Coverage for python/lsst/ip/diffim/utils.py: 5%

Shortcuts 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

689 statements  

1# This file is part of ip_diffim. 

2# 

3# Developed for the LSST Data Management System. 

4# This product includes software developed by the LSST Project 

5# (https://www.lsst.org). 

6# See the COPYRIGHT file at the top-level directory of this distribution 

7# for details of code ownership. 

8# 

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

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

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

12# (at your option) any later version. 

13# 

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

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

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

17# GNU General Public License for more details. 

18# 

19# You should have received a copy of the GNU General Public License 

20# along with this program. If not, see <https://www.gnu.org/licenses/>. 

21 

22"""Support utilities for Measuring sources""" 

23 

24# Export DipoleTestImage to expose fake image generating funcs 

25__all__ = ["DipoleTestImage"] 

26 

27import numpy as np 

28 

29import lsst.geom as geom 

30import lsst.afw.detection as afwDet 

31import lsst.afw.display as afwDisplay 

32import lsst.afw.geom as afwGeom 

33import lsst.afw.image as afwImage 

34import lsst.afw.math as afwMath 

35import lsst.afw.table as afwTable 

36from lsst.log import Log 

37import lsst.meas.algorithms as measAlg 

38import lsst.meas.base as measBase 

39from .dipoleFitTask import DipoleFitAlgorithm 

40from . import diffimLib 

41from . import diffimTools 

42 

43afwDisplay.setDefaultMaskTransparency(75) 

44keptPlots = False # Have we arranged to keep spatial plots open? 

45 

46 

47def showSourceSet(sSet, xy0=(0, 0), frame=0, ctype=afwDisplay.GREEN, symb="+", size=2): 

48 """Draw the (XAstrom, YAstrom) positions of a set of Sources. 

49 

50 Image has the given XY0. 

51 """ 

52 disp = afwDisplay.afwDisplay(frame=frame) 

53 with disp.Buffering(): 

54 for s in sSet: 

55 xc, yc = s.getXAstrom() - xy0[0], s.getYAstrom() - xy0[1] 

56 

57 if symb == "id": 

58 disp.dot(str(s.getId()), xc, yc, ctype=ctype, size=size) 

59 else: 

60 disp.dot(symb, xc, yc, ctype=ctype, size=size) 

61 

62 

63# Kernel display utilities 

64# 

65 

66 

67def showKernelSpatialCells(maskedIm, kernelCellSet, showChi2=False, symb="o", 

68 ctype=None, ctypeUnused=None, ctypeBad=None, size=3, 

69 frame=None, title="Spatial Cells"): 

70 """Show the SpatialCells. 

71 

72 If symb is something that display.dot understands (e.g. "o"), the top 

73 nMaxPerCell candidates will be indicated with that symbol, using ctype 

74 and size. 

75 """ 

76 disp = afwDisplay.Display(frame=frame) 

77 disp.mtv(maskedIm, title=title) 

78 with disp.Buffering(): 

79 origin = [-maskedIm.getX0(), -maskedIm.getY0()] 

80 for cell in kernelCellSet.getCellList(): 

81 afwDisplay.utils.drawBBox(cell.getBBox(), origin=origin, display=disp) 

82 

83 goodies = ctypeBad is None 

84 for cand in cell.begin(goodies): 

85 xc, yc = cand.getXCenter() + origin[0], cand.getYCenter() + origin[1] 

86 if cand.getStatus() == afwMath.SpatialCellCandidate.BAD: 

87 color = ctypeBad 

88 elif cand.getStatus() == afwMath.SpatialCellCandidate.GOOD: 

89 color = ctype 

90 elif cand.getStatus() == afwMath.SpatialCellCandidate.UNKNOWN: 

91 color = ctypeUnused 

92 else: 

93 continue 

94 

95 if color: 

96 disp.dot(symb, xc, yc, ctype=color, size=size) 

97 

98 if showChi2: 

99 rchi2 = cand.getChi2() 

100 if rchi2 > 1e100: 

101 rchi2 = np.nan 

102 disp.dot("%d %.1f" % (cand.getId(), rchi2), 

103 xc - size, yc - size - 4, ctype=color, size=size) 

104 

105 

106def showDiaSources(sources, exposure, isFlagged, isDipole, frame=None): 

107 """Display Dia Sources. 

108 """ 

109 # 

110 # Show us the ccandidates 

111 # 

112 # Too many mask planes in diffims 

113 disp = afwDisplay.Display(frame=frame) 

114 for plane in ("BAD", "CR", "EDGE", "INTERPOlATED", "INTRP", "SAT", "SATURATED"): 

115 disp.setMaskPlaneColor(plane, color="ignore") 

116 

117 mos = afwDisplay.utils.Mosaic() 

118 for i in range(len(sources)): 

119 source = sources[i] 

120 badFlag = isFlagged[i] 

121 dipoleFlag = isDipole[i] 

122 bbox = source.getFootprint().getBBox() 

123 stamp = exposure.Factory(exposure, bbox, True) 

124 im = afwDisplay.utils.Mosaic(gutter=1, background=0, mode="x") 

125 im.append(stamp.getMaskedImage()) 

126 lab = "%.1f,%.1f:" % (source.getX(), source.getY()) 

127 if badFlag: 

128 ctype = afwDisplay.RED 

129 lab += "BAD" 

130 if dipoleFlag: 

131 ctype = afwDisplay.YELLOW 

132 lab += "DIPOLE" 

133 if not badFlag and not dipoleFlag: 

134 ctype = afwDisplay.GREEN 

135 lab += "OK" 

136 mos.append(im.makeMosaic(), lab, ctype) 

137 title = "Dia Sources" 

138 mosaicImage = mos.makeMosaic(display=disp, title=title) 

139 return mosaicImage 

140 

141 

142def showKernelCandidates(kernelCellSet, kernel, background, frame=None, showBadCandidates=True, 

143 resids=False, kernels=False): 

144 """Display the Kernel candidates. 

145 

146 If kernel is provided include spatial model and residuals; 

147 If chi is True, generate a plot of residuals/sqrt(variance), i.e. chi. 

148 """ 

149 # 

150 # Show us the ccandidates 

151 # 

152 if kernels: 

153 mos = afwDisplay.utils.Mosaic(gutter=5, background=0) 

154 else: 

155 mos = afwDisplay.utils.Mosaic(gutter=5, background=-1) 

156 # 

157 candidateCenters = [] 

158 candidateCentersBad = [] 

159 candidateIndex = 0 

160 for cell in kernelCellSet.getCellList(): 

161 for cand in cell.begin(False): # include bad candidates 

162 # Original difference image; if does not exist, skip candidate 

163 try: 

164 resid = cand.getDifferenceImage(diffimLib.KernelCandidateF.ORIG) 

165 except Exception: 

166 continue 

167 

168 rchi2 = cand.getChi2() 

169 if rchi2 > 1e100: 

170 rchi2 = np.nan 

171 

172 if not showBadCandidates and cand.isBad(): 

173 continue 

174 

175 im_resid = afwDisplay.utils.Mosaic(gutter=1, background=-0.5, mode="x") 

176 

177 try: 

178 im = cand.getScienceMaskedImage() 

179 im = im.Factory(im, True) 

180 im.setXY0(cand.getScienceMaskedImage().getXY0()) 

181 except Exception: 

182 continue 

183 if (not resids and not kernels): 

184 im_resid.append(im.Factory(im, True)) 

185 try: 

186 im = cand.getTemplateMaskedImage() 

187 im = im.Factory(im, True) 

188 im.setXY0(cand.getTemplateMaskedImage().getXY0()) 

189 except Exception: 

190 continue 

191 if (not resids and not kernels): 

192 im_resid.append(im.Factory(im, True)) 

193 

194 # Difference image with original basis 

195 if resids: 

196 var = resid.getVariance() 

197 var = var.Factory(var, True) 

198 np.sqrt(var.getArray(), var.getArray()) # inplace sqrt 

199 resid = resid.getImage() 

200 resid /= var 

201 bbox = kernel.shrinkBBox(resid.getBBox()) 

202 resid = resid.Factory(resid, bbox, deep=True) 

203 elif kernels: 

204 kim = cand.getKernelImage(diffimLib.KernelCandidateF.ORIG).convertF() 

205 resid = kim.Factory(kim, True) 

206 im_resid.append(resid) 

207 

208 # residuals using spatial model 

209 ski = afwImage.ImageD(kernel.getDimensions()) 

210 kernel.computeImage(ski, False, int(cand.getXCenter()), int(cand.getYCenter())) 

211 sk = afwMath.FixedKernel(ski) 

212 sbg = 0.0 

213 if background: 

214 sbg = background(int(cand.getXCenter()), int(cand.getYCenter())) 

215 sresid = cand.getDifferenceImage(sk, sbg) 

216 resid = sresid 

217 if resids: 

218 resid = sresid.getImage() 

219 resid /= var 

220 bbox = kernel.shrinkBBox(resid.getBBox()) 

221 resid = resid.Factory(resid, bbox, deep=True) 

222 elif kernels: 

223 kim = ski.convertF() 

224 resid = kim.Factory(kim, True) 

225 im_resid.append(resid) 

226 

227 im = im_resid.makeMosaic() 

228 

229 lab = "%d chi^2 %.1f" % (cand.getId(), rchi2) 

230 ctype = afwDisplay.RED if cand.isBad() else afwDisplay.GREEN 

231 

232 mos.append(im, lab, ctype) 

233 

234 if False and np.isnan(rchi2): 

235 disp = afwDisplay.Display(frame=1) 

236 disp.mtv(cand.getScienceMaskedImage.getImage(), title="candidate") 

237 print("rating", cand.getCandidateRating()) 

238 

239 im = cand.getScienceMaskedImage() 

240 center = (candidateIndex, cand.getXCenter() - im.getX0(), cand.getYCenter() - im.getY0()) 

241 candidateIndex += 1 

242 if cand.isBad(): 

243 candidateCentersBad.append(center) 

244 else: 

245 candidateCenters.append(center) 

246 

247 if resids: 

248 title = "chi Diffim" 

249 elif kernels: 

250 title = "Kernels" 

251 else: 

252 title = "Candidates & residuals" 

253 

254 disp = afwDisplay.Display(frame=frame) 

255 mosaicImage = mos.makeMosaic(display=disp, title=title) 

256 

257 return mosaicImage 

258 

259 

260def showKernelBasis(kernel, frame=None): 

261 """Display a Kernel's basis images. 

262 """ 

263 mos = afwDisplay.utils.Mosaic() 

264 

265 for k in kernel.getKernelList(): 

266 im = afwImage.ImageD(k.getDimensions()) 

267 k.computeImage(im, False) 

268 mos.append(im) 

269 

270 disp = afwDisplay.Display(frame=frame) 

271 mos.makeMosaic(display=disp, title="Kernel Basis Images") 

272 

273 return mos 

274 

275############### 

276 

277 

278def plotKernelSpatialModel(kernel, kernelCellSet, showBadCandidates=True, 

279 numSample=128, keepPlots=True, maxCoeff=10): 

280 """Plot the Kernel spatial model. 

281 """ 

282 try: 

283 import matplotlib.pyplot as plt 

284 import matplotlib.colors 

285 except ImportError as e: 

286 print("Unable to import numpy and matplotlib: %s" % e) 

287 return 

288 

289 x0 = kernelCellSet.getBBox().getBeginX() 

290 y0 = kernelCellSet.getBBox().getBeginY() 

291 

292 candPos = list() 

293 candFits = list() 

294 badPos = list() 

295 badFits = list() 

296 candAmps = list() 

297 badAmps = list() 

298 for cell in kernelCellSet.getCellList(): 

299 for cand in cell.begin(False): 

300 if not showBadCandidates and cand.isBad(): 

301 continue 

302 candCenter = geom.PointD(cand.getXCenter(), cand.getYCenter()) 

303 try: 

304 im = cand.getTemplateMaskedImage() 

305 except Exception: 

306 continue 

307 

308 targetFits = badFits if cand.isBad() else candFits 

309 targetPos = badPos if cand.isBad() else candPos 

310 targetAmps = badAmps if cand.isBad() else candAmps 

311 

312 # compare original and spatial kernel coefficients 

313 kp0 = np.array(cand.getKernel(diffimLib.KernelCandidateF.ORIG).getKernelParameters()) 

314 amp = cand.getCandidateRating() 

315 

316 targetFits = badFits if cand.isBad() else candFits 

317 targetPos = badPos if cand.isBad() else candPos 

318 targetAmps = badAmps if cand.isBad() else candAmps 

319 

320 targetFits.append(kp0) 

321 targetPos.append(candCenter) 

322 targetAmps.append(amp) 

323 

324 xGood = np.array([pos.getX() for pos in candPos]) - x0 

325 yGood = np.array([pos.getY() for pos in candPos]) - y0 

326 zGood = np.array(candFits) 

327 

328 xBad = np.array([pos.getX() for pos in badPos]) - x0 

329 yBad = np.array([pos.getY() for pos in badPos]) - y0 

330 zBad = np.array(badFits) 

331 numBad = len(badPos) 

332 

333 xRange = np.linspace(0, kernelCellSet.getBBox().getWidth(), num=numSample) 

334 yRange = np.linspace(0, kernelCellSet.getBBox().getHeight(), num=numSample) 

335 

336 if maxCoeff: 

337 maxCoeff = min(maxCoeff, kernel.getNKernelParameters()) 

338 else: 

339 maxCoeff = kernel.getNKernelParameters() 

340 

341 for k in range(maxCoeff): 

342 func = kernel.getSpatialFunction(k) 

343 dfGood = zGood[:, k] - np.array([func(pos.getX(), pos.getY()) for pos in candPos]) 

344 yMin = dfGood.min() 

345 yMax = dfGood.max() 

346 if numBad > 0: 

347 dfBad = zBad[:, k] - np.array([func(pos.getX(), pos.getY()) for pos in badPos]) 

348 # Can really screw up the range... 

349 yMin = min([yMin, dfBad.min()]) 

350 yMax = max([yMax, dfBad.max()]) 

351 yMin -= 0.05*(yMax - yMin) 

352 yMax += 0.05*(yMax - yMin) 

353 

354 fRange = np.ndarray((len(xRange), len(yRange))) 

355 for j, yVal in enumerate(yRange): 

356 for i, xVal in enumerate(xRange): 

357 fRange[j][i] = func(xVal, yVal) 

358 

359 fig = plt.figure(k) 

360 

361 fig.clf() 

362 try: 

363 fig.canvas._tkcanvas._root().lift() # == Tk's raise, but raise is a python reserved word 

364 except Exception: # protect against API changes 

365 pass 

366 

367 fig.suptitle('Kernel component %d' % k) 

368 

369 # LL 

370 ax = fig.add_axes((0.1, 0.05, 0.35, 0.35)) 

371 vmin = fRange.min() # - 0.05*np.fabs(fRange.min()) 

372 vmax = fRange.max() # + 0.05*np.fabs(fRange.max()) 

373 norm = matplotlib.colors.Normalize(vmin=vmin, vmax=vmax) 

374 im = ax.imshow(fRange, aspect='auto', norm=norm, 

375 extent=[0, kernelCellSet.getBBox().getWidth() - 1, 

376 0, kernelCellSet.getBBox().getHeight() - 1]) 

377 ax.set_title('Spatial polynomial') 

378 plt.colorbar(im, orientation='horizontal', ticks=[vmin, vmax]) 

379 

380 # UL 

381 ax = fig.add_axes((0.1, 0.55, 0.35, 0.35)) 

382 ax.plot(-2.5*np.log10(candAmps), zGood[:, k], 'b+') 

383 if numBad > 0: 

384 ax.plot(-2.5*np.log10(badAmps), zBad[:, k], 'r+') 

385 ax.set_title("Basis Coefficients") 

386 ax.set_xlabel("Instr mag") 

387 ax.set_ylabel("Coeff") 

388 

389 # LR 

390 ax = fig.add_axes((0.55, 0.05, 0.35, 0.35)) 

391 ax.set_autoscale_on(False) 

392 ax.set_xbound(lower=0, upper=kernelCellSet.getBBox().getHeight()) 

393 ax.set_ybound(lower=yMin, upper=yMax) 

394 ax.plot(yGood, dfGood, 'b+') 

395 if numBad > 0: 

396 ax.plot(yBad, dfBad, 'r+') 

397 ax.axhline(0.0) 

398 ax.set_title('dCoeff (indiv-spatial) vs. y') 

399 

400 # UR 

401 ax = fig.add_axes((0.55, 0.55, 0.35, 0.35)) 

402 ax.set_autoscale_on(False) 

403 ax.set_xbound(lower=0, upper=kernelCellSet.getBBox().getWidth()) 

404 ax.set_ybound(lower=yMin, upper=yMax) 

405 ax.plot(xGood, dfGood, 'b+') 

406 if numBad > 0: 

407 ax.plot(xBad, dfBad, 'r+') 

408 ax.axhline(0.0) 

409 ax.set_title('dCoeff (indiv-spatial) vs. x') 

410 

411 fig.show() 

412 

413 global keptPlots 

414 if keepPlots and not keptPlots: 

415 # Keep plots open when done 

416 def show(): 

417 print("%s: Please close plots when done." % __name__) 

418 try: 

419 plt.show() 

420 except Exception: 

421 pass 

422 print("Plots closed, exiting...") 

423 import atexit 

424 atexit.register(show) 

425 keptPlots = True 

426 

427 

428def plotKernelCoefficients(spatialKernel, kernelCellSet, showBadCandidates=False, keepPlots=True): 

429 """Plot the individual kernel candidate and the spatial kernel solution coefficients. 

430 

431 Parameters 

432 ---------- 

433 

434 spatialKernel : `lsst.afw.math.LinearCombinationKernel` 

435 The spatial spatialKernel solution model which is a spatially varying linear combination 

436 of the spatialKernel basis functions. 

437 Typically returned by `lsst.ip.diffim.SpatialKernelSolution.getSolutionPair()`. 

438 

439 kernelCellSet : `lsst.afw.math.SpatialCellSet` 

440 The spatial cells that was used for solution for the spatialKernel. They contain the 

441 local solutions of the AL kernel for the selected sources. 

442 

443 showBadCandidates : `bool`, optional 

444 If True, plot the coefficient values for kernel candidates where the solution was marked 

445 bad by the numerical algorithm. Defaults to False. 

446 

447 keepPlots: `bool`, optional 

448 If True, sets ``plt.show()`` to be called before the task terminates, so that the plots 

449 can be explored interactively. Defaults to True. 

450 

451 Notes 

452 ----- 

453 This function produces 3 figures per image subtraction operation. 

454 * A grid plot of the local solutions. Each grid cell corresponds to a proportional area in 

455 the image. In each cell, local kernel solution coefficients are plotted of kernel candidates (color) 

456 that fall into this area as a function of the kernel basis function number. 

457 * A grid plot of the spatial solution. Each grid cell corresponds to a proportional area in 

458 the image. In each cell, the spatial solution coefficients are evaluated for the center of the cell. 

459 * Histogram of the local solution coefficients. Red line marks the spatial solution value at 

460 center of the image. 

461 

462 This function is called if ``lsst.ip.diffim.psfMatch.plotKernelCoefficients==True`` in lsstDebug. This 

463 function was implemented as part of DM-17825. 

464 """ 

465 try: 

466 import matplotlib.pyplot as plt 

467 except ImportError as e: 

468 print("Unable to import matplotlib: %s" % e) 

469 return 

470 

471 # Image dimensions 

472 imgBBox = kernelCellSet.getBBox() 

473 x0 = imgBBox.getBeginX() 

474 y0 = imgBBox.getBeginY() 

475 wImage = imgBBox.getWidth() 

476 hImage = imgBBox.getHeight() 

477 imgCenterX = imgBBox.getCenterX() 

478 imgCenterY = imgBBox.getCenterY() 

479 

480 # Plot the local solutions 

481 # ---- 

482 

483 # Grid size 

484 nX = 8 

485 nY = 8 

486 wCell = wImage / nX 

487 hCell = hImage / nY 

488 

489 fig = plt.figure() 

490 fig.suptitle("Kernel candidate parameters on an image grid") 

491 arrAx = fig.subplots(nrows=nY, ncols=nX, sharex=True, sharey=True, gridspec_kw=dict( 

492 wspace=0, hspace=0)) 

493 

494 # Bottom left panel is for bottom left part of the image 

495 arrAx = arrAx[::-1, :] 

496 

497 allParams = [] 

498 for cell in kernelCellSet.getCellList(): 

499 cellBBox = afwGeom.Box2D(cell.getBBox()) 

500 # Determine which panel this spatial cell belongs to 

501 iX = int((cellBBox.getCenterX() - x0)//wCell) 

502 iY = int((cellBBox.getCenterY() - y0)//hCell) 

503 

504 for cand in cell.begin(False): 

505 try: 

506 kernel = cand.getKernel(cand.ORIG) 

507 except Exception: 

508 continue 

509 

510 if not showBadCandidates and cand.isBad(): 

511 continue 

512 

513 nKernelParams = kernel.getNKernelParameters() 

514 kernelParams = np.array(kernel.getKernelParameters()) 

515 allParams.append(kernelParams) 

516 

517 if cand.isBad(): 

518 color = 'red' 

519 else: 

520 color = None 

521 arrAx[iY, iX].plot(np.arange(nKernelParams), kernelParams, '.-', 

522 color=color, drawstyle='steps-mid', linewidth=0.1) 

523 for ax in arrAx.ravel(): 

524 ax.grid(True, axis='y') 

525 

526 # Plot histogram of the local parameters and the global solution at the image center 

527 # ---- 

528 

529 spatialFuncs = spatialKernel.getSpatialFunctionList() 

530 nKernelParams = spatialKernel.getNKernelParameters() 

531 nX = 8 

532 fig = plt.figure() 

533 fig.suptitle("Hist. of parameters marked with spatial solution at img center") 

534 arrAx = fig.subplots(nrows=int(nKernelParams//nX)+1, ncols=nX) 

535 arrAx = arrAx[::-1, :] 

536 allParams = np.array(allParams) 

537 for k in range(nKernelParams): 

538 ax = arrAx.ravel()[k] 

539 ax.hist(allParams[:, k], bins=20, edgecolor='black') 

540 ax.set_xlabel('P{}'.format(k)) 

541 valueParam = spatialFuncs[k](imgCenterX, imgCenterY) 

542 ax.axvline(x=valueParam, color='red') 

543 ax.text(0.1, 0.9, '{:.1f}'.format(valueParam), 

544 transform=ax.transAxes, backgroundcolor='lightsteelblue') 

545 

546 # Plot grid of the spatial solution 

547 # ---- 

548 

549 nX = 8 

550 nY = 8 

551 wCell = wImage / nX 

552 hCell = hImage / nY 

553 x0 += wCell / 2 

554 y0 += hCell / 2 

555 

556 fig = plt.figure() 

557 fig.suptitle("Spatial solution of kernel parameters on an image grid") 

558 arrAx = fig.subplots(nrows=nY, ncols=nX, sharex=True, sharey=True, gridspec_kw=dict( 

559 wspace=0, hspace=0)) 

560 arrAx = arrAx[::-1, :] 

561 kernelParams = np.zeros(nKernelParams, dtype=float) 

562 

563 for iX in range(nX): 

564 for iY in range(nY): 

565 x = x0 + iX * wCell 

566 y = y0 + iY * hCell 

567 # Evaluate the spatial solution functions for this x,y location 

568 kernelParams = [f(x, y) for f in spatialFuncs] 

569 arrAx[iY, iX].plot(np.arange(nKernelParams), kernelParams, '.-', drawstyle='steps-mid') 

570 arrAx[iY, iX].grid(True, axis='y') 

571 

572 global keptPlots 

573 if keepPlots and not keptPlots: 

574 # Keep plots open when done 

575 def show(): 

576 print("%s: Please close plots when done." % __name__) 

577 try: 

578 plt.show() 

579 except Exception: 

580 pass 

581 print("Plots closed, exiting...") 

582 import atexit 

583 atexit.register(show) 

584 keptPlots = True 

585 

586 

587def showKernelMosaic(bbox, kernel, nx=7, ny=None, frame=None, title=None, 

588 showCenter=True, showEllipticity=True): 

589 """Show a mosaic of Kernel images. 

590 """ 

591 mos = afwDisplay.utils.Mosaic() 

592 

593 x0 = bbox.getBeginX() 

594 y0 = bbox.getBeginY() 

595 width = bbox.getWidth() 

596 height = bbox.getHeight() 

597 

598 if not ny: 

599 ny = int(nx*float(height)/width + 0.5) 

600 if not ny: 

601 ny = 1 

602 

603 schema = afwTable.SourceTable.makeMinimalSchema() 

604 centroidName = "base_SdssCentroid" 

605 shapeName = "base_SdssShape" 

606 control = measBase.SdssCentroidControl() 

607 schema.getAliasMap().set("slot_Centroid", centroidName) 

608 schema.getAliasMap().set("slot_Centroid_flag", centroidName + "_flag") 

609 centroider = measBase.SdssCentroidAlgorithm(control, centroidName, schema) 

610 sdssShape = measBase.SdssShapeControl() 

611 shaper = measBase.SdssShapeAlgorithm(sdssShape, shapeName, schema) 

612 table = afwTable.SourceTable.make(schema) 

613 table.defineCentroid(centroidName) 

614 table.defineShape(shapeName) 

615 

616 centers = [] 

617 shapes = [] 

618 for iy in range(ny): 

619 for ix in range(nx): 

620 x = int(ix*(width - 1)/(nx - 1)) + x0 

621 y = int(iy*(height - 1)/(ny - 1)) + y0 

622 

623 im = afwImage.ImageD(kernel.getDimensions()) 

624 ksum = kernel.computeImage(im, False, x, y) 

625 lab = "Kernel(%d,%d)=%.2f" % (x, y, ksum) if False else "" 

626 mos.append(im, lab) 

627 

628 # SdssCentroidAlgorithm.measure requires an exposure of floats 

629 exp = afwImage.makeExposure(afwImage.makeMaskedImage(im.convertF())) 

630 

631 w, h = im.getWidth(), im.getHeight() 

632 centerX = im.getX0() + w//2 

633 centerY = im.getY0() + h//2 

634 src = table.makeRecord() 

635 spans = afwGeom.SpanSet(exp.getBBox()) 

636 foot = afwDet.Footprint(spans) 

637 foot.addPeak(centerX, centerY, 1) 

638 src.setFootprint(foot) 

639 

640 try: # The centroider requires a psf, so this will fail if none is attached to exp 

641 centroider.measure(src, exp) 

642 centers.append((src.getX(), src.getY())) 

643 

644 shaper.measure(src, exp) 

645 shapes.append((src.getIxx(), src.getIxy(), src.getIyy())) 

646 except Exception: 

647 pass 

648 

649 disp = afwDisplay.Display(frame=frame) 

650 mos.makeMosaic(display=disp, title=title if title else "Model Kernel", mode=nx) 

651 

652 if centers and frame is not None: 

653 disp = afwDisplay.Display(frame=frame) 

654 i = 0 

655 with disp.Buffering(): 

656 for cen, shape in zip(centers, shapes): 

657 bbox = mos.getBBox(i) 

658 i += 1 

659 xc, yc = cen[0] + bbox.getMinX(), cen[1] + bbox.getMinY() 

660 if showCenter: 

661 disp.dot("+", xc, yc, ctype=afwDisplay.BLUE) 

662 

663 if showEllipticity: 

664 ixx, ixy, iyy = shape 

665 disp.dot("@:%g,%g,%g" % (ixx, ixy, iyy), xc, yc, ctype=afwDisplay.RED) 

666 

667 return mos 

668 

669 

670def plotPixelResiduals(exposure, warpedTemplateExposure, diffExposure, kernelCellSet, 

671 kernel, background, testSources, config, 

672 origVariance=False, nptsFull=1e6, keepPlots=True, titleFs=14): 

673 """Plot diffim residuals for LOCAL and SPATIAL models. 

674 """ 

675 candidateResids = [] 

676 spatialResids = [] 

677 nonfitResids = [] 

678 

679 for cell in kernelCellSet.getCellList(): 

680 for cand in cell.begin(True): # only look at good ones 

681 # Be sure 

682 if not (cand.getStatus() == afwMath.SpatialCellCandidate.GOOD): 

683 continue 

684 

685 diffim = cand.getDifferenceImage(diffimLib.KernelCandidateF.ORIG) 

686 orig = cand.getScienceMaskedImage() 

687 

688 ski = afwImage.ImageD(kernel.getDimensions()) 

689 kernel.computeImage(ski, False, int(cand.getXCenter()), int(cand.getYCenter())) 

690 sk = afwMath.FixedKernel(ski) 

691 sbg = background(int(cand.getXCenter()), int(cand.getYCenter())) 

692 sdiffim = cand.getDifferenceImage(sk, sbg) 

693 

694 # trim edgs due to convolution 

695 bbox = kernel.shrinkBBox(diffim.getBBox()) 

696 tdiffim = diffim.Factory(diffim, bbox) 

697 torig = orig.Factory(orig, bbox) 

698 tsdiffim = sdiffim.Factory(sdiffim, bbox) 

699 

700 if origVariance: 

701 candidateResids.append(np.ravel(tdiffim.getImage().getArray() 

702 / np.sqrt(torig.getVariance().getArray()))) 

703 spatialResids.append(np.ravel(tsdiffim.getImage().getArray() 

704 / np.sqrt(torig.getVariance().getArray()))) 

705 else: 

706 candidateResids.append(np.ravel(tdiffim.getImage().getArray() 

707 / np.sqrt(tdiffim.getVariance().getArray()))) 

708 spatialResids.append(np.ravel(tsdiffim.getImage().getArray() 

709 / np.sqrt(tsdiffim.getVariance().getArray()))) 

710 

711 fullIm = diffExposure.getMaskedImage().getImage().getArray() 

712 fullMask = diffExposure.getMaskedImage().getMask().getArray() 

713 if origVariance: 

714 fullVar = exposure.getMaskedImage().getVariance().getArray() 

715 else: 

716 fullVar = diffExposure.getMaskedImage().getVariance().getArray() 

717 

718 bitmaskBad = 0 

719 bitmaskBad |= afwImage.Mask.getPlaneBitMask('NO_DATA') 

720 bitmaskBad |= afwImage.Mask.getPlaneBitMask('SAT') 

721 idx = np.where((fullMask & bitmaskBad) == 0) 

722 stride = int(len(idx[0])//nptsFull) 

723 sidx = idx[0][::stride], idx[1][::stride] 

724 allResids = fullIm[sidx]/np.sqrt(fullVar[sidx]) 

725 

726 testFootprints = diffimTools.sourceToFootprintList(testSources, warpedTemplateExposure, 

727 exposure, config, Log.getDefaultLogger()) 

728 for fp in testFootprints: 

729 subexp = diffExposure.Factory(diffExposure, fp["footprint"].getBBox()) 

730 subim = subexp.getMaskedImage().getImage() 

731 if origVariance: 

732 subvar = afwImage.ExposureF(exposure, fp["footprint"].getBBox()).getMaskedImage().getVariance() 

733 else: 

734 subvar = subexp.getMaskedImage().getVariance() 

735 nonfitResids.append(np.ravel(subim.getArray()/np.sqrt(subvar.getArray()))) 

736 

737 candidateResids = np.ravel(np.array(candidateResids)) 

738 spatialResids = np.ravel(np.array(spatialResids)) 

739 nonfitResids = np.ravel(np.array(nonfitResids)) 

740 

741 try: 

742 import pylab 

743 from matplotlib.font_manager import FontProperties 

744 except ImportError as e: 

745 print("Unable to import pylab: %s" % e) 

746 return 

747 

748 fig = pylab.figure() 

749 fig.clf() 

750 try: 

751 fig.canvas._tkcanvas._root().lift() # == Tk's raise, but raise is a python reserved word 

752 except Exception: # protect against API changes 

753 pass 

754 if origVariance: 

755 fig.suptitle("Diffim residuals: Normalized by sqrt(input variance)", fontsize=titleFs) 

756 else: 

757 fig.suptitle("Diffim residuals: Normalized by sqrt(diffim variance)", fontsize=titleFs) 

758 

759 sp1 = pylab.subplot(221) 

760 sp2 = pylab.subplot(222, sharex=sp1, sharey=sp1) 

761 sp3 = pylab.subplot(223, sharex=sp1, sharey=sp1) 

762 sp4 = pylab.subplot(224, sharex=sp1, sharey=sp1) 

763 xs = np.arange(-5, 5.05, 0.1) 

764 ys = 1./np.sqrt(2*np.pi)*np.exp(-0.5*xs**2) 

765 

766 sp1.hist(candidateResids, bins=xs, normed=True, alpha=0.5, label="N(%.2f, %.2f)" 

767 % (np.mean(candidateResids), np.var(candidateResids))) 

768 sp1.plot(xs, ys, "r-", lw=2, label="N(0,1)") 

769 sp1.set_title("Candidates: basis fit", fontsize=titleFs - 2) 

770 sp1.legend(loc=1, fancybox=True, shadow=True, prop=FontProperties(size=titleFs - 6)) 

771 

772 sp2.hist(spatialResids, bins=xs, normed=True, alpha=0.5, label="N(%.2f, %.2f)" 

773 % (np.mean(spatialResids), np.var(spatialResids))) 

774 sp2.plot(xs, ys, "r-", lw=2, label="N(0,1)") 

775 sp2.set_title("Candidates: spatial fit", fontsize=titleFs - 2) 

776 sp2.legend(loc=1, fancybox=True, shadow=True, prop=FontProperties(size=titleFs - 6)) 

777 

778 sp3.hist(nonfitResids, bins=xs, normed=True, alpha=0.5, label="N(%.2f, %.2f)" 

779 % (np.mean(nonfitResids), np.var(nonfitResids))) 

780 sp3.plot(xs, ys, "r-", lw=2, label="N(0,1)") 

781 sp3.set_title("Control sample: spatial fit", fontsize=titleFs - 2) 

782 sp3.legend(loc=1, fancybox=True, shadow=True, prop=FontProperties(size=titleFs - 6)) 

783 

784 sp4.hist(allResids, bins=xs, normed=True, alpha=0.5, label="N(%.2f, %.2f)" 

785 % (np.mean(allResids), np.var(allResids))) 

786 sp4.plot(xs, ys, "r-", lw=2, label="N(0,1)") 

787 sp4.set_title("Full image (subsampled)", fontsize=titleFs - 2) 

788 sp4.legend(loc=1, fancybox=True, shadow=True, prop=FontProperties(size=titleFs - 6)) 

789 

790 pylab.setp(sp1.get_xticklabels() + sp1.get_yticklabels(), fontsize=titleFs - 4) 

791 pylab.setp(sp2.get_xticklabels() + sp2.get_yticklabels(), fontsize=titleFs - 4) 

792 pylab.setp(sp3.get_xticklabels() + sp3.get_yticklabels(), fontsize=titleFs - 4) 

793 pylab.setp(sp4.get_xticklabels() + sp4.get_yticklabels(), fontsize=titleFs - 4) 

794 

795 sp1.set_xlim(-5, 5) 

796 sp1.set_ylim(0, 0.5) 

797 fig.show() 

798 

799 global keptPlots 

800 if keepPlots and not keptPlots: 

801 # Keep plots open when done 

802 def show(): 

803 print("%s: Please close plots when done." % __name__) 

804 try: 

805 pylab.show() 

806 except Exception: 

807 pass 

808 print("Plots closed, exiting...") 

809 import atexit 

810 atexit.register(show) 

811 keptPlots = True 

812 

813 

814def calcCentroid(arr): 

815 """Calculate first moment of a (kernel) image. 

816 """ 

817 y, x = arr.shape 

818 sarr = arr*arr 

819 xarr = np.asarray([[el for el in range(x)] for el2 in range(y)]) 

820 yarr = np.asarray([[el2 for el in range(x)] for el2 in range(y)]) 

821 narr = xarr*sarr 

822 sarrSum = sarr.sum() 

823 centx = narr.sum()/sarrSum 

824 narr = yarr*sarr 

825 centy = narr.sum()/sarrSum 

826 return centx, centy 

827 

828 

829def calcWidth(arr, centx, centy): 

830 """Calculate second moment of a (kernel) image. 

831 """ 

832 y, x = arr.shape 

833 # Square the flux so we don't have to deal with negatives 

834 sarr = arr*arr 

835 xarr = np.asarray([[el for el in range(x)] for el2 in range(y)]) 

836 yarr = np.asarray([[el2 for el in range(x)] for el2 in range(y)]) 

837 narr = sarr*np.power((xarr - centx), 2.) 

838 sarrSum = sarr.sum() 

839 xstd = np.sqrt(narr.sum()/sarrSum) 

840 narr = sarr*np.power((yarr - centy), 2.) 

841 ystd = np.sqrt(narr.sum()/sarrSum) 

842 return xstd, ystd 

843 

844 

845def printSkyDiffs(sources, wcs): 

846 """Print differences in sky coordinates. 

847 

848 The difference is that between the source Position and its Centroid mapped 

849 through Wcs. 

850 """ 

851 for s in sources: 

852 sCentroid = s.getCentroid() 

853 sPosition = s.getCoord().getPosition(geom.degrees) 

854 dra = 3600*(sPosition.getX() - wcs.pixelToSky(sCentroid).getPosition(geom.degrees).getX())/0.2 

855 ddec = 3600*(sPosition.getY() - wcs.pixelToSky(sCentroid).getPosition(geom.degrees).getY())/0.2 

856 if np.isfinite(dra) and np.isfinite(ddec): 

857 print(dra, ddec) 

858 

859 

860def makeRegions(sources, outfilename, wcs=None): 

861 """Create regions file for display from input source list. 

862 """ 

863 fh = open(outfilename, "w") 

864 fh.write("global color=red font=\"helvetica 10 normal\" " 

865 "select=1 highlite=1 edit=1 move=1 delete=1 include=1 fixed=0 source\nfk5\n") 

866 for s in sources: 

867 if wcs: 

868 (ra, dec) = wcs.pixelToSky(s.getCentroid()).getPosition(geom.degrees) 

869 else: 

870 (ra, dec) = s.getCoord().getPosition(geom.degrees) 

871 if np.isfinite(ra) and np.isfinite(dec): 

872 fh.write("circle(%f,%f,2\")\n"%(ra, dec)) 

873 fh.flush() 

874 fh.close() 

875 

876 

877def showSourceSetSky(sSet, wcs, xy0, frame=0, ctype=afwDisplay.GREEN, symb="+", size=2): 

878 """Draw the (RA, Dec) positions of a set of Sources. Image has the XY0. 

879 """ 

880 disp = afwDisplay.Display(frame=frame) 

881 with disp.Buffering(): 

882 for s in sSet: 

883 (xc, yc) = wcs.skyToPixel(s.getCoord().getRa(), s.getCoord().getDec()) 

884 xc -= xy0[0] 

885 yc -= xy0[1] 

886 disp.dot(symb, xc, yc, ctype=ctype, size=size) 

887 

888 

889def plotWhisker(results, newWcs): 

890 """Plot whisker diagram of astromeric offsets between results.matches. 

891 """ 

892 refCoordKey = results.matches[0].first.getTable().getCoordKey() 

893 inCentroidKey = results.matches[0].second.getTable().getCentroidSlot().getMeasKey() 

894 positions = [m.first.get(refCoordKey) for m in results.matches] 

895 residuals = [m.first.get(refCoordKey).getOffsetFrom( 

896 newWcs.pixelToSky(m.second.get(inCentroidKey))) for 

897 m in results.matches] 

898 import matplotlib.pyplot as plt 

899 fig = plt.figure() 

900 sp = fig.add_subplot(1, 1, 0) 

901 xpos = [x[0].asDegrees() for x in positions] 

902 ypos = [x[1].asDegrees() for x in positions] 

903 xpos.append(0.02*(max(xpos) - min(xpos)) + min(xpos)) 

904 ypos.append(0.98*(max(ypos) - min(ypos)) + min(ypos)) 

905 xidxs = np.isfinite(xpos) 

906 yidxs = np.isfinite(ypos) 

907 X = np.asarray(xpos)[xidxs] 

908 Y = np.asarray(ypos)[yidxs] 

909 distance = [x[1].asArcseconds() for x in residuals] 

910 distance.append(0.2) 

911 distance = np.asarray(distance)[xidxs] 

912 # NOTE: This assumes that the bearing is measured positive from +RA through North. 

913 # From the documentation this is not clear. 

914 bearing = [x[0].asRadians() for x in residuals] 

915 bearing.append(0) 

916 bearing = np.asarray(bearing)[xidxs] 

917 U = (distance*np.cos(bearing)) 

918 V = (distance*np.sin(bearing)) 

919 sp.quiver(X, Y, U, V) 

920 sp.set_title("WCS Residual") 

921 plt.show() 

922 

923 

924class DipoleTestImage(object): 

925 """Utility class for dipole measurement testing. 

926 

927 Generate an image with simulated dipoles and noise; store the original 

928 "pre-subtraction" images and catalogs as well. 

929 Used to generate test data for DMTN-007 (http://dmtn-007.lsst.io). 

930 """ 

931 

932 def __init__(self, w=101, h=101, xcenPos=[27.], ycenPos=[25.], xcenNeg=[23.], ycenNeg=[25.], 

933 psfSigma=2., flux=[30000.], fluxNeg=None, noise=10., gradientParams=None): 

934 self.w = w 

935 self.h = h 

936 self.xcenPos = xcenPos 

937 self.ycenPos = ycenPos 

938 self.xcenNeg = xcenNeg 

939 self.ycenNeg = ycenNeg 

940 self.psfSigma = psfSigma 

941 self.flux = flux 

942 self.fluxNeg = fluxNeg 

943 if fluxNeg is None: 

944 self.fluxNeg = self.flux 

945 self.noise = noise 

946 self.gradientParams = gradientParams 

947 self._makeDipoleImage() 

948 

949 def _makeDipoleImage(self): 

950 """Generate an exposure and catalog with the given dipole source(s). 

951 """ 

952 # Must seed the pos/neg images with different values to ensure they get different noise realizations 

953 posImage, posCatalog = self._makeStarImage( 

954 xc=self.xcenPos, yc=self.ycenPos, flux=self.flux, randomSeed=111) 

955 

956 negImage, negCatalog = self._makeStarImage( 

957 xc=self.xcenNeg, yc=self.ycenNeg, flux=self.fluxNeg, randomSeed=222) 

958 

959 dipole = posImage.clone() 

960 di = dipole.getMaskedImage() 

961 di -= negImage.getMaskedImage() 

962 

963 # Carry through pos/neg detection masks to new planes in diffim 

964 dm = di.getMask() 

965 posDetectedBits = posImage.getMaskedImage().getMask().getArray() == dm.getPlaneBitMask("DETECTED") 

966 negDetectedBits = negImage.getMaskedImage().getMask().getArray() == dm.getPlaneBitMask("DETECTED") 

967 pos_det = dm.addMaskPlane("DETECTED_POS") # new mask plane -- different from "DETECTED" 

968 neg_det = dm.addMaskPlane("DETECTED_NEG") # new mask plane -- different from "DETECTED_NEGATIVE" 

969 dma = dm.getArray() 

970 # set the two custom mask planes to these new masks 

971 dma[:, :] = posDetectedBits*pos_det + negDetectedBits*neg_det 

972 self.diffim, self.posImage, self.posCatalog, self.negImage, self.negCatalog \ 

973 = dipole, posImage, posCatalog, negImage, negCatalog 

974 

975 def _makeStarImage(self, xc=[15.3], yc=[18.6], flux=[2500], schema=None, randomSeed=None): 

976 """Generate an exposure and catalog with the given stellar source(s). 

977 """ 

978 from lsst.meas.base.tests import TestDataset 

979 bbox = geom.Box2I(geom.Point2I(0, 0), geom.Point2I(self.w - 1, self.h - 1)) 

980 dataset = TestDataset(bbox, psfSigma=self.psfSigma, threshold=1.) 

981 

982 for i in range(len(xc)): 

983 dataset.addSource(instFlux=flux[i], centroid=geom.Point2D(xc[i], yc[i])) 

984 

985 if schema is None: 

986 schema = TestDataset.makeMinimalSchema() 

987 exposure, catalog = dataset.realize(noise=self.noise, schema=schema, randomSeed=randomSeed) 

988 

989 if self.gradientParams is not None: 

990 y, x = np.mgrid[:self.w, :self.h] 

991 gp = self.gradientParams 

992 gradient = gp[0] + gp[1]*x + gp[2]*y 

993 if len(self.gradientParams) > 3: # it includes a set of 2nd-order polynomial params 

994 gradient += gp[3]*x*y + gp[4]*x*x + gp[5]*y*y 

995 imgArr = exposure.getMaskedImage().getArrays()[0] 

996 imgArr += gradient 

997 

998 return exposure, catalog 

999 

1000 def fitDipoleSource(self, source, **kwds): 

1001 alg = DipoleFitAlgorithm(self.diffim, self.posImage, self.negImage) 

1002 fitResult = alg.fitDipole(source, **kwds) 

1003 return fitResult 

1004 

1005 def detectDipoleSources(self, doMerge=True, diffim=None, detectSigma=5.5, grow=3, minBinSize=32): 

1006 """Utility function for detecting dipoles. 

1007 

1008 Detect pos/neg sources in the diffim, then merge them. A 

1009 bigger "grow" parameter leads to a larger footprint which 

1010 helps with dipole measurement for faint dipoles. 

1011 

1012 Parameters 

1013 ---------- 

1014 doMerge : `bool` 

1015 Whether to merge the positive and negagive detections into a single 

1016 source table. 

1017 diffim : `lsst.afw.image.exposure.exposure.ExposureF` 

1018 Difference image on which to perform detection. 

1019 detectSigma : `float` 

1020 Threshold for object detection. 

1021 grow : `int` 

1022 Number of pixels to grow the footprints before merging. 

1023 minBinSize : `int` 

1024 Minimum bin size for the background (re)estimation (only applies if 

1025 the default leads to min(nBinX, nBinY) < fit order so the default 

1026 config parameter needs to be decreased, but not to a value smaller 

1027 than ``minBinSize``, in which case the fitting algorithm will take 

1028 over and decrease the fit order appropriately.) 

1029 

1030 Returns 

1031 ------- 

1032 sources : `lsst.afw.table.SourceCatalog` 

1033 If doMerge=True, the merged source catalog is returned OR 

1034 detectTask : `lsst.meas.algorithms.SourceDetectionTask` 

1035 schema : `lsst.afw.table.Schema` 

1036 If doMerge=False, the source detection task and its schema are 

1037 returned. 

1038 """ 

1039 if diffim is None: 

1040 diffim = self.diffim 

1041 

1042 # Start with a minimal schema - only the fields all SourceCatalogs need 

1043 schema = afwTable.SourceTable.makeMinimalSchema() 

1044 

1045 # Customize the detection task a bit (optional) 

1046 detectConfig = measAlg.SourceDetectionConfig() 

1047 detectConfig.returnOriginalFootprints = False # should be the default 

1048 

1049 diffimPsf = diffim.getPsf() 

1050 psfSigma = diffimPsf.computeShape(diffimPsf.getAveragePosition()).getDeterminantRadius() 

1051 

1052 # code from imageDifference.py: 

1053 detectConfig.thresholdPolarity = "both" 

1054 detectConfig.thresholdValue = detectSigma 

1055 # detectConfig.nSigmaToGrow = psfSigma 

1056 detectConfig.reEstimateBackground = True # if False, will fail often for faint sources on gradients? 

1057 detectConfig.thresholdType = "pixel_stdev" 

1058 # Test images are often quite small, so may need to adjust background binSize 

1059 while ((min(diffim.getWidth(), diffim.getHeight()))//detectConfig.background.binSize 

1060 < detectConfig.background.approxOrderX and detectConfig.background.binSize > minBinSize): 

1061 detectConfig.background.binSize = max(minBinSize, detectConfig.background.binSize//2) 

1062 

1063 # Create the detection task. We pass the schema so the task can declare a few flag fields 

1064 detectTask = measAlg.SourceDetectionTask(schema, config=detectConfig) 

1065 

1066 table = afwTable.SourceTable.make(schema) 

1067 catalog = detectTask.run(table, diffim, sigma=psfSigma) 

1068 

1069 # Now do the merge. 

1070 if doMerge: 

1071 fpSet = catalog.fpSets.positive 

1072 fpSet.merge(catalog.fpSets.negative, grow, grow, False) 

1073 sources = afwTable.SourceCatalog(table) 

1074 fpSet.makeSources(sources) 

1075 

1076 return sources 

1077 

1078 else: 

1079 return detectTask, schema