.eval(expr_new, truediv=True) 

581 except Exception: # Should check what actually gets raised 

582 expr_new = re.sub(r'mag\((\w+)\)', r'-2.5*log(\g<1>_instFlux)/log(10)', expr) 

583 val = df.eval(expr_new, truediv=True) 

584 return val 

585 

586 

587class CustomFunctor(Functor): 

588 """Arbitrary computation on a catalog 

589 

590 Column names (and thus the columns to be loaded from catalog) are found 

591 by finding all words and trying to ignore all "math-y" words. 

592 

593 Parameters 

594 ---------- 

595 expr : str 

596 Expression to evaluate, to be parsed and executed by `mag_aware_eval`. 

597 """ 

598 _ignore_words = ('mag', 'sin', 'cos', 'exp', 'log', 'sqrt') 

599 

600 def __init__(self, expr, **kwargs): 

601 self.expr = expr 

602 super().__init__(**kwargs) 

603 

604 @property 

605 def name(self): 

606 return self.expr 

607 

608 @property 

609 def columns(self): 

610 flux_cols = re.findall(r'mag\(\s*(\w+)\s*\)', self.expr) 

611 

612 cols = [c for c in re.findall(r'[a-zA-Z_]+', self.expr) if c not in self._ignore_words] 

613 not_a_col = [] 

614 for c in flux_cols: 

615 if not re.search('_instFlux$', c): 

616 cols.append(f'{c}_instFlux') 

617 not_a_col.append(c) 

618 else: 

619 cols.append(c) 

620 

621 return list(set([c for c in cols if c not in not_a_col])) 

622 

623 def _func(self, df): 

624 return mag_aware_eval(df, self.expr) 

625 

626 

627class Column(Functor): 

628 """Get column with specified name 

629 """ 

630 

631 def __init__(self, col, **kwargs): 

632 self.col = col 

633 super().__init__(**kwargs) 

634 

635 @property 

636 def name(self): 

637 return self.col 

638 

639 @property 

640 def columns(self): 

641 return [self.col] 

642 

643 def _func(self, df): 

644 return df[self.col] 

645 

646 

647class Index(Functor): 

648 """Return the value of the index for each object 

649 """ 

650 

651 columns = ['coord_ra'] # just a dummy; something has to be here 

652 _defaultDataset = 'ref' 

653 _defaultNoDup = True 

654 

655 def _func(self, df): 

656 return pd.Series(df.index, index=df.index) 

657 

658 

659class IDColumn(Column): 

660 col = 'id' 

661 _allow_difference = False 

662 _defaultNoDup = True 

663 

664 def _func(self, df): 

665 return pd.Series(df.index, index=df.index) 

666 

667 

668class FootprintNPix(Column): 

669 col = 'base_Footprint_nPix' 

670 

671 

672class CoordColumn(Column): 

673 """Base class for coordinate column, in degrees 

674 """ 

675 _radians = True 

676 

677 def __init__(self, col, **kwargs): 

678 super().__init__(col, **kwargs) 

679 

680 def _func(self, df): 

681 # Must not modify original column in case that column is used by another functor 

682 output = df[self.col] * 180 / np.pi if self._radians else df[self.col] 

683 return output 

684 

685 

686class RAColumn(CoordColumn): 

687 """Right Ascension, in degrees 

688 """ 

689 name = 'RA' 

690 _defaultNoDup = True 

691 

692 def __init__(self, **kwargs): 

693 super().__init__('coord_ra', **kwargs) 

694 

695 def __call__(self, catalog, **kwargs): 

696 return super().__call__(catalog, **kwargs) 

697 

698 

699class DecColumn(CoordColumn): 

700 """Declination, in degrees 

701 """ 

702 name = 'Dec' 

703 _defaultNoDup = True 

704 

705 def __init__(self, **kwargs): 

706 super().__init__('coord_dec', **kwargs) 

707 

708 def __call__(self, catalog, **kwargs): 

709 return super().__call__(catalog, **kwargs) 

710 

711 

712class HtmIndex20(Functor): 

713 """Compute the level 20 HtmIndex for the catalog. 

714 """ 

715 name = "Htm20" 

716 htmLevel = 20 

717 _radians = True 

718 

719 def __init__(self, ra, decl, **kwargs): 

720 self.pixelator = sphgeom.HtmPixelization(self.htmLevel) 

721 self.ra = ra 

722 self.decl = decl 

723 self._columns = [self.ra, self.decl] 

724 super().__init__(**kwargs) 

725 

726 def _func(self, df): 

727 

728 def computePixel(row): 

729 if self._radians: 

730 sphPoint = geom.SpherePoint(row[self.ra], 

731 row[self.decl], 

732 geom.radians) 

733 else: 

734 sphPoint = geom.SpherePoint(row[self.ra], 

735 row[self.decl], 

736 geom.degrees) 

737 return self.pixelator.index(sphPoint.getVector()) 

738 

739 return df.apply(computePixel, axis=1) 

740 

741 

742def fluxName(col): 

743 if not col.endswith('_instFlux'): 

744 col += '_instFlux' 

745 return col 

746 

747 

748def fluxErrName(col): 

749 if not col.endswith('_instFluxErr'): 

750 col += '_instFluxErr' 

751 return col 

752 

753 

754class Mag(Functor): 

755 """Compute calibrated magnitude 

756 

757 Takes a `calib` argument, which returns the flux at mag=0 

758 as `calib.getFluxMag0()`. If not provided, then the default 

759 `fluxMag0` is 63095734448.0194, which is default for HSC. 

760 This default should be removed in DM-21955 

761 

762 This calculation hides warnings about invalid values and dividing by zero. 

763 

764 As for all functors, a `dataset` and `filt` kwarg should be provided upon 

765 initialization. Unlike the default `Functor`, however, the default dataset 

766 for a `Mag` is `'meas'`, rather than `'ref'`. 

767 

768 Parameters 

769 ---------- 

770 col : `str` 

771 Name of flux column from which to compute magnitude. Can be parseable 

772 by `lsst.pipe.tasks.functors.fluxName` function---that is, you can pass 

773 `'modelfit_CModel'` instead of `'modelfit_CModel_instFlux'`) and it will 

774 understand. 

775 calib : `lsst.afw.image.calib.Calib` (optional) 

776 Object that knows zero point. 

777 """ 

778 _defaultDataset = 'meas' 

779 

780 def __init__(self, col, calib=None, **kwargs): 

781 self.col = fluxName(col) 

782 self.calib = calib 

783 if calib is not None: 

784 self.fluxMag0 = calib.getFluxMag0()[0] 

785 else: 

786 # TO DO: DM-21955 Replace hard coded photometic calibration values 

787 self.fluxMag0 = 63095734448.0194 

788 

789 super().__init__(**kwargs) 

790 

791 @property 

792 def columns(self): 

793 return [self.col] 

794 

795 def _func(self, df): 

796 with np.warnings.catch_warnings(): 

797 np.warnings.filterwarnings('ignore', r'invalid value encountered') 

798 np.warnings.filterwarnings('ignore', r'divide by zero') 

799 return -2.5*np.log10(df[self.col] / self.fluxMag0) 

800 

801 @property 

802 def name(self): 

803 return f'mag_{self.col}' 

804 

805 

806class MagErr(Mag): 

807 """Compute calibrated magnitude uncertainty 

808 

809 Takes the same `calib` object as `lsst.pipe.tasks.functors.Mag`. 

810 

811 Parameters 

812 col : `str` 

813 Name of flux column 

814 calib : `lsst.afw.image.calib.Calib` (optional) 

815 Object that knows zero point. 

816 """ 

817 

818 def __init__(self, *args, **kwargs): 

819 super().__init__(*args, **kwargs) 

820 if self.calib is not None: 

821 self.fluxMag0Err = self.calib.getFluxMag0()[1] 

822 else: 

823 self.fluxMag0Err = 0. 

824 

825 @property 

826 def columns(self): 

827 return [self.col, self.col + 'Err'] 

828 

829 def _func(self, df): 

830 with np.warnings.catch_warnings(): 

831 np.warnings.filterwarnings('ignore', r'invalid value encountered') 

832 np.warnings.filterwarnings('ignore', r'divide by zero') 

833 fluxCol, fluxErrCol = self.columns 

834 x = df[fluxErrCol] / df[fluxCol] 

835 y = self.fluxMag0Err / self.fluxMag0 

836 magErr = (2.5 / np.log(10.)) * np.sqrt(x*x + y*y) 

837 return magErr 

838 

839 @property 

840 def name(self): 

841 return super().name + '_err' 

842 

843 

844class NanoMaggie(Mag): 

845 """ 

846 """ 

847 

848 def _func(self, df): 

849 return (df[self.col] / self.fluxMag0) * 1e9 

850 

851 

852class MagDiff(Functor): 

853 _defaultDataset = 'meas' 

854 

855 """Functor to calculate magnitude difference""" 

856 

857 def __init__(self, col1, col2, **kwargs): 

858 self.col1 = fluxName(col1) 

859 self.col2 = fluxName(col2) 

860 super().__init__(**kwargs) 

861 

862 @property 

863 def columns(self): 

864 return [self.col1, self.col2] 

865 

866 def _func(self, df): 

867 with np.warnings.catch_warnings(): 

868 np.warnings.filterwarnings('ignore', r'invalid value encountered') 

869 np.warnings.filterwarnings('ignore', r'divide by zero') 

870 return -2.5*np.log10(df[self.col1]/df[self.col2]) 

871 

872 @property 

873 def name(self): 

874 return f'(mag_{self.col1} - mag_{self.col2})' 

875 

876 @property 

877 def shortname(self): 

878 return f'magDiff_{self.col1}_{self.col2}' 

879 

880 

881class Color(Functor): 

882 """Compute the color between two filters 

883 

884 Computes color by initializing two different `Mag` 

885 functors based on the `col` and filters provided, and 

886 then returning the difference. 

887 

888 This is enabled by the `_func` expecting a dataframe with a 

889 multilevel column index, with both `'band'` and `'column'`, 

890 instead of just `'column'`, which is the `Functor` default. 

891 This is controlled by the `_dfLevels` attribute. 

892 

893 Also of note, the default dataset for `Color` is `forced_src'`, 

894 whereas for `Mag` it is `'meas'`. 

895 

896 Parameters 

897 ---------- 

898 col : str 

899 Name of flux column from which to compute; same as would be passed to 

900 `lsst.pipe.tasks.functors.Mag`. 

901 

902 filt2, filt1 : str 

903 Filters from which to compute magnitude difference. 

904 Color computed is `Mag(filt2) - Mag(filt1)`. 

905 """ 

906 _defaultDataset = 'forced_src' 

907 _dfLevels = ('band', 'column') 

908 _defaultNoDup = True 

909 

910 def __init__(self, col, filt2, filt1, **kwargs): 

911 self.col = fluxName(col) 

912 if filt2 == filt1: 

913 raise RuntimeError("Cannot compute Color for %s: %s - %s " % (col, filt2, filt1)) 

914 self.filt2 = filt2 

915 self.filt1 = filt1 

916 

917 self.mag2 = Mag(col, filt=filt2, **kwargs) 

918 self.mag1 = Mag(col, filt=filt1, **kwargs) 

919 

920 super().__init__(**kwargs) 

921 

922 @property 

923 def filt(self): 

924 return None 

925 

926 @filt.setter 

927 def filt(self, filt): 

928 pass 

929 

930 def _func(self, df): 

931 mag2 = self.mag2._func(df[self.filt2]) 

932 mag1 = self.mag1._func(df[self.filt1]) 

933 return mag2 - mag1 

934 

935 @property 

936 def columns(self): 

937 return [self.mag1.col, self.mag2.col] 

938 

939 def multilevelColumns(self, parq, **kwargs): 

940 return [(self.dataset, self.filt1, self.col), (self.dataset, self.filt2, self.col)] 

941 

942 @property 

943 def name(self): 

944 return f'{self.filt2} - {self.filt1} ({self.col})' 

945 

946 @property 

947 def shortname(self): 

948 return f"{self.col}_{self.filt2.replace('-', '')}m{self.filt1.replace('-', '')}" 

949 

950 

951class Labeller(Functor): 

952 """Main function of this subclass is to override the dropna=True 

953 """ 

954 _null_label = 'null' 

955 _allow_difference = False 

956 name = 'label' 

957 _force_str = False 

958 

959 def __call__(self, parq, dropna=False, **kwargs): 

960 return super().__call__(parq, dropna=False, **kwargs) 

961 

962 

963class StarGalaxyLabeller(Labeller): 

964 _columns = ["base_ClassificationExtendedness_value"] 

965 _column = "base_ClassificationExtendedness_value" 

966 

967 def _func(self, df): 

968 x = df[self._columns][self._column] 

969 mask = x.isnull() 

970 test = (x < 0.5).astype(int) 

971 test = test.mask(mask, 2) 

972 

973 # TODO: DM-21954 Look into veracity of inline comment below 

974 # are these backwards? 

975 categories = ['galaxy', 'star', self._null_label] 

976 label = pd.Series(pd.Categorical.from_codes(test, categories=categories), 

977 index=x.index, name='label') 

978 if self._force_str: 

979 label = label.astype(str) 

980 return label 

981 

982 

983class NumStarLabeller(Labeller): 

984 _columns = ['numStarFlags'] 

985 labels = {"star": 0, "maybe": 1, "notStar": 2} 

986 

987 def _func(self, df): 

988 x = df[self._columns][self._columns[0]] 

989 

990 # Number of filters 

991 n = len(x.unique()) - 1 

992 

993 labels = ['noStar', 'maybe', 'star'] 

994 label = pd.Series(pd.cut(x, [-1, 0, n-1, n], labels=labels), 

995 index=x.index, name='label') 

996 

997 if self._force_str: 

998 label = label.astype(str) 

999 

1000 return label 

1001 

1002 

1003class DeconvolvedMoments(Functor): 

1004 name = 'Deconvolved Moments' 

1005 shortname = 'deconvolvedMoments' 

1006 _columns = ("ext_shapeHSM_HsmSourceMoments_xx", 

1007 "ext_shapeHSM_HsmSourceMoments_yy", 

1008 "base_SdssShape_xx", "base_SdssShape_yy", 

1009 "ext_shapeHSM_HsmPsfMoments_xx", 

1010 "ext_shapeHSM_HsmPsfMoments_yy") 

1011 

1012 def _func(self, df): 

1013 """Calculate deconvolved moments""" 

1014 if "ext_shapeHSM_HsmSourceMoments_xx" in df.columns: # _xx added by tdm 

1015 hsm = df["ext_shapeHSM_HsmSourceMoments_xx"] + df["ext_shapeHSM_HsmSourceMoments_yy"] 

1016 else: 

1017 hsm = np.ones(len(df))*np.nan 

1018 sdss = df["base_SdssShape_xx"] + df["base_SdssShape_yy"] 

1019 if "ext_shapeHSM_HsmPsfMoments_xx" in df.columns: 

1020 psf = df["ext_shapeHSM_HsmPsfMoments_xx"] + df["ext_shapeHSM_HsmPsfMoments_yy"] 

1021 else: 

1022 # LSST does not have shape.sdss.psf. Could instead add base_PsfShape to catalog using 

1023 # exposure.getPsf().computeShape(s.getCentroid()).getIxx() 

1024 # raise TaskError("No psf shape parameter found in catalog") 

1025 raise RuntimeError('No psf shape parameter found in catalog') 

1026 

1027 return hsm.where(np.isfinite(hsm), sdss) - psf 

1028 

1029 

1030class SdssTraceSize(Functor): 

1031 """Functor to calculate SDSS trace radius size for sources""" 

1032 name = "SDSS Trace Size" 

1033 shortname = 'sdssTrace' 

1034 _columns = ("base_SdssShape_xx", "base_SdssShape_yy") 

1035 

1036 def _func(self, df): 

1037 srcSize = np.sqrt(0.5*(df["base_SdssShape_xx"] + df["base_SdssShape_yy"])) 

1038 return srcSize 

1039 

1040 

1041class PsfSdssTraceSizeDiff(Functor): 

1042 """Functor to calculate SDSS trace radius size difference (%) between object and psf model""" 

1043 name = "PSF - SDSS Trace Size" 

1044 shortname = 'psf_sdssTrace' 

1045 _columns = ("base_SdssShape_xx", "base_SdssShape_yy", 

1046 "base_SdssShape_psf_xx", "base_SdssShape_psf_yy") 

1047 

1048 def _func(self, df): 

1049 srcSize = np.sqrt(0.5*(df["base_SdssShape_xx"] + df["base_SdssShape_yy"])) 

1050 psfSize = np.sqrt(0.5*(df["base_SdssShape_psf_xx"] + df["base_SdssShape_psf_yy"])) 

1051 sizeDiff = 100*(srcSize - psfSize)/(0.5*(srcSize + psfSize)) 

1052 return sizeDiff 

1053 

1054 

1055class HsmTraceSize(Functor): 

1056 """Functor to calculate HSM trace radius size for sources""" 

1057 name = 'HSM Trace Size' 

1058 shortname = 'hsmTrace' 

1059 _columns = ("ext_shapeHSM_HsmSourceMoments_xx", 

1060 "ext_shapeHSM_HsmSourceMoments_yy") 

1061 

1062 def _func(self, df): 

1063 srcSize = np.sqrt(0.5*(df["ext_shapeHSM_HsmSourceMoments_xx"] 

1064 + df["ext_shapeHSM_HsmSourceMoments_yy"])) 

1065 return srcSize 

1066 

1067 

1068class PsfHsmTraceSizeDiff(Functor): 

1069 """Functor to calculate HSM trace radius size difference (%) between object and psf model""" 

1070 name = 'PSF - HSM Trace Size' 

1071 shortname = 'psf_HsmTrace' 

1072 _columns = ("ext_shapeHSM_HsmSourceMoments_xx", 

1073 "ext_shapeHSM_HsmSourceMoments_yy", 

1074 "ext_shapeHSM_HsmPsfMoments_xx", 

1075 "ext_shapeHSM_HsmPsfMoments_yy") 

1076 

1077 def _func(self, df): 

1078 srcSize = np.sqrt(0.5*(df["ext_shapeHSM_HsmSourceMoments_xx"] 

1079 + df["ext_shapeHSM_HsmSourceMoments_yy"])) 

1080 psfSize = np.sqrt(0.5*(df["ext_shapeHSM_HsmPsfMoments_xx"] 

1081 + df["ext_shapeHSM_HsmPsfMoments_yy"])) 

1082 sizeDiff = 100*(srcSize - psfSize)/(0.5*(srcSize + psfSize)) 

1083 return sizeDiff 

1084 

1085 

1086class HsmFwhm(Functor): 

1087 name = 'HSM Psf FWHM' 

1088 _columns = ('ext_shapeHSM_HsmPsfMoments_xx', 'ext_shapeHSM_HsmPsfMoments_yy') 

1089 # TODO: DM-21403 pixel scale should be computed from the CD matrix or transform matrix 

1090 pixelScale = 0.168 

1091 SIGMA2FWHM = 2*np.sqrt(2*np.log(2)) 

1092 

1093 def _func(self, df): 

1094 return self.pixelScale*self.SIGMA2FWHM*np.sqrt( 

1095 0.5*(df['ext_shapeHSM_HsmPsfMoments_xx'] + df['ext_shapeHSM_HsmPsfMoments_yy'])) 

1096 

1097 

1098class E1(Functor): 

1099 name = "Distortion Ellipticity (e1)" 

1100 shortname = "Distortion" 

1101 

1102 def __init__(self, colXX, colXY, colYY, **kwargs): 

1103 self.colXX = colXX 

1104 self.colXY = colXY 

1105 self.colYY = colYY 

1106 self._columns = [self.colXX, self.colXY, self.colYY] 

1107 super().__init__(**kwargs) 

1108 

1109 @property 

1110 def columns(self): 

1111 return [self.colXX, self.colXY, self.colYY] 

1112 

1113 def _func(self, df): 

1114 return df[self.colXX] - df[self.colYY] / (df[self.colXX] + df[self.colYY]) 

1115 

1116 

1117class E2(Functor): 

1118 name = "Ellipticity e2" 

1119 

1120 def __init__(self, colXX, colXY, colYY, **kwargs): 

1121 self.colXX = colXX 

1122 self.colXY = colXY 

1123 self.colYY = colYY 

1124 super().__init__(**kwargs) 

1125 

1126 @property 

1127 def columns(self): 

1128 return [self.colXX, self.colXY, self.colYY] 

1129 

1130 def _func(self, df): 

1131 return 2*df[self.colXY] / (df[self.colXX] + df[self.colYY]) 

1132 

1133 

1134class RadiusFromQuadrupole(Functor): 

1135 

1136 def __init__(self, colXX, colXY, colYY, **kwargs): 

1137 self.colXX = colXX 

1138 self.colXY = colXY 

1139 self.colYY = colYY 

1140 super().__init__(**kwargs) 

1141 

1142 @property 

1143 def columns(self): 

1144 return [self.colXX, self.colXY, self.colYY] 

1145 

1146 def _func(self, df): 

1147 return (df[self.colXX]*df[self.colYY] - df[self.colXY]**2)**0.25 

1148 

1149 

1150class LocalWcs(Functor): 

1151 """Computations using the stored localWcs. 

1152 """ 

1153 name = "LocalWcsOperations" 

1154 

1155 def __init__(self, 

1156 colCD_1_1, 

1157 colCD_1_2, 

1158 colCD_2_1, 

1159 colCD_2_2, 

1160 **kwargs): 

1161 self.colCD_1_1 = colCD_1_1 

1162 self.colCD_1_2 = colCD_1_2 

1163 self.colCD_2_1 = colCD_2_1 

1164 self.colCD_2_2 = colCD_2_2 

1165 super().__init__(**kwargs) 

1166 

1167 def computeDeltaRaDec(self, x, y, cd11, cd12, cd21, cd22): 

1168 """Compute the distance on the sphere from x2, y1 to x1, y1. 

1169 

1170 Parameters 

1171 ---------- 

1172 x : `pandas.Series` 

1173 X pixel coordinate. 

1174 y : `pandas.Series` 

1175 Y pixel coordinate. 

1176 cd11 : `pandas.Series` 

1177 [1, 1] element of the local Wcs affine transform. 

1178 cd11 : `pandas.Series` 

1179 [1, 1] element of the local Wcs affine transform. 

1180 cd12 : `pandas.Series` 

1181 [1, 2] element of the local Wcs affine transform. 

1182 cd21 : `pandas.Series` 

1183 [2, 1] element of the local Wcs affine transform. 

1184 cd22 : `pandas.Series` 

1185 [2, 2] element of the local Wcs affine transform. 

1186 

1187 Returns 

1188 ------- 

1189 raDecTuple : tuple 

1190 RA and dec conversion of x and y given the local Wcs. Returned 

1191 units are in radians. 

1192 

1193 """ 

1194 return (x * cd11 + y * cd12, x * cd21 + y * cd22) 

1195 

1196 def computeSkySeperation(self, ra1, dec1, ra2, dec2): 

1197 """Compute the local pixel scale conversion. 

1198 

1199 Parameters 

1200 ---------- 

1201 ra1 : `pandas.Series` 

1202 Ra of the first coordinate in radians. 

1203 dec1 : `pandas.Series` 

1204 Dec of the first coordinate in radians. 

1205 ra2 : `pandas.Series` 

1206 Ra of the second coordinate in radians. 

1207 dec2 : `pandas.Series` 

1208 Dec of the second coordinate in radians. 

1209 

1210 Returns 

1211 ------- 

1212 dist : `pandas.Series` 

1213 Distance on the sphere in radians. 

1214 """ 

1215 deltaDec = dec2 - dec1 

1216 deltaRa = ra2 - ra1 

1217 return 2 * np.arcsin( 

1218 np.sqrt( 

1219 np.sin(deltaDec / 2) ** 2 

1220 + np.cos(dec2) * np.cos(dec1) * np.sin(deltaRa / 2) ** 2)) 

1221 

1222 def getSkySeperationFromPixel(self, x1, y1, x2, y2, cd11, cd12, cd21, cd22): 

1223 """Compute the distance on the sphere from x2, y1 to x1, y1. 

1224 

1225 Parameters 

1226 ---------- 

1227 x1 : `pandas.Series` 

1228 X pixel coordinate. 

1229 y1 : `pandas.Series` 

1230 Y pixel coordinate. 

1231 x2 : `pandas.Series` 

1232 X pixel coordinate. 

1233 y2 : `pandas.Series` 

1234 Y pixel coordinate. 

1235 cd11 : `pandas.Series` 

1236 [1, 1] element of the local Wcs affine transform. 

1237 cd11 : `pandas.Series` 

1238 [1, 1] element of the local Wcs affine transform. 

1239 cd12 : `pandas.Series` 

1240 [1, 2] element of the local Wcs affine transform. 

1241 cd21 : `pandas.Series` 

1242 [2, 1] element of the local Wcs affine transform. 

1243 cd22 : `pandas.Series` 

1244 [2, 2] element of the local Wcs affine transform. 

1245 

1246 Returns 

1247 ------- 

1248 Distance : `pandas.Series` 

1249 Arcseconds per pixel at the location of the local WC 

1250 """ 

1251 ra1, dec1 = self.computeDeltaRaDec(x1, y1, cd11, cd12, cd21, cd22) 

1252 ra2, dec2 = self.computeDeltaRaDec(x2, y2, cd11, cd12, cd21, cd22) 

1253 # Great circle distance for small separations. 

1254 return self.computeSkySeperation(ra1, dec1, ra2, dec2) 

1255 

1256 

1257class ComputePixelScale(LocalWcs): 

1258 """Compute the local pixel scale from the stored CDMatrix. 

1259 """ 

1260 name = "PixelScale" 

1261 

1262 @property 

1263 def columns(self): 

1264 return [self.colCD_1_1, 

1265 self.colCD_1_2, 

1266 self.colCD_2_1, 

1267 self.colCD_2_2] 

1268 

1269 def pixelScaleArcseconds(self, cd11, cd12, cd21, cd22): 

1270 """Compute the local pixel to scale conversion in arcseconds. 

1271 

1272 Parameters 

1273 ---------- 

1274 cd11 : `pandas.Series` 

1275 [1, 1] element of the local Wcs affine transform in radians. 

1276 cd11 : `pandas.Series` 

1277 [1, 1] element of the local Wcs affine transform in radians. 

1278 cd12 : `pandas.Series` 

1279 [1, 2] element of the local Wcs affine transform in radians. 

1280 cd21 : `pandas.Series` 

1281 [2, 1] element of the local Wcs affine transform in radians. 

1282 cd22 : `pandas.Series` 

1283 [2, 2] element of the local Wcs affine transform in radians. 

1284 

1285 Returns 

1286 ------- 

1287 pixScale : `pandas.Series` 

1288 Arcseconds per pixel at the location of the local WC 

1289 """ 

1290 return 3600 * np.degrees(np.sqrt(np.fabs(cd11 * cd22 - cd12 * cd21))) 

1291 

1292 def _func(self, df): 

1293 return self.pixelScaleArcseconds(df[self.colCD_1_1], 

1294 df[self.colCD_1_2], 

1295 df[self.colCD_2_1], 

1296 df[self.colCD_2_2]) 

1297 

1298 

1299class ConvertPixelToArcseconds(ComputePixelScale): 

1300 """Convert a value in units pixels squared to units arcseconds squared. 

1301 """ 

1302 

1303 def __init__(self, 

1304 col, 

1305 colCD_1_1, 

1306 colCD_1_2, 

1307 colCD_2_1, 

1308 colCD_2_2, 

1309 **kwargs): 

1310 self.col = col 

1311 super().__init__(colCD_1_1, 

1312 colCD_1_2, 

1313 colCD_2_1, 

1314 colCD_2_2, 

1315 **kwargs) 

1316 

1317 @property 

1318 def name(self): 

1319 return f"{self.col}_asArcseconds" 

1320 

1321 @property 

1322 def columns(self): 

1323 return [self.col, 

1324 self.colCD_1_1, 

1325 self.colCD_1_2, 

1326 self.colCD_2_1, 

1327 self.colCD_2_2] 

1328 

1329 def _func(self, df): 

1330 return df[self.col] * self.pixelScaleArcseconds(df[self.colCD_1_1], 

1331 df[self.colCD_1_2], 

1332 df[self.colCD_2_1], 

1333 df[self.colCD_2_2]) 

1334 

1335 

1336class ConvertPixelSqToArcsecondsSq(ComputePixelScale): 

1337 """Convert a value in units pixels to units arcseconds. 

1338 """ 

1339 

1340 def __init__(self, 

1341 col, 

1342 colCD_1_1, 

1343 colCD_1_2, 

1344 colCD_2_1, 

1345 colCD_2_2, 

1346 **kwargs): 

1347 self.col = col 

1348 super().__init__(colCD_1_1, 

1349 colCD_1_2, 

1350 colCD_2_1, 

1351 colCD_2_2, 

1352 **kwargs) 

1353 

1354 @property 

1355 def name(self): 

1356 return f"{self.col}_asArcsecondsSq" 

1357 

1358 @property 

1359 def columns(self): 

1360 return [self.col, 

1361 self.colCD_1_1, 

1362 self.colCD_1_2, 

1363 self.colCD_2_1, 

1364 self.colCD_2_2] 

1365 

1366 def _func(self, df): 

1367 pixScale = self.pixelScaleArcseconds(df[self.colCD_1_1], 

1368 df[self.colCD_1_2], 

1369 df[self.colCD_2_1], 

1370 df[self.colCD_2_2]) 

1371 return df[self.col] * pixScale * pixScale 

1372 

1373 

1374class ReferenceBand(Functor): 

1375 name = 'Reference Band' 

1376 shortname = 'refBand' 

1377 

1378 @property 

1379 def columns(self): 

1380 return ["merge_measurement_i", 

1381 "merge_measurement_r", 

1382 "merge_measurement_z", 

1383 "merge_measurement_y", 

1384 "merge_measurement_g"] 

1385 

1386 def _func(self, df): 

1387 def getFilterAliasName(row): 

1388 # get column name with the max value (True > False) 

1389 colName = row.idxmax() 

1390 return colName.replace('merge_measurement_', '') 

1391 

1392 return df[self.columns].apply(getFilterAliasName, axis=1) 

1393 

1394 

1395class Photometry(Functor): 

1396 # AB to NanoJansky (3631 Jansky) 

1397 AB_FLUX_SCALE = (0 * u.ABmag).to_value(u.nJy) 

1398 LOG_AB_FLUX_SCALE = 12.56 

1399 FIVE_OVER_2LOG10 = 1.085736204758129569 

1400 # TO DO: DM-21955 Replace hard coded photometic calibration values 

1401 COADD_ZP = 27 

1402 

1403 def __init__(self, colFlux, colFluxErr=None, calib=None, **kwargs): 

1404 self.vhypot = np.vectorize(self.hypot) 

1405 self.col = colFlux 

1406 self.colFluxErr = colFluxErr 

1407 

1408 self.calib = calib 

1409 if calib is not None: 

1410 self.fluxMag0, self.fluxMag0Err = calib.getFluxMag0() 

1411 else: 

1412 self.fluxMag0 = 1./np.power(10, -0.4*self.COADD_ZP) 

1413 self.fluxMag0Err = 0. 

1414 

1415 super().__init__(**kwargs) 

1416 

1417 @property 

1418 def columns(self): 

1419 return [self.col] 

1420 

1421 @property 

1422 def name(self): 

1423 return f'mag_{self.col}' 

1424 

1425 @classmethod 

1426 def hypot(cls, a, b): 

1427 if np.abs(a) < np.abs(b): 

1428 a, b = b, a 

1429 if a == 0.: 

1430 return 0. 

1431 q = b/a 

1432 return np.abs(a) * np.sqrt(1. + q*q) 

1433 

1434 def dn2flux(self, dn, fluxMag0): 

1435 return self.AB_FLUX_SCALE * dn / fluxMag0 

1436 

1437 def dn2mag(self, dn, fluxMag0): 

1438 with np.warnings.catch_warnings(): 

1439 np.warnings.filterwarnings('ignore', r'invalid value encountered') 

1440 np.warnings.filterwarnings('ignore', r'divide by zero') 

1441 return -2.5 * np.log10(dn/fluxMag0) 

1442 

1443 def dn2fluxErr(self, dn, dnErr, fluxMag0, fluxMag0Err): 

1444 retVal = self.vhypot(dn * fluxMag0Err, dnErr * fluxMag0) 

1445 retVal *= self.AB_FLUX_SCALE / fluxMag0 / fluxMag0 

1446 return retVal 

1447 

1448 def dn2MagErr(self, dn, dnErr, fluxMag0, fluxMag0Err): 

1449 retVal = self.dn2fluxErr(dn, dnErr, fluxMag0, fluxMag0Err) / self.dn2flux(dn, fluxMag0) 

1450 return self.FIVE_OVER_2LOG10 * retVal 

1451 

1452 

1453class NanoJansky(Photometry): 

1454 def _func(self, df): 

1455 return self.dn2flux(df[self.col], self.fluxMag0) 

1456 

1457 

1458class NanoJanskyErr(Photometry): 

1459 @property 

1460 def columns(self): 

1461 return [self.col, self.colFluxErr] 

1462 

1463 def _func(self, df): 

1464 retArr = self.dn2fluxErr(df[self.col], df[self.colFluxErr], self.fluxMag0, self.fluxMag0Err) 

1465 return pd.Series(retArr, index=df.index) 

1466 

1467 

1468class Magnitude(Photometry): 

1469 def _func(self, df): 

1470 return self.dn2mag(df[self.col], self.fluxMag0) 

1471 

1472 

1473class MagnitudeErr(Photometry): 

1474 @property 

1475 def columns(self): 

1476 return [self.col, self.colFluxErr] 

1477 

1478 def _func(self, df): 

1479 retArr = self.dn2MagErr(df[self.col], df[self.colFluxErr], self.fluxMag0, self.fluxMag0Err) 

1480 return pd.Series(retArr, index=df.index) 

1481 

1482 

1483class LocalPhotometry(Functor): 

1484 """Base class for calibrating the specified instrument flux column using 

1485 the local photometric calibration. 

1486 

1487 Parameters 

1488 ---------- 

1489 instFluxCol : `str` 

1490 Name of the instrument flux column. 

1491 instFluxErrCol : `str` 

1492 Name of the assocated error columns for ``instFluxCol``. 

1493 photoCalibCol : `str` 

1494 Name of local calibration column. 

1495 photoCalibErrCol : `str` 

1496 Error associated with ``photoCalibCol`` 

1497 

1498 See also 

1499 -------- 

1500 LocalPhotometry 

1501 LocalNanojansky 

1502 LocalNanojanskyErr 

1503 LocalMagnitude 

1504 LocalMagnitudeErr 

1505 """ 

1506 logNJanskyToAB = (1 * u.nJy).to_value(u.ABmag) 

1507 

1508 def __init__(self, 

1509 instFluxCol, 

1510 instFluxErrCol, 

1511 photoCalibCol, 

1512 photoCalibErrCol, 

1513 **kwargs): 

1514 self.instFluxCol = instFluxCol 

1515 self.instFluxErrCol = instFluxErrCol 

1516 self.photoCalibCol = photoCalibCol 

1517 self.photoCalibErrCol = photoCalibErrCol 

1518 super().__init__(**kwargs) 

1519 

1520 def instFluxToNanojansky(self, instFlux, localCalib): 

1521 """Convert instrument flux to nanojanskys. 

1522 

1523 Parameters 

1524 ---------- 

1525 instFlux : `numpy.ndarray` or `pandas.Series` 

1526 Array of instrument flux measurements 

1527 localCalib : `numpy.ndarray` or `pandas.Series` 

1528 Array of local photometric calibration estimates. 

1529 

1530 Returns 

1531 ------- 

1532 calibFlux : `numpy.ndarray` or `pandas.Series` 

1533 Array of calibrated flux measurements. 

1534 """ 

1535 return instFlux * localCalib 

1536 

1537 def instFluxErrToNanojanskyErr(self, instFlux, instFluxErr, localCalib, localCalibErr): 

1538 """Convert instrument flux to nanojanskys. 

1539 

1540 Parameters 

1541 ---------- 

1542 instFlux : `numpy.ndarray` or `pandas.Series` 

1543 Array of instrument flux measurements 

1544 instFluxErr : `numpy.ndarray` or `pandas.Series` 

1545 Errors on associated ``instFlux`` values 

1546 localCalib : `numpy.ndarray` or `pandas.Series` 

1547 Array of local photometric calibration estimates. 

1548 localCalibErr : `numpy.ndarray` or `pandas.Series` 

1549 Errors on associated ``localCalib`` values 

1550 

1551 Returns 

1552 ------- 

1553 calibFluxErr : `numpy.ndarray` or `pandas.Series` 

1554 Errors on calibrated flux measurements. 

1555 """ 

1556 return np.hypot(instFluxErr * localCalib, instFlux * localCalibErr) 

1557 

1558 def instFluxToMagnitude(self, instFlux, localCalib): 

1559 """Convert instrument flux to nanojanskys. 

1560 

1561 Parameters 

1562 ---------- 

1563 instFlux : `numpy.ndarray` or `pandas.Series` 

1564 Array of instrument flux measurements 

1565 localCalib : `numpy.ndarray` or `pandas.Series` 

1566 Array of local photometric calibration estimates. 

1567 

1568 Returns 

1569 ------- 

1570 calibMag : `numpy.ndarray` or `pandas.Series` 

1571 Array of calibrated AB magnitudes. 

1572 """ 

1573 return -2.5 * np.log10(self.instFluxToNanojansky(instFlux, localCalib)) + self.logNJanskyToAB 

1574 

1575 def instFluxErrToMagnitudeErr(self, instFlux, instFluxErr, localCalib, localCalibErr): 

1576 """Convert instrument flux err to nanojanskys. 

1577 

1578 Parameters 

1579 ---------- 

1580 instFlux : `numpy.ndarray` or `pandas.Series` 

1581 Array of instrument flux measurements 

1582 instFluxErr : `numpy.ndarray` or `pandas.Series` 

1583 Errors on associated ``instFlux`` values 

1584 localCalib : `numpy.ndarray` or `pandas.Series` 

1585 Array of local photometric calibration estimates. 

1586 localCalibErr : `numpy.ndarray` or `pandas.Series` 

1587 Errors on associated ``localCalib`` values 

1588 

1589 Returns 

1590 ------- 

1591 calibMagErr: `numpy.ndarray` or `pandas.Series` 

1592 Error on calibrated AB magnitudes. 

1593 """ 

1594 err = self.instFluxErrToNanojanskyErr(instFlux, instFluxErr, localCalib, localCalibErr) 

1595 return 2.5 / np.log(10) * err / self.instFluxToNanojansky(instFlux, instFluxErr) 

1596 

1597 

1598class LocalNanojansky(LocalPhotometry): 

1599 """Compute calibrated fluxes using the local calibration value. 

1600 

1601 See also 

1602 -------- 

1603 LocalNanojansky 

1604 LocalNanojanskyErr 

1605 LocalMagnitude 

1606 LocalMagnitudeErr 

1607 """ 

1608 

1609 @property 

1610 def columns(self): 

1611 return [self.instFluxCol, self.photoCalibCol] 

1612 

1613 @property 

1614 def name(self): 

1615 return f'flux_{self.instFluxCol}' 

1616 

1617 def _func(self, df): 

1618 return self.instFluxToNanojansky(df[self.instFluxCol], df[self.photoCalibCol]) 

1619 

1620 

1621class LocalNanojanskyErr(LocalPhotometry): 

1622 """Compute calibrated flux errors using the local calibration value. 

1623 

1624 See also 

1625 -------- 

1626 LocalNanojansky 

1627 LocalNanojanskyErr 

1628 LocalMagnitude 

1629 LocalMagnitudeErr 

1630 """ 

1631 

1632 @property 

1633 def columns(self): 

1634 return [self.instFluxCol, self.instFluxErrCol, 

1635 self.photoCalibCol, self.photoCalibErrCol] 

1636 

1637 @property 

1638 def name(self): 

1639 return f'fluxErr_{self.instFluxCol}' 

1640 

1641 def _func(self, df): 

1642 return self.instFluxErrToNanojanskyErr(df[self.instFluxCol], df[self.instFluxErrCol], 

1643 df[self.photoCalibCol], df[self.photoCalibErrCol]) 

1644 

1645 

1646class LocalMagnitude(LocalPhotometry): 

1647 """Compute calibrated AB magnitudes using the local calibration value. 

1648 

1649 See also 

1650 -------- 

1651 LocalNanojansky 

1652 LocalNanojanskyErr 

1653 LocalMagnitude 

1654 LocalMagnitudeErr 

1655 """ 

1656 

1657 @property 

1658 def columns(self): 

1659 return [self.instFluxCol, self.photoCalibCol] 

1660 

1661 @property 

1662 def name(self): 

1663 return f'mag_{self.instFluxCol}' 

1664 

1665 def _func(self, df): 

1666 return self.instFluxToMagnitude(df[self.instFluxCol], 

1667 df[self.photoCalibCol]) 

1668 

1669 

1670class LocalMagnitudeErr(LocalPhotometry): 

1671 """Compute calibrated AB magnitude errors using the local calibration value. 

1672 

1673 See also 

1674 -------- 

1675 LocalNanojansky 

1676 LocalNanojanskyErr 

1677 LocalMagnitude 

1678 LocalMagnitudeErr 

1679 """ 

1680 

1681 @property 

1682 def columns(self): 

1683 return [self.instFluxCol, self.instFluxErrCol, 

1684 self.photoCalibCol, self.photoCalibErrCol] 

1685 

1686 @property 

1687 def name(self): 

1688 return f'magErr_{self.instFluxCol}' 

1689 

1690 def _func(self, df): 

1691 return self.instFluxErrToMagnitudeErr(df[self.instFluxCol], 

1692 df[self.instFluxErrCol], 

1693 df[self.photoCalibCol], 

1694 df[self.photoCalibErrCol]) 

1695 

1696 

1697class LocalDipoleMeanFlux(LocalPhotometry): 

1698 """Compute absolute mean of dipole fluxes. 

1699 

1700 See also 

1701 -------- 

1702 LocalNanojansky 

1703 LocalNanojanskyErr 

1704 LocalMagnitude 

1705 LocalMagnitudeErr 

1706 LocalDipoleMeanFlux 

1707 LocalDipoleMeanFluxErr 

1708 LocalDipoleDiffFlux 

1709 LocalDipoleDiffFluxErr 

1710 """ 

1711 def __init__(self, 

1712 instFluxPosCol, 

1713 instFluxNegCol, 

1714 instFluxPosErrCol, 

1715 instFluxNegErrCol, 

1716 photoCalibCol, 

1717 photoCalibErrCol, 

1718 **kwargs): 

1719 self.instFluxNegCol = instFluxNegCol 

1720 self.instFluxPosCol = instFluxPosCol 

1721 self.instFluxNegErrCol = instFluxNegErrCol 

1722 self.instFluxPosErrCol = instFluxPosErrCol 

1723 self.photoCalibCol = photoCalibCol 

1724 self.photoCalibErrCol = photoCalibErrCol 

1725 super().__init__(instFluxNegCol, 

1726 instFluxNegErrCol, 

1727 photoCalibCol, 

1728 photoCalibErrCol, 

1729 **kwargs) 

1730 

1731 @property 

1732 def columns(self): 

1733 return [self.instFluxPosCol, 

1734 self.instFluxNegCol, 

1735 self.photoCalibCol] 

1736 

1737 @property 

1738 def name(self): 

1739 return f'dipMeanFlux_{self.instFluxPosCol}_{self.instFluxNegCol}' 

1740 

1741 def _func(self, df): 

1742 return 0.5*(np.fabs(self.instFluxToNanojansky(df[self.instFluxNegCol], df[self.photoCalibCol])) 

1743 + np.fabs(self.instFluxToNanojansky(df[self.instFluxPosCol], df[self.photoCalibCol]))) 

1744 

1745 

1746class LocalDipoleMeanFluxErr(LocalDipoleMeanFlux): 

1747 """Compute the error on the absolute mean of dipole fluxes. 

1748 

1749 See also 

1750 -------- 

1751 LocalNanojansky 

1752 LocalNanojanskyErr 

1753 LocalMagnitude 

1754 LocalMagnitudeErr 

1755 LocalDipoleMeanFlux 

1756 LocalDipoleMeanFluxErr 

1757 LocalDipoleDiffFlux 

1758 LocalDipoleDiffFluxErr 

1759 """ 

1760 

1761 @property 

1762 def columns(self): 

1763 return [self.instFluxPosCol, 

1764 self.instFluxNegCol, 

1765 self.instFluxPosErrCol, 

1766 self.instFluxNegErrCol, 

1767 self.photoCalibCol, 

1768 self.photoCalibErrCol] 

1769 

1770 @property 

1771 def name(self): 

1772 return f'dipMeanFluxErr_{self.instFluxPosCol}_{self.instFluxNegCol}' 

1773 

1774 def _func(self, df): 

1775 return 0.5*np.sqrt( 

1776 (np.fabs(df[self.instFluxNegCol]) + np.fabs(df[self.instFluxPosCol]) 

1777 * df[self.photoCalibErrCol])**2 

1778 + (df[self.instFluxNegErrCol]**2 + df[self.instFluxPosErrCol]**2) 

1779 * df[self.photoCalibCol]**2) 

1780 

1781 

1782class LocalDipoleDiffFlux(LocalDipoleMeanFlux): 

1783 """Compute the absolute difference of dipole fluxes. 

1784 

1785 Value is (abs(pos) - abs(neg)) 

1786 

1787 See also 

1788 -------- 

1789 LocalNanojansky 

1790 LocalNanojanskyErr 

1791 LocalMagnitude 

1792 LocalMagnitudeErr 

1793 LocalDipoleMeanFlux 

1794 LocalDipoleMeanFluxErr 

1795 LocalDipoleDiffFlux 

1796 LocalDipoleDiffFluxErr 

1797 """ 

1798 

1799 @property 

1800 def columns(self): 

1801 return [self.instFluxPosCol, 

1802 self.instFluxNegCol, 

1803 self.photoCalibCol] 

1804 

1805 @property 

1806 def name(self): 

1807 return f'dipDiffFlux_{self.instFluxPosCol}_{self.instFluxNegCol}' 

1808 

1809 def _func(self, df): 

1810 return (np.fabs(self.instFluxToNanojansky(df[self.instFluxPosCol], df[self.photoCalibCol])) 

1811 - np.fabs(self.instFluxToNanojansky(df[self.instFluxNegCol], df[self.photoCalibCol]))) 

1812 

1813 

1814class LocalDipoleDiffFluxErr(LocalDipoleMeanFlux): 

1815 """Compute the error on the absolute difference of dipole fluxes. 

1816 

1817 See also 

1818 -------- 

1819 LocalNanojansky 

1820 LocalNanojanskyErr 

1821 LocalMagnitude 

1822 LocalMagnitudeErr 

1823 LocalDipoleMeanFlux 

1824 LocalDipoleMeanFluxErr 

1825 LocalDipoleDiffFlux 

1826 LocalDipoleDiffFluxErr 

1827 """ 

1828 

1829 @property 

1830 def columns(self): 

1831 return [self.instFluxPosCol, 

1832 self.instFluxNegCol, 

1833 self.instFluxPosErrCol, 

1834 self.instFluxNegErrCol, 

1835 self.photoCalibCol, 

1836 self.photoCalibErrCol] 

1837 

1838 @property 

1839 def name(self): 

1840 return f'dipDiffFluxErr_{self.instFluxPosCol}_{self.instFluxNegCol}' 

1841 

1842 def _func(self, df): 

1843 return np.sqrt( 

1844 ((np.fabs(df[self.instFluxPosCol]) - np.fabs(df[self.instFluxNegCol])) 

1845 * df[self.photoCalibErrCol])**2 

1846 + (df[self.instFluxPosErrCol]**2 + df[self.instFluxNegErrCol]**2) 

1847 * df[self.photoCalibCol]**2) 

1848 

1849 

1850class Ratio(Functor): 

1851 """Base class for returning the ratio of 2 columns. 

1852 

1853 Can be used to compute a Signal to Noise ratio for any input flux. 

1854 

1855 Parameters 

1856 ---------- 

1857 numerator : `str` 

1858 Name of the column to use at the numerator in the ratio 

1859 denominator : `str` 

1860 Name of the column to use as the denominator in the ratio. 

1861 """ 

1862 def __init__(self, 

1863 numerator, 

1864 denominator, 

1865 **kwargs): 

1866 self.numerator = numerator 

1867 self.denominator = denominator 

1868 super().__init__(**kwargs) 

1869 

1870 @property 

1871 def columns(self): 

1872 return [self.numerator, self.denominator] 

1873 

1874 @property 

1875 def name(self): 

1876 return f'ratio_{self.numerator}_{self.denominator}' 

1877 

1878 def _func(self, df): 

1879 with np.warnings.catch_warnings(): 

1880 np.warnings.filterwarnings('ignore', r'invalid value encountered') 

1881 np.warnings.filterwarnings('ignore', r'divide by zero') 

1882 return df[self.numerator] / df[self.denominator]