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 

33from .parquetTable import ParquetTable, MultilevelParquetTable 

34 

35 

36def init_fromDict(initDict, basePath='lsst.pipe.tasks.functors', 

37 typeKey='functor', name=None): 

38 """Initialize an object defined in a dictionary 

39 

40 The object needs to be importable as 

41 f'{basePath}.{initDict[typeKey]}' 

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

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

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

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

46 

47 Parameters 

48 ---------- 

49 initDict : dictionary 

50 Dictionary describing object's initialization. Must contain 

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

52 relative to ``basePath``. 

53 basePath : str 

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

55 typeKey : str 

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

57 (relative to `basePath`). 

58 """ 

59 initDict = initDict.copy() 

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

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

62 args = [] 

63 if 'args' in initDict: 

64 args = initDict.pop('args') 

65 if isinstance(args, str): 

66 args = [args] 

67 try: 

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

69 except Exception as e: 

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

71 raise type(e)(message, e.args) 

72 return element 

73 

74 

75class Functor(object): 

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

77 

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

79 `DeferredDatasetHandle`, and returns the 

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

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

82 from the `ParquetTable`. 

83 

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

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

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

87 result. 

88 

89 

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

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

92 In addition, it must define the following attributes 

93 

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

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

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

97 

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

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

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

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

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

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

104 perform the calculation. 

105 

106 As currently implemented, `Functor` is only set up to expect a 

107 dataset of the format of the `deepCoadd_obj` dataset; that is, a 

108 dataframe with a multi-level column index, 

109 with the levels of the column index being `band`, 

110 `dataset`, and `column`. This is defined in the `_columnLevels` attribute, 

111 as well as being implicit in the role of the `filt` and `dataset` attributes 

112 defined at initialization. In addition, the `_get_data` method that reads 

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

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

115 `column`. 

116 

117 The `_columnLevels` and `_dfLevels` attributes should generally not need to 

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

119 to do the calculation. 

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

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

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

123 

124 Parameters 

125 ---------- 

126 filt : str 

127 Filter upon which to do the calculation 

128 

129 dataset : str 

130 Dataset upon which to do the calculation 

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

132 

133 """ 

134 

135 _defaultDataset = 'ref' 

136 _columnLevels = ('band', 'dataset', 'column') 

137 _dfLevels = ('column',) 

138 _defaultNoDup = False 

139 

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

141 self.filt = filt 

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

143 self._noDup = noDup 

144 

145 @property 

146 def noDup(self): 

147 if self._noDup is not None: 

148 return self._noDup 

149 else: 

150 return self._defaultNoDup 

151 

152 @property 

153 def columns(self): 

154 """Columns required to perform calculation 

155 """ 

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

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

158 return self._columns 

159 

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

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

162 

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

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

165 and with the gen3 `DeferredDatasetHandle`. 

166 

167 Parameters 

168 ---------- 

169 data : `MultilevelParquetTable` or `DeferredDatasetHandle` 

170 

171 columnnIndex (optional): pandas `Index` object 

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

173 """ 

174 if isinstance(data, DeferredDatasetHandle): 

175 if columnIndex is None: 

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

177 if columnIndex is not None: 

178 return columnIndex.names 

179 if isinstance(data, MultilevelParquetTable): 

180 return data.columnLevels 

181 else: 

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

183 

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

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

186 

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

188 

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

190 """ 

191 if isinstance(data, DeferredDatasetHandle): 

192 if columnIndex is None: 

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

194 if columnIndex is not None: 

195 columnLevels = columnIndex.names 

196 columnLevelNames = { 

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

198 for i, level in enumerate(columnLevels) 

199 } 

200 return columnLevelNames 

201 if isinstance(data, MultilevelParquetTable): 

202 return data.columnLevelNames 

203 else: 

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

205 

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

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

208 

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

210 """ 

211 new_colDict = {} 

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

213 

214 for i, lev in enumerate(columnLevels): 

215 if lev in colDict: 

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

217 new_colDict[lev] = [colDict[lev]] 

218 else: 

219 new_colDict[lev] = colDict[lev] 

220 else: 

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

222 

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

224 cols = product(*levelCols) 

225 return list(cols) 

226 

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

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

229 

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

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

232 dictionary. 

233 

234 Parameters 

235 ---------- 

236 data : `MultilevelParquetTable` or `DeferredDatasetHandle` 

237 

238 columnIndex (optional): pandas `Index` object 

239 either passed or read in from `DeferredDatasetHandle`. 

240 

241 `returnTuple` : bool 

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

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

244 combine columns from the various component functors. 

245 

246 """ 

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

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

249 

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

251 columnLevels = self._get_data_columnLevels(data, columnIndex) 

252 

253 if not set(columnLevels) == set(self._columnLevels): 

254 raise ValueError( 

255 "ParquetTable does not have the expected column levels. " 

256 f"Got {columnLevels}; expected {self._columnLevels}." 

257 ) 

258 

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

260 'dataset': self.dataset} 

261 if self.filt is None: 

262 columnLevelNames = self._get_data_columnLevelNames(data, columnIndex) 

263 if "band" in columnLevels: 

264 if self.dataset == "ref": 

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

266 else: 

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

268 f"(dataset {self.dataset}) " 

269 "and ParquetTable " 

270 "contains multiple filters in column index. " 

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

272 else: 

273 columnDict['band'] = self.filt 

274 

275 if isinstance(data, MultilevelParquetTable): 

276 return data._colsFromDict(columnDict) 

277 elif isinstance(data, DeferredDatasetHandle): 

278 if returnTuple: 

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

280 else: 

281 return columnDict 

282 

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

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

285 

286 def _get_columnIndex(self, data): 

287 """Return columnIndex 

288 """ 

289 

290 if isinstance(data, DeferredDatasetHandle): 

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

292 else: 

293 return None 

294 

295 def _get_data(self, data): 

296 """Retrieve dataframe necessary for calculation. 

297 

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

299 

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

301 

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

303 """ 

304 if isinstance(data, pd.DataFrame): 

305 return data 

306 

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

308 columnIndex = self._get_columnIndex(data) 

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

310 

311 # Simple single-level parquet table, gen2 

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

313 columns = self.columns 

314 df = data.toDataFrame(columns=columns) 

315 return df 

316 

317 # Get proper columns specification for this functor 

318 if is_multiLevel: 

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

320 else: 

321 columns = self.columns 

322 

323 if isinstance(data, MultilevelParquetTable): 

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

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

326 elif isinstance(data, DeferredDatasetHandle): 

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

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

329 

330 # Drop unnecessary column levels 

331 if is_multiLevel: 

332 df = self._setLevels(df) 

333 

334 return df 

335 

336 def _setLevels(self, df): 

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

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

339 return df 

340 

341 def _dropna(self, vals): 

342 return vals.dropna() 

343 

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

345 try: 

346 df = self._get_data(data) 

347 vals = self._func(df) 

348 except Exception: 

349 vals = self.fail(df) 

350 if dropna: 

351 vals = self._dropna(vals) 

352 

353 return vals 

354 

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

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

357 """ 

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

359 

360 def fail(self, df): 

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

362 

363 @property 

364 def name(self): 

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

366 """ 

367 return NotImplementedError 

368 

369 @property 

370 def shortname(self): 

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

372 """ 

373 return self.name 

374 

375 

376class CompositeFunctor(Functor): 

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

378 

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

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

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

382 being the keys of `funcDict`. 

383 

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

385 in all the component functors. 

386 

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

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

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

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

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

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

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

394 

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

396 file specifying a collection of functors. 

397 

398 Parameters 

399 ---------- 

400 funcs : `dict` or `list` 

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

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

403 

404 """ 

405 dataset = None 

406 

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

408 

409 if type(funcs) == dict: 

410 self.funcDict = funcs 

411 else: 

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

413 

414 self._filt = None 

415 

416 super().__init__(**kwargs) 

417 

418 @property 

419 def filt(self): 

420 return self._filt 

421 

422 @filt.setter 

423 def filt(self, filt): 

424 if filt is not None: 

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

426 f.filt = filt 

427 self._filt = filt 

428 

429 def update(self, new): 

430 if isinstance(new, dict): 

431 self.funcDict.update(new) 

432 elif isinstance(new, CompositeFunctor): 

433 self.funcDict.update(new.funcDict) 

434 else: 

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

436 

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

438 if self.filt is not None: 

439 self.filt = self.filt 

440 

441 @property 

442 def columns(self): 

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

444 

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

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

447 return list( 

448 set( 

449 [ 

450 x 

451 for y in [ 

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

453 ] 

454 for x in y 

455 ] 

456 ) 

457 ) 

458 

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

460 """Apply the functor to the data table 

461 

462 Parameters 

463 ---------- 

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

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

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

467 or `pandas.DataFrame`. 

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

469 be accessed 

470 """ 

471 columnIndex = self._get_columnIndex(data) 

472 

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

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

475 

476 # Multilevel index, gen2 or gen3 

477 if is_multiLevel: 

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

479 

480 if isinstance(data, MultilevelParquetTable): 

481 # Read data into memory the gen2 way 

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

483 elif isinstance(data, DeferredDatasetHandle): 

484 # Read data into memory the gen3 way 

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

486 

487 valDict = {} 

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

489 try: 

490 subdf = f._setLevels( 

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

492 ) 

493 valDict[k] = f._func(subdf) 

494 except Exception: 

495 valDict[k] = f.fail(subdf) 

496 

497 else: 

498 if isinstance(data, DeferredDatasetHandle): 

499 # input if Gen3 deferLoad=True 

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

501 elif isinstance(data, pd.DataFrame): 

502 # input if Gen3 deferLoad=False 

503 df = data 

504 else: 

505 # Original Gen2 input is type ParquetTable and the fallback 

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

507 

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

509 

510 try: 

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

512 except TypeError: 

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

514 raise 

515 

516 if kwargs.get('dropna', False): 

517 valDf = valDf.dropna(how='any') 

518 

519 return valDf 

520 

521 @classmethod 

522 def renameCol(cls, col, renameRules): 

523 if renameRules is None: 

524 return col 

525 for old, new in renameRules: 

526 if col.startswith(old): 

527 col = col.replace(old, new) 

528 return col 

529 

530 @classmethod 

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

532 # Allow environment variables in the filename. 

533 filename = os.path.expandvars(filename) 

534 with open(filename) as f: 

535 translationDefinition = yaml.safe_load(f) 

536 

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

538 

539 @classmethod 

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

541 funcs = {} 

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

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

544 

545 if 'flag_rename_rules' in translationDefinition: 

546 renameRules = translationDefinition['flag_rename_rules'] 

547 else: 

548 renameRules = None 

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 'flags' in translationDefinition: 

555 for flag in translationDefinition['flags']: 

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

557 

558 return cls(funcs, **kwargs) 

559 

560 

561def mag_aware_eval(df, expr): 

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

563 

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

565 

566 Parameters 

567 ---------- 

568 df : pandas.DataFrame 

569 Dataframe on which to evaluate expression. 

570 

571 expr : str 

572 Expression. 

573 """ 

574 try: 

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

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

577 except Exception: # Should check what actually gets raised 

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

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

580 return val 

581 

582 

583class CustomFunctor(Functor): 

584 """Arbitrary computation on a catalog 

585 

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

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

588 

589 Parameters 

590 ---------- 

591 expr : str 

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

593 """ 

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

595 

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

597 self.expr = expr 

598 super().__init__(**kwargs) 

599 

600 @property 

601 def name(self): 

602 return self.expr 

603 

604 @property 

605 def columns(self): 

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

607 

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

609 not_a_col = [] 

610 for c in flux_cols: 

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

612 cols.append(f'{c}_instFlux') 

613 not_a_col.append(c) 

614 else: 

615 cols.append(c) 

616 

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

618 

619 def _func(self, df): 

620 return mag_aware_eval(df, self.expr) 

621 

622 

623class Column(Functor): 

624 """Get column with specified name 

625 """ 

626 

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

628 self.col = col 

629 super().__init__(**kwargs) 

630 

631 @property 

632 def name(self): 

633 return self.col 

634 

635 @property 

636 def columns(self): 

637 return [self.col] 

638 

639 def _func(self, df): 

640 return df[self.col] 

641 

642 

643class Index(Functor): 

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

645 """ 

646 

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

648 _defaultDataset = 'ref' 

649 _defaultNoDup = True 

650 

651 def _func(self, df): 

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

653 

654 

655class IDColumn(Column): 

656 col = 'id' 

657 _allow_difference = False 

658 _defaultNoDup = True 

659 

660 def _func(self, df): 

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

662 

663 

664class FootprintNPix(Column): 

665 col = 'base_Footprint_nPix' 

666 

667 

668class CoordColumn(Column): 

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

670 """ 

671 _radians = True 

672 

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

674 super().__init__(col, **kwargs) 

675 

676 def _func(self, df): 

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

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

679 return output 

680 

681 

682class RAColumn(CoordColumn): 

683 """Right Ascension, in degrees 

684 """ 

685 name = 'RA' 

686 _defaultNoDup = True 

687 

688 def __init__(self, **kwargs): 

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

690 

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

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

693 

694 

695class DecColumn(CoordColumn): 

696 """Declination, in degrees 

697 """ 

698 name = 'Dec' 

699 _defaultNoDup = True 

700 

701 def __init__(self, **kwargs): 

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

703 

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

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

706 

707 

708def fluxName(col): 

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

710 col += '_instFlux' 

711 return col 

712 

713 

714def fluxErrName(col): 

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

716 col += '_instFluxErr' 

717 return col 

718 

719 

720class Mag(Functor): 

721 """Compute calibrated magnitude 

722 

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

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

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

726 This default should be removed in DM-21955 

727 

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

729 

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

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

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

733 

734 Parameters 

735 ---------- 

736 col : `str` 

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

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

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

740 understand. 

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

742 Object that knows zero point. 

743 """ 

744 _defaultDataset = 'meas' 

745 

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

747 self.col = fluxName(col) 

748 self.calib = calib 

749 if calib is not None: 

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

751 else: 

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

753 self.fluxMag0 = 63095734448.0194 

754 

755 super().__init__(**kwargs) 

756 

757 @property 

758 def columns(self): 

759 return [self.col] 

760 

761 def _func(self, df): 

762 with np.warnings.catch_warnings(): 

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

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

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

766 

767 @property 

768 def name(self): 

769 return f'mag_{self.col}' 

770 

771 

772class MagErr(Mag): 

773 """Compute calibrated magnitude uncertainty 

774 

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

776 

777 Parameters 

778 col : `str` 

779 Name of flux column 

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

781 Object that knows zero point. 

782 """ 

783 

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

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

786 if self.calib is not None: 

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

788 else: 

789 self.fluxMag0Err = 0. 

790 

791 @property 

792 def columns(self): 

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

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 fluxCol, fluxErrCol = self.columns 

800 x = df[fluxErrCol] / df[fluxCol] 

801 y = self.fluxMag0Err / self.fluxMag0 

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

803 return magErr 

804 

805 @property 

806 def name(self): 

807 return super().name + '_err' 

808 

809 

810class NanoMaggie(Mag): 

811 """ 

812 """ 

813 

814 def _func(self, df): 

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

816 

817 

818class MagDiff(Functor): 

819 _defaultDataset = 'meas' 

820 

821 """Functor to calculate magnitude difference""" 

822 

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

824 self.col1 = fluxName(col1) 

825 self.col2 = fluxName(col2) 

826 super().__init__(**kwargs) 

827 

828 @property 

829 def columns(self): 

830 return [self.col1, self.col2] 

831 

832 def _func(self, df): 

833 with np.warnings.catch_warnings(): 

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

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

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

837 

838 @property 

839 def name(self): 

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

841 

842 @property 

843 def shortname(self): 

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

845 

846 

847class Color(Functor): 

848 """Compute the color between two filters 

849 

850 Computes color by initializing two different `Mag` 

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

852 then returning the difference. 

853 

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

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

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

857 This is controlled by the `_dfLevels` attribute. 

858 

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

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

861 

862 Parameters 

863 ---------- 

864 col : str 

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

866 `lsst.pipe.tasks.functors.Mag`. 

867 

868 filt2, filt1 : str 

869 Filters from which to compute magnitude difference. 

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

871 """ 

872 _defaultDataset = 'forced_src' 

873 _dfLevels = ('band', 'column') 

874 _defaultNoDup = True 

875 

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

877 self.col = fluxName(col) 

878 if filt2 == filt1: 

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

880 self.filt2 = filt2 

881 self.filt1 = filt1 

882 

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

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

885 

886 super().__init__(**kwargs) 

887 

888 @property 

889 def filt(self): 

890 return None 

891 

892 @filt.setter 

893 def filt(self, filt): 

894 pass 

895 

896 def _func(self, df): 

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

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

899 return mag2 - mag1 

900 

901 @property 

902 def columns(self): 

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

904 

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

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

907 

908 @property 

909 def name(self): 

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

911 

912 @property 

913 def shortname(self): 

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

915 

916 

917class Labeller(Functor): 

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

919 """ 

920 _null_label = 'null' 

921 _allow_difference = False 

922 name = 'label' 

923 _force_str = False 

924 

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

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

927 

928 

929class StarGalaxyLabeller(Labeller): 

930 _columns = ["base_ClassificationExtendedness_value"] 

931 _column = "base_ClassificationExtendedness_value" 

932 

933 def _func(self, df): 

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

935 mask = x.isnull() 

936 test = (x < 0.5).astype(int) 

937 test = test.mask(mask, 2) 

938 

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

940 # are these backwards? 

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

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

943 index=x.index, name='label') 

944 if self._force_str: 

945 label = label.astype(str) 

946 return label 

947 

948 

949class NumStarLabeller(Labeller): 

950 _columns = ['numStarFlags'] 

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

952 

953 def _func(self, df): 

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

955 

956 # Number of filters 

957 n = len(x.unique()) - 1 

958 

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

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

961 index=x.index, name='label') 

962 

963 if self._force_str: 

964 label = label.astype(str) 

965 

966 return label 

967 

968 

969class DeconvolvedMoments(Functor): 

970 name = 'Deconvolved Moments' 

971 shortname = 'deconvolvedMoments' 

972 _columns = ("ext_shapeHSM_HsmSourceMoments_xx", 

973 "ext_shapeHSM_HsmSourceMoments_yy", 

974 "base_SdssShape_xx", "base_SdssShape_yy", 

975 "ext_shapeHSM_HsmPsfMoments_xx", 

976 "ext_shapeHSM_HsmPsfMoments_yy") 

977 

978 def _func(self, df): 

979 """Calculate deconvolved moments""" 

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

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

982 else: 

983 hsm = np.ones(len(df))*np.nan 

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

985 if "ext_shapeHSM_HsmPsfMoments_xx" in df.columns: 

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

987 else: 

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

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

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

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

992 

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

994 

995 

996class SdssTraceSize(Functor): 

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

998 name = "SDSS Trace Size" 

999 shortname = 'sdssTrace' 

1000 _columns = ("base_SdssShape_xx", "base_SdssShape_yy") 

1001 

1002 def _func(self, df): 

1003 srcSize = np.sqrt(0.5*(df["base_SdssShape_xx"] + df["base_SdssShape_yy"])) 

1004 return srcSize 

1005 

1006 

1007class PsfSdssTraceSizeDiff(Functor): 

1008 """Functor to calculate SDSS trace radius size difference (%) between object and psf model""" 

1009 name = "PSF - SDSS Trace Size" 

1010 shortname = 'psf_sdssTrace' 

1011 _columns = ("base_SdssShape_xx", "base_SdssShape_yy", 

1012 "base_SdssShape_psf_xx", "base_SdssShape_psf_yy") 

1013 

1014 def _func(self, df): 

1015 srcSize = np.sqrt(0.5*(df["base_SdssShape_xx"] + df["base_SdssShape_yy"])) 

1016 psfSize = np.sqrt(0.5*(df["base_SdssShape_psf_xx"] + df["base_SdssShape_psf_yy"])) 

1017 sizeDiff = 100*(srcSize - psfSize)/(0.5*(srcSize + psfSize)) 

1018 return sizeDiff 

1019 

1020 

1021class HsmTraceSize(Functor): 

1022 """Functor to calculate HSM trace radius size for sources""" 

1023 name = 'HSM Trace Size' 

1024 shortname = 'hsmTrace' 

1025 _columns = ("ext_shapeHSM_HsmSourceMoments_xx", 

1026 "ext_shapeHSM_HsmSourceMoments_yy") 

1027 

1028 def _func(self, df): 

1029 srcSize = np.sqrt(0.5*(df["ext_shapeHSM_HsmSourceMoments_xx"] 

1030 + df["ext_shapeHSM_HsmSourceMoments_yy"])) 

1031 return srcSize 

1032 

1033 

1034class PsfHsmTraceSizeDiff(Functor): 

1035 """Functor to calculate HSM trace radius size difference (%) between object and psf model""" 

1036 name = 'PSF - HSM Trace Size' 

1037 shortname = 'psf_HsmTrace' 

1038 _columns = ("ext_shapeHSM_HsmSourceMoments_xx", 

1039 "ext_shapeHSM_HsmSourceMoments_yy", 

1040 "ext_shapeHSM_HsmPsfMoments_xx", 

1041 "ext_shapeHSM_HsmPsfMoments_yy") 

1042 

1043 def _func(self, df): 

1044 srcSize = np.sqrt(0.5*(df["ext_shapeHSM_HsmSourceMoments_xx"] 

1045 + df["ext_shapeHSM_HsmSourceMoments_yy"])) 

1046 psfSize = np.sqrt(0.5*(df["ext_shapeHSM_HsmPsfMoments_xx"] 

1047 + df["ext_shapeHSM_HsmPsfMoments_yy"])) 

1048 sizeDiff = 100*(srcSize - psfSize)/(0.5*(srcSize + psfSize)) 

1049 return sizeDiff 

1050 

1051 

1052class HsmFwhm(Functor): 

1053 name = 'HSM Psf FWHM' 

1054 _columns = ('ext_shapeHSM_HsmPsfMoments_xx', 'ext_shapeHSM_HsmPsfMoments_yy') 

1055 # TODO: DM-21403 pixel scale should be computed from the CD matrix or transform matrix 

1056 pixelScale = 0.168 

1057 SIGMA2FWHM = 2*np.sqrt(2*np.log(2)) 

1058 

1059 def _func(self, df): 

1060 return self.pixelScale*self.SIGMA2FWHM*np.sqrt( 

1061 0.5*(df['ext_shapeHSM_HsmPsfMoments_xx'] + df['ext_shapeHSM_HsmPsfMoments_yy'])) 

1062 

1063 

1064class E1(Functor): 

1065 name = "Distortion Ellipticity (e1)" 

1066 shortname = "Distortion" 

1067 

1068 def __init__(self, colXX, colXY, colYY, **kwargs): 

1069 self.colXX = colXX 

1070 self.colXY = colXY 

1071 self.colYY = colYY 

1072 self._columns = [self.colXX, self.colXY, self.colYY] 

1073 super().__init__(**kwargs) 

1074 

1075 @property 

1076 def columns(self): 

1077 return [self.colXX, self.colXY, self.colYY] 

1078 

1079 def _func(self, df): 

1080 return df[self.colXX] - df[self.colYY] / (df[self.colXX] + df[self.colYY]) 

1081 

1082 

1083class E2(Functor): 

1084 name = "Ellipticity e2" 

1085 

1086 def __init__(self, colXX, colXY, colYY, **kwargs): 

1087 self.colXX = colXX 

1088 self.colXY = colXY 

1089 self.colYY = colYY 

1090 super().__init__(**kwargs) 

1091 

1092 @property 

1093 def columns(self): 

1094 return [self.colXX, self.colXY, self.colYY] 

1095 

1096 def _func(self, df): 

1097 return 2*df[self.colXY] / (df[self.colXX] + df[self.colYY]) 

1098 

1099 

1100class RadiusFromQuadrupole(Functor): 

1101 

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

1103 self.colXX = colXX 

1104 self.colXY = colXY 

1105 self.colYY = colYY 

1106 super().__init__(**kwargs) 

1107 

1108 @property 

1109 def columns(self): 

1110 return [self.colXX, self.colXY, self.colYY] 

1111 

1112 def _func(self, df): 

1113 return (df[self.colXX]*df[self.colYY] - df[self.colXY]**2)**0.25 

1114 

1115 

1116class LocalWcs(Functor): 

1117 """Computations using the stored localWcs. 

1118 """ 

1119 name = "LocalWcsOperations" 

1120 

1121 def __init__(self, 

1122 colCD_1_1, 

1123 colCD_1_2, 

1124 colCD_2_1, 

1125 colCD_2_2, 

1126 **kwargs): 

1127 self.colCD_1_1 = colCD_1_1 

1128 self.colCD_1_2 = colCD_1_2 

1129 self.colCD_2_1 = colCD_2_1 

1130 self.colCD_2_2 = colCD_2_2 

1131 super().__init__(**kwargs) 

1132 

1133 def computeDeltaRaDec(self, x, y, cd11, cd12, cd21, cd22): 

1134 """Compute the distance on the sphere from x2, y1 to x1, y1. 

1135 

1136 Parameters 

1137 ---------- 

1138 x : `pandas.Series` 

1139 X pixel coordinate. 

1140 y : `pandas.Series` 

1141 Y pixel coordinate. 

1142 cd11 : `pandas.Series` 

1143 [1, 1] element of the local Wcs affine transform. 

1144 cd11 : `pandas.Series` 

1145 [1, 1] element of the local Wcs affine transform. 

1146 cd12 : `pandas.Series` 

1147 [1, 2] element of the local Wcs affine transform. 

1148 cd21 : `pandas.Series` 

1149 [2, 1] element of the local Wcs affine transform. 

1150 cd22 : `pandas.Series` 

1151 [2, 2] element of the local Wcs affine transform. 

1152 

1153 Returns 

1154 ------- 

1155 raDecTuple : tuple 

1156 RA and dec conversion of x and y given the local Wcs. Returned 

1157 units are in radians. 

1158 

1159 """ 

1160 return (x * cd11 + y * cd12, x * cd21 + y * cd22) 

1161 

1162 def computeSkySeperation(self, ra1, dec1, ra2, dec2): 

1163 """Compute the local pixel scale conversion. 

1164 

1165 Parameters 

1166 ---------- 

1167 ra1 : `pandas.Series` 

1168 Ra of the first coordinate in radians. 

1169 dec1 : `pandas.Series` 

1170 Dec of the first coordinate in radians. 

1171 ra2 : `pandas.Series` 

1172 Ra of the second coordinate in radians. 

1173 dec2 : `pandas.Series` 

1174 Dec of the second coordinate in radians. 

1175 

1176 Returns 

1177 ------- 

1178 dist : `pandas.Series` 

1179 Distance on the sphere in radians. 

1180 """ 

1181 deltaDec = dec2 - dec1 

1182 deltaRa = ra2 - ra1 

1183 return 2 * np.arcsin( 

1184 np.sqrt( 

1185 np.sin(deltaDec / 2) ** 2 

1186 + np.cos(dec2) * np.cos(dec1) * np.sin(deltaRa / 2) ** 2)) 

1187 

1188 def getSkySeperationFromPixel(self, x1, y1, x2, y2, cd11, cd12, cd21, cd22): 

1189 """Compute the distance on the sphere from x2, y1 to x1, y1. 

1190 

1191 Parameters 

1192 ---------- 

1193 x1 : `pandas.Series` 

1194 X pixel coordinate. 

1195 y1 : `pandas.Series` 

1196 Y pixel coordinate. 

1197 x2 : `pandas.Series` 

1198 X pixel coordinate. 

1199 y2 : `pandas.Series` 

1200 Y pixel coordinate. 

1201 cd11 : `pandas.Series` 

1202 [1, 1] element of the local Wcs affine transform. 

1203 cd11 : `pandas.Series` 

1204 [1, 1] element of the local Wcs affine transform. 

1205 cd12 : `pandas.Series` 

1206 [1, 2] element of the local Wcs affine transform. 

1207 cd21 : `pandas.Series` 

1208 [2, 1] element of the local Wcs affine transform. 

1209 cd22 : `pandas.Series` 

1210 [2, 2] element of the local Wcs affine transform. 

1211 

1212 Returns 

1213 ------- 

1214 Distance : `pandas.Series` 

1215 Arcseconds per pixel at the location of the local WC 

1216 """ 

1217 ra1, dec1 = self.computeDeltaRaDec(x1, y1, cd11, cd12, cd21, cd22) 

1218 ra2, dec2 = self.computeDeltaRaDec(x2, y2, cd11, cd12, cd21, cd22) 

1219 # Great circle distance for small separations. 

1220 return self.computeSkySeperation(ra1, dec1, ra2, dec2) 

1221 

1222 

1223class ComputePixelScale(LocalWcs): 

1224 """Compute the local pixel scale from the stored CDMatrix. 

1225 """ 

1226 name = "PixelScale" 

1227 

1228 @property 

1229 def columns(self): 

1230 return [self.colCD_1_1, 

1231 self.colCD_1_2, 

1232 self.colCD_2_1, 

1233 self.colCD_2_2] 

1234 

1235 def pixelScaleArcseconds(self, cd11, cd12, cd21, cd22): 

1236 """Compute the local pixel to scale conversion in arcseconds. 

1237 

1238 Parameters 

1239 ---------- 

1240 cd11 : `pandas.Series` 

1241 [1, 1] element of the local Wcs affine transform in radians. 

1242 cd11 : `pandas.Series` 

1243 [1, 1] element of the local Wcs affine transform in radians. 

1244 cd12 : `pandas.Series` 

1245 [1, 2] element of the local Wcs affine transform in radians. 

1246 cd21 : `pandas.Series` 

1247 [2, 1] element of the local Wcs affine transform in radians. 

1248 cd22 : `pandas.Series` 

1249 [2, 2] element of the local Wcs affine transform in radians. 

1250 

1251 Returns 

1252 ------- 

1253 pixScale : `pandas.Series` 

1254 Arcseconds per pixel at the location of the local WC 

1255 """ 

1256 return 3600 * np.degrees(np.sqrt(np.fabs(cd11 * cd22 - cd12 * cd21))) 

1257 

1258 def _func(self, df): 

1259 return self.pixelScaleArcseconds(df[self.colCD_1_1], 

1260 df[self.colCD_1_2], 

1261 df[self.colCD_2_1], 

1262 df[self.colCD_2_2]) 

1263 

1264 

1265class ConvertPixelToArcseconds(ComputePixelScale): 

1266 """Convert a value in units pixels squared to units arcseconds squared. 

1267 """ 

1268 

1269 def __init__(self, 

1270 col, 

1271 colCD_1_1, 

1272 colCD_1_2, 

1273 colCD_2_1, 

1274 colCD_2_2, 

1275 **kwargs): 

1276 self.col = col 

1277 super().__init__(colCD_1_1, 

1278 colCD_1_2, 

1279 colCD_2_1, 

1280 colCD_2_2, 

1281 **kwargs) 

1282 

1283 @property 

1284 def name(self): 

1285 return f"{self.col}_asArcseconds" 

1286 

1287 @property 

1288 def columns(self): 

1289 return [self.col, 

1290 self.colCD_1_1, 

1291 self.colCD_1_2, 

1292 self.colCD_2_1, 

1293 self.colCD_2_2] 

1294 

1295 def _func(self, df): 

1296 return df[self.col] * self.pixelScaleArcseconds(df[self.colCD_1_1], 

1297 df[self.colCD_1_2], 

1298 df[self.colCD_2_1], 

1299 df[self.colCD_2_2]) 

1300 

1301 

1302class ConvertPixelSqToArcsecondsSq(ComputePixelScale): 

1303 """Convert a value in units pixels to units arcseconds. 

1304 """ 

1305 

1306 def __init__(self, 

1307 col, 

1308 colCD_1_1, 

1309 colCD_1_2, 

1310 colCD_2_1, 

1311 colCD_2_2, 

1312 **kwargs): 

1313 self.col = col 

1314 super().__init__(colCD_1_1, 

1315 colCD_1_2, 

1316 colCD_2_1, 

1317 colCD_2_2, 

1318 **kwargs) 

1319 

1320 @property 

1321 def name(self): 

1322 return f"{self.col}_asArcsecondsSq" 

1323 

1324 @property 

1325 def columns(self): 

1326 return [self.col, 

1327 self.colCD_1_1, 

1328 self.colCD_1_2, 

1329 self.colCD_2_1, 

1330 self.colCD_2_2] 

1331 

1332 def _func(self, df): 

1333 pixScale = self.pixelScaleArcseconds(df[self.colCD_1_1], 

1334 df[self.colCD_1_2], 

1335 df[self.colCD_2_1], 

1336 df[self.colCD_2_2]) 

1337 return df[self.col] * pixScale * pixScale 

1338 

1339 

1340class ReferenceBand(Functor): 

1341 name = 'Reference Band' 

1342 shortname = 'refBand' 

1343 

1344 @property 

1345 def columns(self): 

1346 return ["merge_measurement_i", 

1347 "merge_measurement_r", 

1348 "merge_measurement_z", 

1349 "merge_measurement_y", 

1350 "merge_measurement_g"] 

1351 

1352 def _func(self, df): 

1353 def getFilterAliasName(row): 

1354 # get column name with the max value (True > False) 

1355 colName = row.idxmax() 

1356 return colName.replace('merge_measurement_', '') 

1357 

1358 return df[self.columns].apply(getFilterAliasName, axis=1) 

1359 

1360 

1361class Photometry(Functor): 

1362 # AB to NanoJansky (3631 Jansky) 

1363 AB_FLUX_SCALE = (0 * u.ABmag).to_value(u.nJy) 

1364 LOG_AB_FLUX_SCALE = 12.56 

1365 FIVE_OVER_2LOG10 = 1.085736204758129569 

1366 # TO DO: DM-21955 Replace hard coded photometic calibration values 

1367 COADD_ZP = 27 

1368 

1369 def __init__(self, colFlux, colFluxErr=None, calib=None, **kwargs): 

1370 self.vhypot = np.vectorize(self.hypot) 

1371 self.col = colFlux 

1372 self.colFluxErr = colFluxErr 

1373 

1374 self.calib = calib 

1375 if calib is not None: 

1376 self.fluxMag0, self.fluxMag0Err = calib.getFluxMag0() 

1377 else: 

1378 self.fluxMag0 = 1./np.power(10, -0.4*self.COADD_ZP) 

1379 self.fluxMag0Err = 0. 

1380 

1381 super().__init__(**kwargs) 

1382 

1383 @property 

1384 def columns(self): 

1385 return [self.col] 

1386 

1387 @property 

1388 def name(self): 

1389 return f'mag_{self.col}' 

1390 

1391 @classmethod 

1392 def hypot(cls, a, b): 

1393 if np.abs(a) < np.abs(b): 

1394 a, b = b, a 

1395 if a == 0.: 

1396 return 0. 

1397 q = b/a 

1398 return np.abs(a) * np.sqrt(1. + q*q) 

1399 

1400 def dn2flux(self, dn, fluxMag0): 

1401 return self.AB_FLUX_SCALE * dn / fluxMag0 

1402 

1403 def dn2mag(self, dn, fluxMag0): 

1404 with np.warnings.catch_warnings(): 

1405 np.warnings.filterwarnings('ignore', r'invalid value encountered') 

1406 np.warnings.filterwarnings('ignore', r'divide by zero') 

1407 return -2.5 * np.log10(dn/fluxMag0) 

1408 

1409 def dn2fluxErr(self, dn, dnErr, fluxMag0, fluxMag0Err): 

1410 retVal = self.vhypot(dn * fluxMag0Err, dnErr * fluxMag0) 

1411 retVal *= self.AB_FLUX_SCALE / fluxMag0 / fluxMag0 

1412 return retVal 

1413 

1414 def dn2MagErr(self, dn, dnErr, fluxMag0, fluxMag0Err): 

1415 retVal = self.dn2fluxErr(dn, dnErr, fluxMag0, fluxMag0Err) / self.dn2flux(dn, fluxMag0) 

1416 return self.FIVE_OVER_2LOG10 * retVal 

1417 

1418 

1419class NanoJansky(Photometry): 

1420 def _func(self, df): 

1421 return self.dn2flux(df[self.col], self.fluxMag0) 

1422 

1423 

1424class NanoJanskyErr(Photometry): 

1425 @property 

1426 def columns(self): 

1427 return [self.col, self.colFluxErr] 

1428 

1429 def _func(self, df): 

1430 retArr = self.dn2fluxErr(df[self.col], df[self.colFluxErr], self.fluxMag0, self.fluxMag0Err) 

1431 return pd.Series(retArr, index=df.index) 

1432 

1433 

1434class Magnitude(Photometry): 

1435 def _func(self, df): 

1436 return self.dn2mag(df[self.col], self.fluxMag0) 

1437 

1438 

1439class MagnitudeErr(Photometry): 

1440 @property 

1441 def columns(self): 

1442 return [self.col, self.colFluxErr] 

1443 

1444 def _func(self, df): 

1445 retArr = self.dn2MagErr(df[self.col], df[self.colFluxErr], self.fluxMag0, self.fluxMag0Err) 

1446 return pd.Series(retArr, index=df.index) 

1447 

1448 

1449class LocalPhotometry(Functor): 

1450 """Base class for calibrating the specified instrument flux column using 

1451 the local photometric calibration. 

1452 

1453 Parameters 

1454 ---------- 

1455 instFluxCol : `str` 

1456 Name of the instrument flux column. 

1457 instFluxErrCol : `str` 

1458 Name of the assocated error columns for ``instFluxCol``. 

1459 photoCalibCol : `str` 

1460 Name of local calibration column. 

1461 photoCalibErrCol : `str` 

1462 Error associated with ``photoCalibCol`` 

1463 

1464 See also 

1465 -------- 

1466 LocalPhotometry 

1467 LocalNanojansky 

1468 LocalNanojanskyErr 

1469 LocalMagnitude 

1470 LocalMagnitudeErr 

1471 """ 

1472 logNJanskyToAB = (1 * u.nJy).to_value(u.ABmag) 

1473 

1474 def __init__(self, 

1475 instFluxCol, 

1476 instFluxErrCol, 

1477 photoCalibCol, 

1478 photoCalibErrCol, 

1479 **kwargs): 

1480 self.instFluxCol = instFluxCol 

1481 self.instFluxErrCol = instFluxErrCol 

1482 self.photoCalibCol = photoCalibCol 

1483 self.photoCalibErrCol = photoCalibErrCol 

1484 super().__init__(**kwargs) 

1485 

1486 def instFluxToNanojansky(self, instFlux, localCalib): 

1487 """Convert instrument flux to nanojanskys. 

1488 

1489 Parameters 

1490 ---------- 

1491 instFlux : `numpy.ndarray` or `pandas.Series` 

1492 Array of instrument flux measurements 

1493 localCalib : `numpy.ndarray` or `pandas.Series` 

1494 Array of local photometric calibration estimates. 

1495 

1496 Returns 

1497 ------- 

1498 calibFlux : `numpy.ndarray` or `pandas.Series` 

1499 Array of calibrated flux measurements. 

1500 """ 

1501 return instFlux * localCalib 

1502 

1503 def instFluxErrToNanojanskyErr(self, instFlux, instFluxErr, localCalib, localCalibErr): 

1504 """Convert instrument flux to nanojanskys. 

1505 

1506 Parameters 

1507 ---------- 

1508 instFlux : `numpy.ndarray` or `pandas.Series` 

1509 Array of instrument flux measurements 

1510 instFluxErr : `numpy.ndarray` or `pandas.Series` 

1511 Errors on associated ``instFlux`` values 

1512 localCalib : `numpy.ndarray` or `pandas.Series` 

1513 Array of local photometric calibration estimates. 

1514 localCalibErr : `numpy.ndarray` or `pandas.Series` 

1515 Errors on associated ``localCalib`` values 

1516 

1517 Returns 

1518 ------- 

1519 calibFluxErr : `numpy.ndarray` or `pandas.Series` 

1520 Errors on calibrated flux measurements. 

1521 """ 

1522 return np.hypot(instFluxErr * localCalib, instFlux * localCalibErr) 

1523 

1524 def instFluxToMagnitude(self, instFlux, localCalib): 

1525 """Convert instrument flux to nanojanskys. 

1526 

1527 Parameters 

1528 ---------- 

1529 instFlux : `numpy.ndarray` or `pandas.Series` 

1530 Array of instrument flux measurements 

1531 localCalib : `numpy.ndarray` or `pandas.Series` 

1532 Array of local photometric calibration estimates. 

1533 

1534 Returns 

1535 ------- 

1536 calibMag : `numpy.ndarray` or `pandas.Series` 

1537 Array of calibrated AB magnitudes. 

1538 """ 

1539 return -2.5 * np.log10(self.instFluxToNanojansky(instFlux, localCalib)) + self.logNJanskyToAB 

1540 

1541 def instFluxErrToMagnitudeErr(self, instFlux, instFluxErr, localCalib, localCalibErr): 

1542 """Convert instrument flux err to nanojanskys. 

1543 

1544 Parameters 

1545 ---------- 

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

1547 Array of instrument flux measurements 

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

1549 Errors on associated ``instFlux`` values 

1550 localCalib : `numpy.ndarray` or `pandas.Series` 

1551 Array of local photometric calibration estimates. 

1552 localCalibErr : `numpy.ndarray` or `pandas.Series` 

1553 Errors on associated ``localCalib`` values 

1554 

1555 Returns 

1556 ------- 

1557 calibMagErr: `numpy.ndarray` or `pandas.Series` 

1558 Error on calibrated AB magnitudes. 

1559 """ 

1560 err = self.instFluxErrToNanojanskyErr(instFlux, instFluxErr, localCalib, localCalibErr) 

1561 return 2.5 / np.log(10) * err / self.instFluxToNanojansky(instFlux, instFluxErr) 

1562 

1563 

1564class LocalNanojansky(LocalPhotometry): 

1565 """Compute calibrated fluxes using the local calibration value. 

1566 

1567 See also 

1568 -------- 

1569 LocalNanojansky 

1570 LocalNanojanskyErr 

1571 LocalMagnitude 

1572 LocalMagnitudeErr 

1573 """ 

1574 

1575 @property 

1576 def columns(self): 

1577 return [self.instFluxCol, self.photoCalibCol] 

1578 

1579 @property 

1580 def name(self): 

1581 return f'flux_{self.instFluxCol}' 

1582 

1583 def _func(self, df): 

1584 return self.instFluxToNanojansky(df[self.instFluxCol], df[self.photoCalibCol]) 

1585 

1586 

1587class LocalNanojanskyErr(LocalPhotometry): 

1588 """Compute calibrated flux errors using the local calibration value. 

1589 

1590 See also 

1591 -------- 

1592 LocalNanojansky 

1593 LocalNanojanskyErr 

1594 LocalMagnitude 

1595 LocalMagnitudeErr 

1596 """ 

1597 

1598 @property 

1599 def columns(self): 

1600 return [self.instFluxCol, self.instFluxErrCol, 

1601 self.photoCalibCol, self.photoCalibErrCol] 

1602 

1603 @property 

1604 def name(self): 

1605 return f'fluxErr_{self.instFluxCol}' 

1606 

1607 def _func(self, df): 

1608 return self.instFluxErrToNanojanskyErr(df[self.instFluxCol], df[self.instFluxErrCol], 

1609 df[self.photoCalibCol], df[self.photoCalibErrCol]) 

1610 

1611 

1612class LocalMagnitude(LocalPhotometry): 

1613 """Compute calibrated AB magnitudes using the local calibration value. 

1614 

1615 See also 

1616 -------- 

1617 LocalNanojansky 

1618 LocalNanojanskyErr 

1619 LocalMagnitude 

1620 LocalMagnitudeErr 

1621 """ 

1622 

1623 @property 

1624 def columns(self): 

1625 return [self.instFluxCol, self.photoCalibCol] 

1626 

1627 @property 

1628 def name(self): 

1629 return f'mag_{self.instFluxCol}' 

1630 

1631 def _func(self, df): 

1632 return self.instFluxToMagnitude(df[self.instFluxCol], 

1633 df[self.photoCalibCol]) 

1634 

1635 

1636class LocalMagnitudeErr(LocalPhotometry): 

1637 """Compute calibrated AB magnitude errors using the local calibration value. 

1638 

1639 See also 

1640 -------- 

1641 LocalNanojansky 

1642 LocalNanojanskyErr 

1643 LocalMagnitude 

1644 LocalMagnitudeErr 

1645 """ 

1646 

1647 @property 

1648 def columns(self): 

1649 return [self.instFluxCol, self.instFluxErrCol, 

1650 self.photoCalibCol, self.photoCalibErrCol] 

1651 

1652 @property 

1653 def name(self): 

1654 return f'magErr_{self.instFluxCol}' 

1655 

1656 def _func(self, df): 

1657 return self.instFluxErrToMagnitudeErr(df[self.instFluxCol], 

1658 df[self.instFluxErrCol], 

1659 df[self.photoCalibCol], 

1660 df[self.photoCalibErrCol]) 

1661 

1662 

1663class LocalDipoleMeanFlux(LocalPhotometry): 

1664 """Compute absolute mean of dipole fluxes. 

1665 

1666 See also 

1667 -------- 

1668 LocalNanojansky 

1669 LocalNanojanskyErr 

1670 LocalMagnitude 

1671 LocalMagnitudeErr 

1672 LocalDipoleMeanFlux 

1673 LocalDipoleMeanFluxErr 

1674 LocalDipoleDiffFlux 

1675 LocalDipoleDiffFluxErr 

1676 """ 

1677 def __init__(self, 

1678 instFluxPosCol, 

1679 instFluxNegCol, 

1680 instFluxPosErrCol, 

1681 instFluxNegErrCol, 

1682 photoCalibCol, 

1683 photoCalibErrCol, 

1684 **kwargs): 

1685 self.instFluxNegCol = instFluxNegCol 

1686 self.instFluxPosCol = instFluxPosCol 

1687 self.instFluxNegErrCol = instFluxNegErrCol 

1688 self.instFluxPosErrCol = instFluxPosErrCol 

1689 self.photoCalibCol = photoCalibCol 

1690 self.photoCalibErrCol = photoCalibErrCol 

1691 super().__init__(instFluxNegCol, 

1692 instFluxNegErrCol, 

1693 photoCalibCol, 

1694 photoCalibErrCol, 

1695 **kwargs) 

1696 

1697 @property 

1698 def columns(self): 

1699 return [self.instFluxPosCol, 

1700 self.instFluxNegCol, 

1701 self.photoCalibCol] 

1702 

1703 @property 

1704 def name(self): 

1705 return f'dipMeanFlux_{self.instFluxPosCol}_{self.instFluxNegCol}' 

1706 

1707 def _func(self, df): 

1708 return 0.5*(np.fabs(self.instFluxToNanojansky(df[self.instFluxNegCol], df[self.photoCalibCol])) 

1709 + np.fabs(self.instFluxToNanojansky(df[self.instFluxPosCol], df[self.photoCalibCol]))) 

1710 

1711 

1712class LocalDipoleMeanFluxErr(LocalDipoleMeanFlux): 

1713 """Compute the error on the absolute mean of dipole fluxes. 

1714 

1715 See also 

1716 -------- 

1717 LocalNanojansky 

1718 LocalNanojanskyErr 

1719 LocalMagnitude 

1720 LocalMagnitudeErr 

1721 LocalDipoleMeanFlux 

1722 LocalDipoleMeanFluxErr 

1723 LocalDipoleDiffFlux 

1724 LocalDipoleDiffFluxErr 

1725 """ 

1726 

1727 @property 

1728 def columns(self): 

1729 return [self.instFluxPosCol, 

1730 self.instFluxNegCol, 

1731 self.instFluxPosErrCol, 

1732 self.instFluxNegErrCol, 

1733 self.photoCalibCol, 

1734 self.photoCalibErrCol] 

1735 

1736 @property 

1737 def name(self): 

1738 return f'dipMeanFluxErr_{self.instFluxPosCol}_{self.instFluxNegCol}' 

1739 

1740 def _func(self, df): 

1741 return 0.5*np.sqrt( 

1742 (np.fabs(df[self.instFluxNegCol]) + np.fabs(df[self.instFluxPosCol]) 

1743 * df[self.photoCalibErrCol])**2 

1744 + (df[self.instFluxNegErrCol]**2 + df[self.instFluxPosErrCol]**2) 

1745 * df[self.photoCalibCol]**2) 

1746 

1747 

1748class LocalDipoleDiffFlux(LocalDipoleMeanFlux): 

1749 """Compute the absolute difference of dipole fluxes. 

1750 

1751 Value is (abs(pos) - abs(neg)) 

1752 

1753 See also 

1754 -------- 

1755 LocalNanojansky 

1756 LocalNanojanskyErr 

1757 LocalMagnitude 

1758 LocalMagnitudeErr 

1759 LocalDipoleMeanFlux 

1760 LocalDipoleMeanFluxErr 

1761 LocalDipoleDiffFlux 

1762 LocalDipoleDiffFluxErr 

1763 """ 

1764 

1765 @property 

1766 def columns(self): 

1767 return [self.instFluxPosCol, 

1768 self.instFluxNegCol, 

1769 self.photoCalibCol] 

1770 

1771 @property 

1772 def name(self): 

1773 return f'dipDiffFlux_{self.instFluxPosCol}_{self.instFluxNegCol}' 

1774 

1775 def _func(self, df): 

1776 return (np.fabs(self.instFluxToNanojansky(df[self.instFluxPosCol], df[self.photoCalibCol])) 

1777 - np.fabs(self.instFluxToNanojansky(df[self.instFluxNegCol], df[self.photoCalibCol]))) 

1778 

1779 

1780class LocalDipoleDiffFluxErr(LocalDipoleMeanFlux): 

1781 """Compute the error on the absolute difference of dipole fluxes. 

1782 

1783 See also 

1784 -------- 

1785 LocalNanojansky 

1786 LocalNanojanskyErr 

1787 LocalMagnitude 

1788 LocalMagnitudeErr 

1789 LocalDipoleMeanFlux 

1790 LocalDipoleMeanFluxErr 

1791 LocalDipoleDiffFlux 

1792 LocalDipoleDiffFluxErr 

1793 """ 

1794 

1795 @property 

1796 def columns(self): 

1797 return [self.instFluxPosCol, 

1798 self.instFluxNegCol, 

1799 self.instFluxPosErrCol, 

1800 self.instFluxNegErrCol, 

1801 self.photoCalibCol, 

1802 self.photoCalibErrCol] 

1803 

1804 @property 

1805 def name(self): 

1806 return f'dipDiffFluxErr_{self.instFluxPosCol}_{self.instFluxNegCol}' 

1807 

1808 def _func(self, df): 

1809 return np.sqrt( 

1810 ((np.fabs(df[self.instFluxPosCol]) - np.fabs(df[self.instFluxNegCol])) 

1811 * df[self.photoCalibErrCol])**2 

1812 + (df[self.instFluxPosErrCol]**2 + df[self.instFluxNegErrCol]**2) 

1813 * df[self.photoCalibCol]**2) 

1814 

1815 

1816class Ratio(Functor): 

1817 """Base class for returning the ratio of 2 columns. 

1818 

1819 Can be used to compute a Signal to Noise ratio for any input flux. 

1820 

1821 Parameters 

1822 ---------- 

1823 numerator : `str` 

1824 Name of the column to use at the numerator in the ratio 

1825 denominator : `str` 

1826 Name of the column to use as the denominator in the ratio. 

1827 """ 

1828 def __init__(self, 

1829 numerator, 

1830 denominator, 

1831 **kwargs): 

1832 self.numerator = numerator 

1833 self.denominator = denominator 

1834 super().__init__(**kwargs) 

1835 

1836 @property 

1837 def columns(self): 

1838 return [self.numerator, self.denominator] 

1839 

1840 @property 

1841 def name(self): 

1842 return f'ratio_{self.numerator}_{self.denominator}' 

1843 

1844 def _func(self, df): 

1845 with np.warnings.catch_warnings(): 

1846 np.warnings.filterwarnings('ignore', r'invalid value encountered') 

1847 np.warnings.filterwarnings('ignore', r'divide by zero') 

1848 return df[self.numerator] / df[self.denominator]