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

1import yaml 

2import re 

3from itertools import product 

4 

5import pandas as pd 

6import numpy as np 

7import astropy.units as u 

8 

9from lsst.daf.persistence import doImport 

10from lsst.daf.butler import DeferredDatasetHandle 

11from .parquetTable import ParquetTable, MultilevelParquetTable 

12 

13 

14def init_fromDict(initDict, basePath='lsst.pipe.tasks.functors', 

15 typeKey='functor', name=None): 

16 """Initialize an object defined in a dictionary 

17 

18 The object needs to be importable as 

19 f'{basePath}.{initDict[typeKey]}' 

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

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

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

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

24 

25 Parameters 

26 ---------- 

27 initDict : dictionary 

28 Dictionary describing object's initialization. Must contain 

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

30 relative to ``basePath``. 

31 basePath : str 

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

33 typeKey : str 

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

35 (relative to `basePath`). 

36 """ 

37 initDict = initDict.copy() 

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

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

40 args = [] 

41 if 'args' in initDict: 

42 args = initDict.pop('args') 

43 if isinstance(args, str): 

44 args = [args] 

45 try: 

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

47 except Exception as e: 

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

49 raise type(e)(message, e.args) 

50 return element 

51 

52 

53class Functor(object): 

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

55 

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

57 `DeferredDatasetHandle`, and returns the 

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

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

60 from the `ParquetTable`. 

61 

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

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

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

65 result. 

66 

67 

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

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

70 In addition, it must define the following attributes 

71 

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

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

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

75 

76 On initialization, a `Functor` should declare what filter (`filt` kwarg) 

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

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

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

80 on the `_defaultDataset`attribute. If filter is not specified and `dataset` 

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

82 perform the calculation. 

83 

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

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

86 dataframe with a multi-level column index, 

87 with the levels of the column index being `filter`, 

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

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

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

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

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

93 `column`. 

94 

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

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

97 to do the calculation. 

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

99 which `_dfLevels = ('filter', 'column')`, and `_func` expects the dataframe 

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

101 

102 Parameters 

103 ---------- 

104 filt : str 

105 Filter upon which to do the calculation 

106 

107 dataset : str 

108 Dataset upon which to do the calculation 

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

110 

111 """ 

112 

113 _defaultDataset = 'ref' 

114 _columnLevels = ('filter', 'dataset', 'column') 

115 _dfLevels = ('column',) 

116 _defaultNoDup = False 

117 

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

119 self.filt = filt 

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

121 self._noDup = noDup 

122 

123 @property 

124 def noDup(self): 

125 if self._noDup is not None: 

126 return self._noDup 

127 else: 

128 return self._defaultNoDup 

129 

130 @property 

131 def columns(self): 

132 """Columns required to perform calculation 

133 """ 

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

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

136 return self._columns 

137 

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

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

140 

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

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

143 and with the gen3 `DeferredDatasetHandle`. 

144 

145 Parameters 

146 ---------- 

147 data : `MultilevelParquetTable` or `DeferredDatasetHandle` 

148 

149 columnnIndex (optional): pandas `Index` object 

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

151 """ 

152 if isinstance(data, DeferredDatasetHandle): 

153 if columnIndex is None: 

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

155 if columnIndex is not None: 

156 return columnIndex.names 

157 if isinstance(data, MultilevelParquetTable): 

158 return data.columnLevels 

159 else: 

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

161 

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

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

164 

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

166 

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

168 """ 

169 if isinstance(data, DeferredDatasetHandle): 

170 if columnIndex is None: 

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

172 if columnIndex is not None: 

173 columnLevels = columnIndex.names 

174 columnLevelNames = { 

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

176 for i, level in enumerate(columnLevels) 

177 } 

178 return columnLevelNames 

179 if isinstance(data, MultilevelParquetTable): 

180 return data.columnLevelNames 

181 else: 

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

183 

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

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

186 

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

188 """ 

189 new_colDict = {} 

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

191 

192 for i, l in enumerate(columnLevels): 

193 if l in colDict: 

194 if isinstance(colDict[l], str): 

195 new_colDict[l] = [colDict[l]] 

196 else: 

197 new_colDict[l] = colDict[l] 

198 else: 

199 new_colDict[l] = columnIndex.levels[i] 

200 

201 levelCols = [new_colDict[l] for l in columnLevels] 

202 cols = product(*levelCols) 

203 return list(cols) 

204 

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

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

207 

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

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

210 dictionary. 

211 

212 Parameters 

213 ---------- 

214 data : `MultilevelParquetTable` or `DeferredDatasetHandle` 

215 

216 columnIndex (optional): pandas `Index` object 

217 either passed or read in from `DeferredDatasetHandle`. 

218 

219 `returnTuple` : bool 

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

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

222 combine columns from the various component functors. 

223 

224 """ 

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

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

227 

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

229 columnLevels = self._get_data_columnLevels(data, columnIndex) 

230 

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

232 raise ValueError( 

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

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

235 ) 

236 

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

238 'dataset': self.dataset} 

239 if self.filt is None: 

240 columnLevelNames = self._get_data_columnLevelNames(data, columnIndex) 

241 if "filter" in columnLevels: 

242 if self.dataset == "ref": 

243 columnDict["filter"] = columnLevelNames["filter"][0] 

244 else: 

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

246 f"(dataset {self.dataset}) " 

247 "and ParquetTable " 

248 "contains multiple filters in column index. " 

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

250 else: 

251 columnDict['filter'] = self.filt 

252 

253 if isinstance(data, MultilevelParquetTable): 

254 return data._colsFromDict(columnDict) 

255 elif isinstance(data, DeferredDatasetHandle): 

256 if returnTuple: 

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

258 else: 

259 return columnDict 

260 

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

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

263 

264 def _get_columnIndex(self, data): 

265 """Return columnIndex 

266 """ 

267 

268 if isinstance(data, DeferredDatasetHandle): 

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

270 else: 

271 return None 

272 

273 def _get_data(self, data): 

274 """Retrieve dataframe necessary for calculation. 

275 

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

277 

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

279 

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

281 """ 

282 if isinstance(data, pd.DataFrame): 

283 return data 

284 

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

286 columnIndex = self._get_columnIndex(data) 

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

288 

289 # Simple single-level parquet table, gen2 

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

291 columns = self.columns 

292 df = data.toDataFrame(columns=columns) 

293 return df 

294 

295 # Get proper multi-level columns specification for this functor 

296 if is_multiLevel: 

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

298 

299 if isinstance(data, MultilevelParquetTable): 

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

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

302 elif isinstance(data, DeferredDatasetHandle): 

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

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

305 

306 # Drop unnecessary column levels 

307 df = self._setLevels(df) 

308 return df 

309 

310 def _setLevels(self, df): 

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

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

313 return df 

314 

315 def _dropna(self, vals): 

316 return vals.dropna() 

317 

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

319 try: 

320 df = self._get_data(data) 

321 vals = self._func(df) 

322 except Exception: 

323 vals = self.fail(df) 

324 if dropna: 

325 vals = self._dropna(vals) 

326 

327 return vals 

328 

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

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

331 """ 

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

333 

334 def fail(self, df): 

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

336 

337 @property 

338 def name(self): 

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

340 """ 

341 return NotImplementedError 

342 

343 @property 

344 def shortname(self): 

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

346 """ 

347 return self.name 

348 

349 

350class CompositeFunctor(Functor): 

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

352 

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

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

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

356 being the keys of `funcDict`. 

357 

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

359 in all the component functors. 

360 

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

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

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

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

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

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

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

368 

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

370 file specifying a collection of functors. 

371 

372 Parameters 

373 ---------- 

374 funcs : `dict` or `list` 

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

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

377 

378 """ 

379 dataset = None 

380 

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

382 

383 if type(funcs) == dict: 

384 self.funcDict = funcs 

385 else: 

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

387 

388 self._filt = None 

389 

390 super().__init__(**kwargs) 

391 

392 @property 

393 def filt(self): 

394 return self._filt 

395 

396 @filt.setter 

397 def filt(self, filt): 

398 if filt is not None: 

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

400 f.filt = filt 

401 self._filt = filt 

402 

403 def update(self, new): 

404 if isinstance(new, dict): 

405 self.funcDict.update(new) 

406 elif isinstance(new, CompositeFunctor): 

407 self.funcDict.update(new.funcDict) 

408 else: 

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

410 

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

412 if self.filt is not None: 

413 self.filt = self.filt 

414 

415 @property 

416 def columns(self): 

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

418 

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

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

421 return list( 

422 set( 

423 [ 

424 x 

425 for y in [ 

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

427 ] 

428 for x in y 

429 ] 

430 ) 

431 ) 

432 

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

434 columnIndex = self._get_columnIndex(data) 

435 

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

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

438 

439 # Simple single-level column index, gen2 

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

441 columns = self.columns 

442 df = data.toDataFrame(columns=columns) 

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

444 

445 # Multilevel index, gen2 or gen3 

446 if is_multiLevel: 

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

448 

449 if isinstance(data, MultilevelParquetTable): 

450 # Read data into memory the gen2 way 

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

452 elif isinstance(data, DeferredDatasetHandle): 

453 # Read data into memory the gen3 way 

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

455 

456 valDict = {} 

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

458 try: 

459 subdf = f._setLevels( 

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

461 ) 

462 valDict[k] = f._func(subdf) 

463 except Exception: 

464 valDict[k] = f.fail(subdf) 

465 

466 # non-multilevel, gen3 (TODO: this should work, but this case is not tested in test_functors.py) 

467 elif isinstance(data, DeferredDatasetHandle): 

468 columns = self.columns 

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

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

471 

472 try: 

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

474 except TypeError: 

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

476 raise 

477 

478 if kwargs.get('dropna', False): 

479 valDf = valDf.dropna(how='any') 

480 

481 return valDf 

482 

483 @classmethod 

484 def renameCol(cls, col, renameRules): 

485 if renameRules is None: 

486 return col 

487 for old, new in renameRules: 

488 if col.startswith(old): 

489 col = col.replace(old, new) 

490 return col 

491 

492 @classmethod 

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

494 with open(filename) as f: 

495 translationDefinition = yaml.safe_load(f) 

496 

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

498 

499 @classmethod 

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

501 funcs = {} 

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

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

504 

505 if 'flag_rename_rules' in translationDefinition: 

506 renameRules = translationDefinition['flag_rename_rules'] 

507 else: 

508 renameRules = None 

509 

510 if 'refFlags' in translationDefinition: 

511 for flag in translationDefinition['refFlags']: 

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

513 

514 if 'flags' in translationDefinition: 

515 for flag in translationDefinition['flags']: 

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

517 

518 return cls(funcs, **kwargs) 

519 

520 

521def mag_aware_eval(df, expr): 

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

523 

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

525 

526 Parameters 

527 ---------- 

528 df : pandas.DataFrame 

529 Dataframe on which to evaluate expression. 

530 

531 expr : str 

532 Expression. 

533 """ 

534 try: 

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

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

537 except Exception: # Should check what actually gets raised 

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

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

540 return val 

541 

542 

543class CustomFunctor(Functor): 

544 """Arbitrary computation on a catalog 

545 

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

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

548 

549 Parameters 

550 ---------- 

551 expr : str 

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

553 """ 

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

555 

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

557 self.expr = expr 

558 super().__init__(**kwargs) 

559 

560 @property 

561 def name(self): 

562 return self.expr 

563 

564 @property 

565 def columns(self): 

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

567 

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

569 not_a_col = [] 

570 for c in flux_cols: 

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

572 cols.append(f'{c}_instFlux') 

573 not_a_col.append(c) 

574 else: 

575 cols.append(c) 

576 

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

578 

579 def _func(self, df): 

580 return mag_aware_eval(df, self.expr) 

581 

582 

583class Column(Functor): 

584 """Get column with specified name 

585 """ 

586 

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

588 self.col = col 

589 super().__init__(**kwargs) 

590 

591 @property 

592 def name(self): 

593 return self.col 

594 

595 @property 

596 def columns(self): 

597 return [self.col] 

598 

599 def _func(self, df): 

600 return df[self.col] 

601 

602 

603class Index(Functor): 

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

605 """ 

606 

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

608 _defaultDataset = 'ref' 

609 _defaultNoDup = True 

610 

611 def _func(self, df): 

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

613 

614 

615class IDColumn(Column): 

616 col = 'id' 

617 _allow_difference = False 

618 _defaultNoDup = True 

619 

620 def _func(self, df): 

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

622 

623 

624class FootprintNPix(Column): 

625 col = 'base_Footprint_nPix' 

626 

627 

628class CoordColumn(Column): 

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

630 """ 

631 _radians = True 

632 

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

634 super().__init__(col, **kwargs) 

635 

636 def _func(self, df): 

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

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

639 return output 

640 

641 

642class RAColumn(CoordColumn): 

643 """Right Ascension, in degrees 

644 """ 

645 name = 'RA' 

646 _defaultNoDup = True 

647 

648 def __init__(self, **kwargs): 

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

650 

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

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

653 

654 

655class DecColumn(CoordColumn): 

656 """Declination, in degrees 

657 """ 

658 name = 'Dec' 

659 _defaultNoDup = True 

660 

661 def __init__(self, **kwargs): 

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

663 

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

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

666 

667 

668def fluxName(col): 

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

670 col += '_instFlux' 

671 return col 

672 

673 

674def fluxErrName(col): 

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

676 col += '_instFluxErr' 

677 return col 

678 

679 

680class Mag(Functor): 

681 """Compute calibrated magnitude 

682 

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

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

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

686 This default should be removed in DM-21955 

687 

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

689 

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

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

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

693 

694 Parameters 

695 ---------- 

696 col : `str` 

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

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

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

700 understand. 

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

702 Object that knows zero point. 

703 """ 

704 _defaultDataset = 'meas' 

705 

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

707 self.col = fluxName(col) 

708 self.calib = calib 

709 if calib is not None: 

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

711 else: 

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

713 self.fluxMag0 = 63095734448.0194 

714 

715 super().__init__(**kwargs) 

716 

717 @property 

718 def columns(self): 

719 return [self.col] 

720 

721 def _func(self, df): 

722 with np.warnings.catch_warnings(): 

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

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

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

726 

727 @property 

728 def name(self): 

729 return f'mag_{self.col}' 

730 

731 

732class MagErr(Mag): 

733 """Compute calibrated magnitude uncertainty 

734 

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

736 

737 Parameters 

738 col : `str` 

739 Name of flux column 

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

741 Object that knows zero point. 

742 """ 

743 

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

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

746 if self.calib is not None: 

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

748 else: 

749 self.fluxMag0Err = 0. 

750 

751 @property 

752 def columns(self): 

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

754 

755 def _func(self, df): 

756 with np.warnings.catch_warnings(): 

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

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

759 fluxCol, fluxErrCol = self.columns 

760 x = df[fluxErrCol] / df[fluxCol] 

761 y = self.fluxMag0Err / self.fluxMag0 

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

763 return magErr 

764 

765 @property 

766 def name(self): 

767 return super().name + '_err' 

768 

769 

770class NanoMaggie(Mag): 

771 """ 

772 """ 

773 

774 def _func(self, df): 

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

776 

777 

778class MagDiff(Functor): 

779 _defaultDataset = 'meas' 

780 

781 """Functor to calculate magnitude difference""" 

782 

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

784 self.col1 = fluxName(col1) 

785 self.col2 = fluxName(col2) 

786 super().__init__(**kwargs) 

787 

788 @property 

789 def columns(self): 

790 return [self.col1, self.col2] 

791 

792 def _func(self, df): 

793 with np.warnings.catch_warnings(): 

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

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

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

797 

798 @property 

799 def name(self): 

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

801 

802 @property 

803 def shortname(self): 

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

805 

806 

807class Color(Functor): 

808 """Compute the color between two filters 

809 

810 Computes color by initializing two different `Mag` 

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

812 then returning the difference. 

813 

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

815 multilevel column index, with both `'filter'` and `'column'`, 

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

817 This is controlled by the `_dfLevels` attribute. 

818 

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

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

821 

822 Parameters 

823 ---------- 

824 col : str 

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

826 `lsst.pipe.tasks.functors.Mag`. 

827 

828 filt2, filt1 : str 

829 Filters from which to compute magnitude difference. 

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

831 """ 

832 _defaultDataset = 'forced_src' 

833 _dfLevels = ('filter', 'column') 

834 _defaultNoDup = True 

835 

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

837 self.col = fluxName(col) 

838 if filt2 == filt1: 

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

840 self.filt2 = filt2 

841 self.filt1 = filt1 

842 

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

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

845 

846 super().__init__(**kwargs) 

847 

848 @property 

849 def filt(self): 

850 return None 

851 

852 @filt.setter 

853 def filt(self, filt): 

854 pass 

855 

856 def _func(self, df): 

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

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

859 return mag2 - mag1 

860 

861 @property 

862 def columns(self): 

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

864 

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

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

867 

868 @property 

869 def name(self): 

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

871 

872 @property 

873 def shortname(self): 

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

875 

876 

877class Labeller(Functor): 

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

879 """ 

880 _null_label = 'null' 

881 _allow_difference = False 

882 name = 'label' 

883 _force_str = False 

884 

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

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

887 

888 

889class StarGalaxyLabeller(Labeller): 

890 _columns = ["base_ClassificationExtendedness_value"] 

891 _column = "base_ClassificationExtendedness_value" 

892 

893 def _func(self, df): 

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

895 mask = x.isnull() 

896 test = (x < 0.5).astype(int) 

897 test = test.mask(mask, 2) 

898 

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

900 # are these backwards? 

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

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

903 index=x.index, name='label') 

904 if self._force_str: 

905 label = label.astype(str) 

906 return label 

907 

908 

909class NumStarLabeller(Labeller): 

910 _columns = ['numStarFlags'] 

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

912 

913 def _func(self, df): 

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

915 

916 # Number of filters 

917 n = len(x.unique()) - 1 

918 

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

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

921 index=x.index, name='label') 

922 

923 if self._force_str: 

924 label = label.astype(str) 

925 

926 return label 

927 

928 

929class DeconvolvedMoments(Functor): 

930 name = 'Deconvolved Moments' 

931 shortname = 'deconvolvedMoments' 

932 _columns = ("ext_shapeHSM_HsmSourceMoments_xx", 

933 "ext_shapeHSM_HsmSourceMoments_yy", 

934 "base_SdssShape_xx", "base_SdssShape_yy", 

935 "ext_shapeHSM_HsmPsfMoments_xx", 

936 "ext_shapeHSM_HsmPsfMoments_yy") 

937 

938 def _func(self, df): 

939 """Calculate deconvolved moments""" 

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

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

942 else: 

943 hsm = np.ones(len(df))*np.nan 

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

945 if "ext_shapeHSM_HsmPsfMoments_xx" in df.columns: 

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

947 else: 

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

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

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

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

952 

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

954 

955 

956class SdssTraceSize(Functor): 

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

958 name = "SDSS Trace Size" 

959 shortname = 'sdssTrace' 

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

961 

962 def _func(self, df): 

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

964 return srcSize 

965 

966 

967class PsfSdssTraceSizeDiff(Functor): 

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

969 name = "PSF - SDSS Trace Size" 

970 shortname = 'psf_sdssTrace' 

971 _columns = ("base_SdssShape_xx", "base_SdssShape_yy", 

972 "base_SdssShape_psf_xx", "base_SdssShape_psf_yy") 

973 

974 def _func(self, df): 

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

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

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

978 return sizeDiff 

979 

980 

981class HsmTraceSize(Functor): 

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

983 name = 'HSM Trace Size' 

984 shortname = 'hsmTrace' 

985 _columns = ("ext_shapeHSM_HsmSourceMoments_xx", 

986 "ext_shapeHSM_HsmSourceMoments_yy") 

987 

988 def _func(self, df): 

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

990 + df["ext_shapeHSM_HsmSourceMoments_yy"])) 

991 return srcSize 

992 

993 

994class PsfHsmTraceSizeDiff(Functor): 

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

996 name = 'PSF - HSM Trace Size' 

997 shortname = 'psf_HsmTrace' 

998 _columns = ("ext_shapeHSM_HsmSourceMoments_xx", 

999 "ext_shapeHSM_HsmSourceMoments_yy", 

1000 "ext_shapeHSM_HsmPsfMoments_xx", 

1001 "ext_shapeHSM_HsmPsfMoments_yy") 

1002 

1003 def _func(self, df): 

1004 srcSize = np.sqrt(0.5*(df["ext_shapeHSM_HsmSourceMoments_xx"] 

1005 + df["ext_shapeHSM_HsmSourceMoments_yy"])) 

1006 psfSize = np.sqrt(0.5*(df["ext_shapeHSM_HsmPsfMoments_xx"] 

1007 + df["ext_shapeHSM_HsmPsfMoments_yy"])) 

1008 sizeDiff = 100*(srcSize - psfSize)/(0.5*(srcSize + psfSize)) 

1009 return sizeDiff 

1010 

1011 

1012class HsmFwhm(Functor): 

1013 name = 'HSM Psf FWHM' 

1014 _columns = ('ext_shapeHSM_HsmPsfMoments_xx', 'ext_shapeHSM_HsmPsfMoments_yy') 

1015 # TODO: DM-21403 pixel scale should be computed from the CD matrix or transform matrix 

1016 pixelScale = 0.168 

1017 SIGMA2FWHM = 2*np.sqrt(2*np.log(2)) 

1018 

1019 def _func(self, df): 

1020 return self.pixelScale*self.SIGMA2FWHM*np.sqrt( 

1021 0.5*(df['ext_shapeHSM_HsmPsfMoments_xx'] + df['ext_shapeHSM_HsmPsfMoments_yy'])) 

1022 

1023 

1024class E1(Functor): 

1025 name = "Distortion Ellipticity (e1)" 

1026 shortname = "Distortion" 

1027 

1028 def __init__(self, colXX, colXY, colYY, **kwargs): 

1029 self.colXX = colXX 

1030 self.colXY = colXY 

1031 self.colYY = colYY 

1032 self._columns = [self.colXX, self.colXY, self.colYY] 

1033 super().__init__(**kwargs) 

1034 

1035 @property 

1036 def columns(self): 

1037 return [self.colXX, self.colXY, self.colYY] 

1038 

1039 def _func(self, df): 

1040 return df[self.colXX] - df[self.colYY] / (df[self.colXX] + df[self.colYY]) 

1041 

1042 

1043class E2(Functor): 

1044 name = "Ellipticity e2" 

1045 

1046 def __init__(self, colXX, colXY, colYY, **kwargs): 

1047 self.colXX = colXX 

1048 self.colXY = colXY 

1049 self.colYY = colYY 

1050 super().__init__(**kwargs) 

1051 

1052 @property 

1053 def columns(self): 

1054 return [self.colXX, self.colXY, self.colYY] 

1055 

1056 def _func(self, df): 

1057 return 2*df[self.colXY] / (df[self.colXX] + df[self.colYY]) 

1058 

1059 

1060class RadiusFromQuadrupole(Functor): 

1061 

1062 def __init__(self, colXX, colXY, colYY, **kwargs): 

1063 self.colXX = colXX 

1064 self.colXY = colXY 

1065 self.colYY = colYY 

1066 super().__init__(**kwargs) 

1067 

1068 @property 

1069 def columns(self): 

1070 return [self.colXX, self.colXY, self.colYY] 

1071 

1072 def _func(self, df): 

1073 return (df[self.colXX]*df[self.colYY] - df[self.colXY]**2)**0.25 

1074 

1075 

1076class LocalWcs(Functor): 

1077 """Computations using the stored localWcs. 

1078 """ 

1079 name = "LocalWcsOperations" 

1080 

1081 def __init__(self, 

1082 colCD_1_1, 

1083 colCD_1_2, 

1084 colCD_2_1, 

1085 colCD_2_2, 

1086 **kwargs): 

1087 self.colCD_1_1 = colCD_1_1 

1088 self.colCD_1_2 = colCD_1_2 

1089 self.colCD_2_1 = colCD_2_1 

1090 self.colCD_2_2 = colCD_2_2 

1091 super().__init__(**kwargs) 

1092 

1093 def computeDeltaRaDec(self, x, y, cd11, cd12, cd21, cd22): 

1094 """Compute the distance on the sphere from x2, y1 to x1, y1. 

1095 

1096 Parameters 

1097 ---------- 

1098 x : `pandas.Series` 

1099 X pixel coordinate. 

1100 y : `pandas.Series` 

1101 Y pixel coordinate. 

1102 cd11 : `pandas.Series` 

1103 [1, 1] element of the local Wcs affine transform. 

1104 cd11 : `pandas.Series` 

1105 [1, 1] element of the local Wcs affine transform. 

1106 cd12 : `pandas.Series` 

1107 [1, 2] element of the local Wcs affine transform. 

1108 cd21 : `pandas.Series` 

1109 [2, 1] element of the local Wcs affine transform. 

1110 cd22 : `pandas.Series` 

1111 [2, 2] element of the local Wcs affine transform. 

1112 

1113 Returns 

1114 ------- 

1115 raDecTuple : tuple 

1116 RA and dec conversion of x and y given the local Wcs. Returned 

1117 units are in radians. 

1118 

1119 """ 

1120 return (x * cd11 + y * cd12, x * cd21 + y * cd22) 

1121 

1122 def computeSkySeperation(self, ra1, dec1, ra2, dec2): 

1123 """Compute the local pixel scale conversion. 

1124 

1125 Parameters 

1126 ---------- 

1127 ra1 : `pandas.Series` 

1128 Ra of the first coordinate in radians. 

1129 dec1 : `pandas.Series` 

1130 Dec of the first coordinate in radians. 

1131 ra2 : `pandas.Series` 

1132 Ra of the second coordinate in radians. 

1133 dec2 : `pandas.Series` 

1134 Dec of the second coordinate in radians. 

1135 

1136 Returns 

1137 ------- 

1138 dist : `pandas.Series` 

1139 Distance on the sphere in radians. 

1140 """ 

1141 deltaDec = dec2 - dec1 

1142 deltaRa = ra2 - ra1 

1143 return 2 * np.arcsin( 

1144 np.sqrt( 

1145 np.sin(deltaDec / 2) ** 2 

1146 + np.cos(dec2) * np.cos(dec1) * np.sin(deltaRa / 2) ** 2)) 

1147 

1148 def getSkySeperationFromPixel(self, x1, y1, x2, y2, cd11, cd12, cd21, cd22): 

1149 """Compute the distance on the sphere from x2, y1 to x1, y1. 

1150 

1151 Parameters 

1152 ---------- 

1153 x1 : `pandas.Series` 

1154 X pixel coordinate. 

1155 y1 : `pandas.Series` 

1156 Y pixel coordinate. 

1157 x2 : `pandas.Series` 

1158 X pixel coordinate. 

1159 y2 : `pandas.Series` 

1160 Y pixel coordinate. 

1161 cd11 : `pandas.Series` 

1162 [1, 1] element of the local Wcs affine transform. 

1163 cd11 : `pandas.Series` 

1164 [1, 1] element of the local Wcs affine transform. 

1165 cd12 : `pandas.Series` 

1166 [1, 2] element of the local Wcs affine transform. 

1167 cd21 : `pandas.Series` 

1168 [2, 1] element of the local Wcs affine transform. 

1169 cd22 : `pandas.Series` 

1170 [2, 2] element of the local Wcs affine transform. 

1171 

1172 Returns 

1173 ------- 

1174 Distance : `pandas.Series` 

1175 Arcseconds per pixel at the location of the local WC 

1176 """ 

1177 ra1, dec1 = self.computeDeltaRaDec(x1, y1, cd11, cd12, cd21, cd22) 

1178 ra2, dec2 = self.computeDeltaRaDec(x2, y2, cd11, cd12, cd21, cd22) 

1179 # Great circle distance for small separations. 

1180 return self.computeSkySeperation(ra1, dec1, ra2, dec2) 

1181 

1182 

1183class ComputePixelScale(LocalWcs): 

1184 """Compute the local pixel scale from the stored CDMatrix. 

1185 """ 

1186 name = "PixelScale" 

1187 

1188 @property 

1189 def columns(self): 

1190 return [self.colCD_1_1, 

1191 self.colCD_1_2, 

1192 self.colCD_2_1, 

1193 self.colCD_2_2] 

1194 

1195 def pixelScaleArcseconds(self, cd11, cd12, cd21, cd22): 

1196 """Compute the local pixel to scale conversion in arcseconds. 

1197 

1198 Parameters 

1199 ---------- 

1200 cd11 : `pandas.Series` 

1201 [1, 1] element of the local Wcs affine transform in radians. 

1202 cd11 : `pandas.Series` 

1203 [1, 1] element of the local Wcs affine transform in radians. 

1204 cd12 : `pandas.Series` 

1205 [1, 2] element of the local Wcs affine transform in radians. 

1206 cd21 : `pandas.Series` 

1207 [2, 1] element of the local Wcs affine transform in radians. 

1208 cd22 : `pandas.Series` 

1209 [2, 2] element of the local Wcs affine transform in radians. 

1210 

1211 Returns 

1212 ------- 

1213 pixScale : `pandas.Series` 

1214 Arcseconds per pixel at the location of the local WC 

1215 """ 

1216 return 3600 * np.degrees(np.sqrt(np.fabs(cd11 * cd22 - cd12 * cd21))) 

1217 

1218 def _func(self, df): 

1219 return self.pixelScaleArcseconds(df[self.colCD_1_1], 

1220 df[self.colCD_1_2], 

1221 df[self.colCD_2_1], 

1222 df[self.colCD_2_2]) 

1223 

1224 

1225class ConvertPixelToArcseconds(ComputePixelScale): 

1226 """Convert a value in units pixels to units arcseconds. 

1227 """ 

1228 

1229 def __init__(self, 

1230 col, 

1231 colCD_1_1, 

1232 colCD_1_2, 

1233 colCD_2_1, 

1234 colCD_2_2, 

1235 **kwargs): 

1236 self.col = col 

1237 super().__init__(colCD_1_1, 

1238 colCD_1_2, 

1239 colCD_2_1, 

1240 colCD_2_2, 

1241 **kwargs) 

1242 

1243 @property 

1244 def name(self): 

1245 return f"{self.col}_asArcseconds" 

1246 

1247 @property 

1248 def columns(self): 

1249 return [self.col, 

1250 self.colCD_1_1, 

1251 self.colCD_1_2, 

1252 self.colCD_2_1, 

1253 self.colCD_2_2] 

1254 

1255 def _func(self, df): 

1256 return df[self.col] * self.pixelScaleArcseconds(df[self.colCD_1_1], 

1257 df[self.colCD_1_2], 

1258 df[self.colCD_2_1], 

1259 df[self.colCD_2_2]) 

1260 

1261 

1262class ReferenceBand(Functor): 

1263 name = 'Reference Band' 

1264 shortname = 'refBand' 

1265 

1266 @property 

1267 def columns(self): 

1268 return ["merge_measurement_i", 

1269 "merge_measurement_r", 

1270 "merge_measurement_z", 

1271 "merge_measurement_y", 

1272 "merge_measurement_g"] 

1273 

1274 def _func(self, df): 

1275 def getFilterAliasName(row): 

1276 # get column name with the max value (True > False) 

1277 colName = row.idxmax() 

1278 return colName.replace('merge_measurement_', '') 

1279 

1280 return df[self.columns].apply(getFilterAliasName, axis=1) 

1281 

1282 

1283class Photometry(Functor): 

1284 # AB to NanoJansky (3631 Jansky) 

1285 AB_FLUX_SCALE = (0 * u.ABmag).to_value(u.nJy) 

1286 LOG_AB_FLUX_SCALE = 12.56 

1287 FIVE_OVER_2LOG10 = 1.085736204758129569 

1288 # TO DO: DM-21955 Replace hard coded photometic calibration values 

1289 COADD_ZP = 27 

1290 

1291 def __init__(self, colFlux, colFluxErr=None, calib=None, **kwargs): 

1292 self.vhypot = np.vectorize(self.hypot) 

1293 self.col = colFlux 

1294 self.colFluxErr = colFluxErr 

1295 

1296 self.calib = calib 

1297 if calib is not None: 

1298 self.fluxMag0, self.fluxMag0Err = calib.getFluxMag0() 

1299 else: 

1300 self.fluxMag0 = 1./np.power(10, -0.4*self.COADD_ZP) 

1301 self.fluxMag0Err = 0. 

1302 

1303 super().__init__(**kwargs) 

1304 

1305 @property 

1306 def columns(self): 

1307 return [self.col] 

1308 

1309 @property 

1310 def name(self): 

1311 return f'mag_{self.col}' 

1312 

1313 @classmethod 

1314 def hypot(cls, a, b): 

1315 if np.abs(a) < np.abs(b): 

1316 a, b = b, a 

1317 if a == 0.: 

1318 return 0. 

1319 q = b/a 

1320 return np.abs(a) * np.sqrt(1. + q*q) 

1321 

1322 def dn2flux(self, dn, fluxMag0): 

1323 return self.AB_FLUX_SCALE * dn / fluxMag0 

1324 

1325 def dn2mag(self, dn, fluxMag0): 

1326 with np.warnings.catch_warnings(): 

1327 np.warnings.filterwarnings('ignore', r'invalid value encountered') 

1328 np.warnings.filterwarnings('ignore', r'divide by zero') 

1329 return -2.5 * np.log10(dn/fluxMag0) 

1330 

1331 def dn2fluxErr(self, dn, dnErr, fluxMag0, fluxMag0Err): 

1332 retVal = self.vhypot(dn * fluxMag0Err, dnErr * fluxMag0) 

1333 retVal *= self.AB_FLUX_SCALE / fluxMag0 / fluxMag0 

1334 return retVal 

1335 

1336 def dn2MagErr(self, dn, dnErr, fluxMag0, fluxMag0Err): 

1337 retVal = self.dn2fluxErr(dn, dnErr, fluxMag0, fluxMag0Err) / self.dn2flux(dn, fluxMag0) 

1338 return self.FIVE_OVER_2LOG10 * retVal 

1339 

1340 

1341class NanoJansky(Photometry): 

1342 def _func(self, df): 

1343 return self.dn2flux(df[self.col], self.fluxMag0) 

1344 

1345 

1346class NanoJanskyErr(Photometry): 

1347 @property 

1348 def columns(self): 

1349 return [self.col, self.colFluxErr] 

1350 

1351 def _func(self, df): 

1352 retArr = self.dn2fluxErr(df[self.col], df[self.colFluxErr], self.fluxMag0, self.fluxMag0Err) 

1353 return pd.Series(retArr, index=df.index) 

1354 

1355 

1356class Magnitude(Photometry): 

1357 def _func(self, df): 

1358 return self.dn2mag(df[self.col], self.fluxMag0) 

1359 

1360 

1361class MagnitudeErr(Photometry): 

1362 @property 

1363 def columns(self): 

1364 return [self.col, self.colFluxErr] 

1365 

1366 def _func(self, df): 

1367 retArr = self.dn2MagErr(df[self.col], df[self.colFluxErr], self.fluxMag0, self.fluxMag0Err) 

1368 return pd.Series(retArr, index=df.index) 

1369 

1370 

1371class LocalPhotometry(Functor): 

1372 """Base class for calibrating the specified instrument flux column using 

1373 the local photometric calibration. 

1374 

1375 Parameters 

1376 ---------- 

1377 instFluxCol : `str` 

1378 Name of the instrument flux column. 

1379 instFluxErrCol : `str` 

1380 Name of the assocated error columns for ``instFluxCol``. 

1381 photoCalibCol : `str` 

1382 Name of local calibration column. 

1383 photoCalibErrCol : `str` 

1384 Error associated with ``photoCalibCol`` 

1385 

1386 See also 

1387 -------- 

1388 LocalPhotometry 

1389 LocalNanojansky 

1390 LocalNanojanskyErr 

1391 LocalMagnitude 

1392 LocalMagnitudeErr 

1393 """ 

1394 logNJanskyToAB = (1 * u.nJy).to_value(u.ABmag) 

1395 

1396 def __init__(self, 

1397 instFluxCol, 

1398 instFluxErrCol, 

1399 photoCalibCol, 

1400 photoCalibErrCol, 

1401 **kwargs): 

1402 self.instFluxCol = instFluxCol 

1403 self.instFluxErrCol = instFluxErrCol 

1404 self.photoCalibCol = photoCalibCol 

1405 self.photoCalibErrCol = photoCalibErrCol 

1406 super().__init__(**kwargs) 

1407 

1408 def instFluxToNanojansky(self, instFlux, localCalib): 

1409 """Convert instrument flux to nanojanskys. 

1410 

1411 Parameters 

1412 ---------- 

1413 instFlux : `numpy.ndarray` or `pandas.Series` 

1414 Array of instrument flux measurements 

1415 localCalib : `numpy.ndarray` or `pandas.Series` 

1416 Array of local photometric calibration estimates. 

1417 

1418 Returns 

1419 ------- 

1420 calibFlux : `numpy.ndarray` or `pandas.Series` 

1421 Array of calibrated flux measurements. 

1422 """ 

1423 return instFlux * localCalib 

1424 

1425 def instFluxErrToNanojanskyErr(self, instFlux, instFluxErr, localCalib, localCalibErr): 

1426 """Convert instrument flux to nanojanskys. 

1427 

1428 Parameters 

1429 ---------- 

1430 instFlux : `numpy.ndarray` or `pandas.Series` 

1431 Array of instrument flux measurements 

1432 instFluxErr : `numpy.ndarray` or `pandas.Series` 

1433 Errors on associated ``instFlux`` values 

1434 localCalib : `numpy.ndarray` or `pandas.Series` 

1435 Array of local photometric calibration estimates. 

1436 localCalibErr : `numpy.ndarray` or `pandas.Series` 

1437 Errors on associated ``localCalib`` values 

1438 

1439 Returns 

1440 ------- 

1441 calibFluxErr : `numpy.ndarray` or `pandas.Series` 

1442 Errors on calibrated flux measurements. 

1443 """ 

1444 return np.hypot(instFluxErr * localCalib, instFlux * localCalibErr) 

1445 

1446 def instFluxToMagnitude(self, instFlux, localCalib): 

1447 """Convert instrument flux to nanojanskys. 

1448 

1449 Parameters 

1450 ---------- 

1451 instFlux : `numpy.ndarray` or `pandas.Series` 

1452 Array of instrument flux measurements 

1453 localCalib : `numpy.ndarray` or `pandas.Series` 

1454 Array of local photometric calibration estimates. 

1455 

1456 Returns 

1457 ------- 

1458 calibMag : `numpy.ndarray` or `pandas.Series` 

1459 Array of calibrated AB magnitudes. 

1460 """ 

1461 return -2.5 * np.log10(self.instFluxToNanojansky(instFlux, localCalib)) + self.logNJanskyToAB 

1462 

1463 def instFluxErrToMagnitudeErr(self, instFlux, instFluxErr, localCalib, localCalibErr): 

1464 """Convert instrument flux err to nanojanskys. 

1465 

1466 Parameters 

1467 ---------- 

1468 instFlux : `numpy.ndarray` or `pandas.Series` 

1469 Array of instrument flux measurements 

1470 instFluxErr : `numpy.ndarray` or `pandas.Series` 

1471 Errors on associated ``instFlux`` values 

1472 localCalib : `numpy.ndarray` or `pandas.Series` 

1473 Array of local photometric calibration estimates. 

1474 localCalibErr : `numpy.ndarray` or `pandas.Series` 

1475 Errors on associated ``localCalib`` values 

1476 

1477 Returns 

1478 ------- 

1479 calibMagErr: `numpy.ndarray` or `pandas.Series` 

1480 Error on calibrated AB magnitudes. 

1481 """ 

1482 err = self.instFluxErrToNanojanskyErr(instFlux, instFluxErr, localCalib, localCalibErr) 

1483 return 2.5 / np.log(10) * err / self.instFluxToNanojansky(instFlux, instFluxErr) 

1484 

1485 

1486class LocalNanojansky(LocalPhotometry): 

1487 """Compute calibrated fluxes using the local calibration value. 

1488 

1489 See also 

1490 -------- 

1491 LocalNanojansky 

1492 LocalNanojanskyErr 

1493 LocalMagnitude 

1494 LocalMagnitudeErr 

1495 """ 

1496 

1497 @property 

1498 def columns(self): 

1499 return [self.instFluxCol, self.photoCalibCol] 

1500 

1501 @property 

1502 def name(self): 

1503 return f'flux_{self.instFluxCol}' 

1504 

1505 def _func(self, df): 

1506 return self.instFluxToNanojansky(df[self.instFluxCol], df[self.photoCalibCol]) 

1507 

1508 

1509class LocalNanojanskyErr(LocalPhotometry): 

1510 """Compute calibrated flux errors using the local calibration value. 

1511 

1512 See also 

1513 -------- 

1514 LocalNanojansky 

1515 LocalNanojanskyErr 

1516 LocalMagnitude 

1517 LocalMagnitudeErr 

1518 """ 

1519 

1520 @property 

1521 def columns(self): 

1522 return [self.instFluxCol, self.instFluxErrCol, 

1523 self.photoCalibCol, self.photoCalibErrCol] 

1524 

1525 @property 

1526 def name(self): 

1527 return f'fluxErr_{self.instFluxCol}' 

1528 

1529 def _func(self, df): 

1530 return self.instFluxErrToNanojanskyErr(df[self.instFluxCol], df[self.instFluxErrCol], 

1531 df[self.photoCalibCol], df[self.photoCalibErrCol]) 

1532 

1533 

1534class LocalMagnitude(LocalPhotometry): 

1535 """Compute calibrated AB magnitudes using the local calibration value. 

1536 

1537 See also 

1538 -------- 

1539 LocalNanojansky 

1540 LocalNanojanskyErr 

1541 LocalMagnitude 

1542 LocalMagnitudeErr 

1543 """ 

1544 

1545 @property 

1546 def columns(self): 

1547 return [self.instFluxCol, self.photoCalibCol] 

1548 

1549 @property 

1550 def name(self): 

1551 return f'mag_{self.instFluxCol}' 

1552 

1553 def _func(self, df): 

1554 return self.instFluxToMagnitude(df[self.instFluxCol], 

1555 df[self.photoCalibCol]) 

1556 

1557 

1558class LocalMagnitudeErr(LocalPhotometry): 

1559 """Compute calibrated AB magnitude errors using the local calibration value. 

1560 

1561 See also 

1562 -------- 

1563 LocalNanojansky 

1564 LocalNanojanskyErr 

1565 LocalMagnitude 

1566 LocalMagnitudeErr 

1567 """ 

1568 

1569 @property 

1570 def columns(self): 

1571 return [self.instFluxCol, self.instFluxErrCol, 

1572 self.photoCalibCol, self.photoCalibErrCol] 

1573 

1574 @property 

1575 def name(self): 

1576 return f'magErr_{self.instFluxCol}' 

1577 

1578 def _func(self, df): 

1579 return self.instFluxErrToMagnitudeErr(df[self.instFluxCol], 

1580 df[self.instFluxErrCol], 

1581 df[self.photoCalibCol], 

1582 df[self.photoCalibErrCol])