1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

757

758

759

760

761

762

763

764

765

766

767

768

769

770

771

772

773

774

775

776

777

778

779

780

781

782

783

784

785

786

787

788

789

790

791

792

793

794

795

796

797

798

799

800

801

802

803

804

805

806

807

808

809

810

811

812

813

814

815

816

817

818

819

820

821

822

823

824

825

826

827

828

829

830

831

832

833

834

835

836

837

838

839

840

841

842

843

844

845

846

847

848

849

850

851

852

853

854

855

856

857

858

859

860

861

862

863

864

865

866

867

868

869

870

871

872

873

874

875

876

877

878

879

880

881

882

883

884

885

886

887

888

889

890

891

892

893

894

895

896

897

898

899

900

901

902

903

904

905

906

907

908

909

910

911

912

913

914

915

916

917

918

919

920

921

922

923

924

925

926

927

928

929

930

931

932

933

934

935

936

937

938

939

940

941

942

943

944

945

946

947

948

949

950

951

952

953

954

955

956

957

958

959

960

961

962

963

964

965

966

967

968

969

970

971

972

973

974

975

976

977

978

979

980

981

982

983

984

985

986

987

988

989

990

991

992

993

994

995

996

997

998

999

1000

1001

1002

1003

1004

1005

1006

1007

1008

1009

1010

1011

1012

1013

1014

1015

1016

1017

1018

1019

1020

1021

1022

1023

1024

1025

1026

1027

1028

1029

1030

1031

1032

1033

1034

1035

1036

1037

1038

1039

1040

1041

1042

1043

1044

1045

1046

1047

1048

1049

1050

1051

1052

1053

1054

1055

1056

1057

1058

1059

1060

1061

1062

1063

1064

1065

1066

1067

1068

1069

1070

1071

1072

1073

1074

1075

1076

1077

1078

1079

1080

1081

1082

1083

1084

1085

1086

1087

1088

1089

1090

1091

1092

1093

1094

1095

1096

1097

1098

1099

1100

1101

1102

1103

1104

1105

1106

1107

1108

1109

1110

1111

1112

1113

1114

1115

1116

1117

1118

1119

1120

1121

1122

1123

1124

1125

1126

1127

1128

1129

1130

1131

1132

1133

1134

1135

1136

1137

1138

1139

1140

1141

1142

1143

1144

1145

1146

1147

1148

1149

1150

1151

1152

1153

1154

1155

1156

1157

1158

1159

1160

1161

1162

1163

1164

1165

1166

1167

1168

1169

1170

1171

1172

1173

1174

1175

1176

1177

1178

1179

1180

1181

1182

1183

1184

1185

1186

1187

1188

1189

1190

1191

1192

1193

1194

1195

1196

1197

1198

1199

1200

1201

1202

1203

1204

1205

1206

1207

1208

1209

1210

1211

1212

1213

1214

1215

1216

1217

1218

1219

1220

1221

1222

1223

1224

1225

1226

1227

1228

1229

1230

1231

1232

1233

1234

1235

1236

1237

1238

1239

1240

1241

1242

1243

1244

1245

1246

1247

1248

1249

1250

1251

1252

1253

1254

1255

1256

1257

1258

1259

1260

1261

1262

1263

1264

1265

1266

1267

1268

1269

1270

1271

1272

1273

1274

1275

1276

1277

1278

1279

1280

1281

1282

1283

1284

1285

1286

1287

1288

1289

1290

1291

1292

1293

1294

1295

1296

1297

1298

1299

1300

1301

1302

1303

1304

1305

1306

1307

1308

1309

1310

1311

1312

1313

1314

1315

1316

1317

1318

1319

1320

1321

1322

1323

1324

1325

1326

1327

1328

1329

1330

1331

1332

1333

1334

1335

1336

1337

1338

1339

1340

1341

1342

1343

1344

1345

1346

1347

1348

1349

1350

1351

1352

1353

1354

1355

1356

1357

1358

1359

1360

1361

1362

1363

1364

1365

1366

1367

1368

1369

1370

1371

1372

1373

1374

1375

1376

1377

1378

1379

1380

1381

1382

1383

1384

1385

1386

1387

1388

1389

1390

1391

1392

1393

1394

1395

1396

1397

1398

1399

1400

1401

1402

1403

1404

1405

1406

1407

1408

1409

1410

1411

1412

1413

1414

1415

1416

1417

1418

1419

1420

1421

1422

1423

1424

1425

1426

1427

1428

1429

1430

1431

1432

1433

1434

1435

1436

1437

1438

1439

1440

1441

1442

1443

1444

1445

1446

1447

1448

1449

1450

1451

1452

1453

1454

1455

1456

1457

1458

1459

1460

1461

1462

1463

1464

1465

1466

1467

1468

1469

1470

1471

1472

1473

1474

1475

1476

1477

1478

1479

1480

1481

1482

1483

1484

1485

1486

1487

1488

1489

1490

1491

1492

1493

1494

1495

1496

1497

1498

1499

1500

1501

1502

1503

1504

1505

1506

1507

1508

1509

1510

1511

1512

1513

1514

1515

1516

1517

1518

1519

1520

1521

1522

1523

1524

1525

1526

1527

1528

1529

1530

1531

1532

1533

1534

1535

1536

1537

1538

1539

1540

1541

1542

1543

1544

1545

1546

1547

1548

1549

1550

1551

1552

1553

1554

1555

1556

1557

1558

1559

1560

1561

1562

1563

1564

1565

1566

1567

1568

1569

1570

1571

1572

1573

1574

1575

1576

1577

1578

1579

1580

1581

1582

1583

1584

1585

1586

1587

1588

1589

1590

1591

1592

1593

1594

1595

1596

1597

1598

1599

1600

1601

1602

1603

1604

1605

1606

1607

1608

1609

1610

1611

1612

1613

1614

1615

1616

1617

1618

1619

1620

1621

1622

1623

1624

1625

1626

1627

1628

1629

1630

1631

1632

1633

1634

1635

1636

1637

1638

1639

1640

1641

1642

1643

1644

1645

1646

1647

1648

1649

1650

1651

1652

1653

1654

1655

1656

1657

1658

1659

1660

1661

1662

1663

1664

1665

1666

1667

1668

1669

1670

1671

1672

1673

1674

1675

1676

1677

1678

1679

1680

1681

1682

1683

1684

1685

1686

1687

1688

1689

1690

1691

1692

1693

1694

1695

1696

1697

1698

1699

1700

1701

1702

1703

1704

1705

1706

1707

1708

1709

1710

""" 

fitpack --- curve and surface fitting with splines 

 

fitpack is based on a collection of Fortran routines DIERCKX 

by P. Dierckx (see http://www.netlib.org/dierckx/) transformed 

to double routines by Pearu Peterson. 

""" 

# Created by Pearu Peterson, June,August 2003 

from __future__ import division, print_function, absolute_import 

 

__all__ = [ 

'UnivariateSpline', 

'InterpolatedUnivariateSpline', 

'LSQUnivariateSpline', 

'BivariateSpline', 

'LSQBivariateSpline', 

'SmoothBivariateSpline', 

'LSQSphereBivariateSpline', 

'SmoothSphereBivariateSpline', 

'RectBivariateSpline', 

'RectSphereBivariateSpline'] 

 

 

import warnings 

 

from numpy import zeros, concatenate, alltrue, ravel, all, diff, array, ones 

import numpy as np 

 

from . import fitpack 

from . import dfitpack 

 

 

################ Univariate spline #################### 

 

_curfit_messages = {1:""" 

The required storage space exceeds the available storage space, as 

specified by the parameter nest: nest too small. If nest is already 

large (say nest > m/2), it may also indicate that s is too small. 

The approximation returned is the weighted least-squares spline 

according to the knots t[0],t[1],...,t[n-1]. (n=nest) the parameter fp 

gives the corresponding weighted sum of squared residuals (fp>s). 

""", 

2:""" 

A theoretically impossible result was found during the iteration 

process for finding a smoothing spline with fp = s: s too small. 

There is an approximation returned but the corresponding weighted sum 

of squared residuals does not satisfy the condition abs(fp-s)/s < tol.""", 

3:""" 

The maximal number of iterations maxit (set to 20 by the program) 

allowed for finding a smoothing spline with fp=s has been reached: s 

too small. 

There is an approximation returned but the corresponding weighted sum 

of squared residuals does not satisfy the condition abs(fp-s)/s < tol.""", 

10:""" 

Error on entry, no approximation returned. The following conditions 

must hold: 

xb<=x[0]<x[1]<...<x[m-1]<=xe, w[i]>0, i=0..m-1 

if iopt=-1: 

xb<t[k+1]<t[k+2]<...<t[n-k-2]<xe""" 

} 

 

 

# UnivariateSpline, ext parameter can be an int or a string 

_extrap_modes = {0: 0, 'extrapolate': 0, 

1: 1, 'zeros': 1, 

2: 2, 'raise': 2, 

3: 3, 'const': 3} 

 

 

class UnivariateSpline(object): 

""" 

One-dimensional smoothing spline fit to a given set of data points. 

 

Fits a spline y = spl(x) of degree `k` to the provided `x`, `y` data. `s` 

specifies the number of knots by specifying a smoothing condition. 

 

Parameters 

---------- 

x : (N,) array_like 

1-D array of independent input data. Must be increasing. 

y : (N,) array_like 

1-D array of dependent input data, of the same length as `x`. 

w : (N,) array_like, optional 

Weights for spline fitting. Must be positive. If None (default), 

weights are all equal. 

bbox : (2,) array_like, optional 

2-sequence specifying the boundary of the approximation interval. If 

None (default), ``bbox=[x[0], x[-1]]``. 

k : int, optional 

Degree of the smoothing spline. Must be <= 5. 

Default is k=3, a cubic spline. 

s : float or None, optional 

Positive smoothing factor used to choose the number of knots. Number 

of knots will be increased until the smoothing condition is satisfied:: 

 

sum((w[i] * (y[i]-spl(x[i])))**2, axis=0) <= s 

 

If None (default), ``s = len(w)`` which should be a good value if 

``1/w[i]`` is an estimate of the standard deviation of ``y[i]``. 

If 0, spline will interpolate through all data points. 

ext : int or str, optional 

Controls the extrapolation mode for elements 

not in the interval defined by the knot sequence. 

 

* if ext=0 or 'extrapolate', return the extrapolated value. 

* if ext=1 or 'zeros', return 0 

* if ext=2 or 'raise', raise a ValueError 

* if ext=3 of 'const', return the boundary value. 

 

The default value is 0. 

 

check_finite : bool, optional 

Whether to check that the input arrays contain only finite numbers. 

Disabling may give a performance gain, but may result in problems 

(crashes, non-termination or non-sensical results) if the inputs 

do contain infinities or NaNs. 

Default is False. 

 

See Also 

-------- 

InterpolatedUnivariateSpline : Subclass with smoothing forced to 0 

LSQUnivariateSpline : Subclass in which knots are user-selected instead of 

being set by smoothing condition 

splrep : An older, non object-oriented wrapping of FITPACK 

splev, sproot, splint, spalde 

BivariateSpline : A similar class for two-dimensional spline interpolation 

 

Notes 

----- 

The number of data points must be larger than the spline degree `k`. 

 

**NaN handling**: If the input arrays contain ``nan`` values, the result 

is not useful, since the underlying spline fitting routines cannot deal 

with ``nan`` . A workaround is to use zero weights for not-a-number 

data points: 

 

>>> from scipy.interpolate import UnivariateSpline 

>>> x, y = np.array([1, 2, 3, 4]), np.array([1, np.nan, 3, 4]) 

>>> w = np.isnan(y) 

>>> y[w] = 0. 

>>> spl = UnivariateSpline(x, y, w=~w) 

 

Notice the need to replace a ``nan`` by a numerical value (precise value 

does not matter as long as the corresponding weight is zero.) 

 

Examples 

-------- 

>>> import matplotlib.pyplot as plt 

>>> from scipy.interpolate import UnivariateSpline 

>>> x = np.linspace(-3, 3, 50) 

>>> y = np.exp(-x**2) + 0.1 * np.random.randn(50) 

>>> plt.plot(x, y, 'ro', ms=5) 

 

Use the default value for the smoothing parameter: 

 

>>> spl = UnivariateSpline(x, y) 

>>> xs = np.linspace(-3, 3, 1000) 

>>> plt.plot(xs, spl(xs), 'g', lw=3) 

 

Manually change the amount of smoothing: 

 

>>> spl.set_smoothing_factor(0.5) 

>>> plt.plot(xs, spl(xs), 'b', lw=3) 

>>> plt.show() 

 

""" 

def __init__(self, x, y, w=None, bbox=[None]*2, k=3, s=None, 

ext=0, check_finite=False): 

 

if check_finite: 

w_finite = np.isfinite(w).all() if w is not None else True 

if (not np.isfinite(x).all() or not np.isfinite(y).all() or 

not w_finite): 

raise ValueError("x and y array must not contain NaNs or infs.") 

if not all(diff(x) > 0.0): 

raise ValueError('x must be strictly increasing') 

 

# _data == x,y,w,xb,xe,k,s,n,t,c,fp,fpint,nrdata,ier 

try: 

self.ext = _extrap_modes[ext] 

except KeyError: 

raise ValueError("Unknown extrapolation mode %s." % ext) 

 

data = dfitpack.fpcurf0(x,y,k,w=w, 

xb=bbox[0],xe=bbox[1],s=s) 

if data[-1] == 1: 

# nest too small, setting to maximum bound 

data = self._reset_nest(data) 

self._data = data 

self._reset_class() 

 

@classmethod 

def _from_tck(cls, tck, ext=0): 

"""Construct a spline object from given tck""" 

self = cls.__new__(cls) 

t, c, k = tck 

self._eval_args = tck 

#_data == x,y,w,xb,xe,k,s,n,t,c,fp,fpint,nrdata,ier 

self._data = (None,None,None,None,None,k,None,len(t),t, 

c,None,None,None,None) 

self.ext = ext 

return self 

 

def _reset_class(self): 

data = self._data 

n,t,c,k,ier = data[7],data[8],data[9],data[5],data[-1] 

self._eval_args = t[:n],c[:n],k 

if ier == 0: 

# the spline returned has a residual sum of squares fp 

# such that abs(fp-s)/s <= tol with tol a relative 

# tolerance set to 0.001 by the program 

pass 

elif ier == -1: 

# the spline returned is an interpolating spline 

self._set_class(InterpolatedUnivariateSpline) 

elif ier == -2: 

# the spline returned is the weighted least-squares 

# polynomial of degree k. In this extreme case fp gives 

# the upper bound fp0 for the smoothing factor s. 

self._set_class(LSQUnivariateSpline) 

else: 

# error 

if ier == 1: 

self._set_class(LSQUnivariateSpline) 

message = _curfit_messages.get(ier,'ier=%s' % (ier)) 

warnings.warn(message) 

 

def _set_class(self, cls): 

self._spline_class = cls 

if self.__class__ in (UnivariateSpline, InterpolatedUnivariateSpline, 

LSQUnivariateSpline): 

self.__class__ = cls 

else: 

# It's an unknown subclass -- don't change class. cf. #731 

pass 

 

def _reset_nest(self, data, nest=None): 

n = data[10] 

if nest is None: 

k,m = data[5],len(data[0]) 

nest = m+k+1 # this is the maximum bound for nest 

else: 

if not n <= nest: 

raise ValueError("`nest` can only be increased") 

t, c, fpint, nrdata = [np.resize(data[j], nest) for j in [8,9,11,12]] 

 

args = data[:8] + (t,c,n,fpint,nrdata,data[13]) 

data = dfitpack.fpcurf1(*args) 

return data 

 

def set_smoothing_factor(self, s): 

""" Continue spline computation with the given smoothing 

factor s and with the knots found at the last call. 

 

This routine modifies the spline in place. 

 

""" 

data = self._data 

if data[6] == -1: 

warnings.warn('smoothing factor unchanged for' 

'LSQ spline with fixed knots') 

return 

args = data[:6] + (s,) + data[7:] 

data = dfitpack.fpcurf1(*args) 

if data[-1] == 1: 

# nest too small, setting to maximum bound 

data = self._reset_nest(data) 

self._data = data 

self._reset_class() 

 

def __call__(self, x, nu=0, ext=None): 

""" 

Evaluate spline (or its nu-th derivative) at positions x. 

 

Parameters 

---------- 

x : array_like 

A 1-D array of points at which to return the value of the smoothed 

spline or its derivatives. Note: x can be unordered but the 

evaluation is more efficient if x is (partially) ordered. 

nu : int 

The order of derivative of the spline to compute. 

ext : int 

Controls the value returned for elements of ``x`` not in the 

interval defined by the knot sequence. 

 

* if ext=0 or 'extrapolate', return the extrapolated value. 

* if ext=1 or 'zeros', return 0 

* if ext=2 or 'raise', raise a ValueError 

* if ext=3 or 'const', return the boundary value. 

 

The default value is 0, passed from the initialization of 

UnivariateSpline. 

 

""" 

x = np.asarray(x) 

# empty input yields empty output 

if x.size == 0: 

return array([]) 

# if nu is None: 

# return dfitpack.splev(*(self._eval_args+(x,))) 

# return dfitpack.splder(nu=nu,*(self._eval_args+(x,))) 

if ext is None: 

ext = self.ext 

else: 

try: 

ext = _extrap_modes[ext] 

except KeyError: 

raise ValueError("Unknown extrapolation mode %s." % ext) 

return fitpack.splev(x, self._eval_args, der=nu, ext=ext) 

 

def get_knots(self): 

""" Return positions of interior knots of the spline. 

 

Internally, the knot vector contains ``2*k`` additional boundary knots. 

""" 

data = self._data 

k,n = data[5],data[7] 

return data[8][k:n-k] 

 

def get_coeffs(self): 

"""Return spline coefficients.""" 

data = self._data 

k,n = data[5],data[7] 

return data[9][:n-k-1] 

 

def get_residual(self): 

"""Return weighted sum of squared residuals of the spline approximation. 

 

This is equivalent to:: 

 

sum((w[i] * (y[i]-spl(x[i])))**2, axis=0) 

 

""" 

return self._data[10] 

 

def integral(self, a, b): 

""" Return definite integral of the spline between two given points. 

 

Parameters 

---------- 

a : float 

Lower limit of integration. 

b : float 

Upper limit of integration. 

 

Returns 

------- 

integral : float 

The value of the definite integral of the spline between limits. 

 

Examples 

-------- 

>>> from scipy.interpolate import UnivariateSpline 

>>> x = np.linspace(0, 3, 11) 

>>> y = x**2 

>>> spl = UnivariateSpline(x, y) 

>>> spl.integral(0, 3) 

9.0 

 

which agrees with :math:`\\int x^2 dx = x^3 / 3` between the limits 

of 0 and 3. 

 

A caveat is that this routine assumes the spline to be zero outside of 

the data limits: 

 

>>> spl.integral(-1, 4) 

9.0 

>>> spl.integral(-1, 0) 

0.0 

 

""" 

return dfitpack.splint(*(self._eval_args+(a,b))) 

 

def derivatives(self, x): 

""" Return all derivatives of the spline at the point x. 

 

Parameters 

---------- 

x : float 

The point to evaluate the derivatives at. 

 

Returns 

------- 

der : ndarray, shape(k+1,) 

Derivatives of the orders 0 to k. 

 

Examples 

-------- 

>>> from scipy.interpolate import UnivariateSpline 

>>> x = np.linspace(0, 3, 11) 

>>> y = x**2 

>>> spl = UnivariateSpline(x, y) 

>>> spl.derivatives(1.5) 

array([2.25, 3.0, 2.0, 0]) 

 

""" 

d,ier = dfitpack.spalde(*(self._eval_args+(x,))) 

if not ier == 0: 

raise ValueError("Error code returned by spalde: %s" % ier) 

return d 

 

def roots(self): 

""" Return the zeros of the spline. 

 

Restriction: only cubic splines are supported by fitpack. 

""" 

k = self._data[5] 

if k == 3: 

z,m,ier = dfitpack.sproot(*self._eval_args[:2]) 

if not ier == 0: 

raise ValueError("Error code returned by spalde: %s" % ier) 

return z[:m] 

raise NotImplementedError('finding roots unsupported for ' 

'non-cubic splines') 

 

def derivative(self, n=1): 

""" 

Construct a new spline representing the derivative of this spline. 

 

Parameters 

---------- 

n : int, optional 

Order of derivative to evaluate. Default: 1 

 

Returns 

------- 

spline : UnivariateSpline 

Spline of order k2=k-n representing the derivative of this 

spline. 

 

See Also 

-------- 

splder, antiderivative 

 

Notes 

----- 

 

.. versionadded:: 0.13.0 

 

Examples 

-------- 

This can be used for finding maxima of a curve: 

 

>>> from scipy.interpolate import UnivariateSpline 

>>> x = np.linspace(0, 10, 70) 

>>> y = np.sin(x) 

>>> spl = UnivariateSpline(x, y, k=4, s=0) 

 

Now, differentiate the spline and find the zeros of the 

derivative. (NB: `sproot` only works for order 3 splines, so we 

fit an order 4 spline): 

 

>>> spl.derivative().roots() / np.pi 

array([ 0.50000001, 1.5 , 2.49999998]) 

 

This agrees well with roots :math:`\\pi/2 + n\\pi` of 

:math:`\\cos(x) = \\sin'(x)`. 

 

""" 

tck = fitpack.splder(self._eval_args, n) 

return UnivariateSpline._from_tck(tck, self.ext) 

 

def antiderivative(self, n=1): 

""" 

Construct a new spline representing the antiderivative of this spline. 

 

Parameters 

---------- 

n : int, optional 

Order of antiderivative to evaluate. Default: 1 

 

Returns 

------- 

spline : UnivariateSpline 

Spline of order k2=k+n representing the antiderivative of this 

spline. 

 

Notes 

----- 

 

.. versionadded:: 0.13.0 

 

See Also 

-------- 

splantider, derivative 

 

Examples 

-------- 

>>> from scipy.interpolate import UnivariateSpline 

>>> x = np.linspace(0, np.pi/2, 70) 

>>> y = 1 / np.sqrt(1 - 0.8*np.sin(x)**2) 

>>> spl = UnivariateSpline(x, y, s=0) 

 

The derivative is the inverse operation of the antiderivative, 

although some floating point error accumulates: 

 

>>> spl(1.7), spl.antiderivative().derivative()(1.7) 

(array(2.1565429877197317), array(2.1565429877201865)) 

 

Antiderivative can be used to evaluate definite integrals: 

 

>>> ispl = spl.antiderivative() 

>>> ispl(np.pi/2) - ispl(0) 

2.2572053588768486 

 

This is indeed an approximation to the complete elliptic integral 

:math:`K(m) = \\int_0^{\\pi/2} [1 - m\\sin^2 x]^{-1/2} dx`: 

 

>>> from scipy.special import ellipk 

>>> ellipk(0.8) 

2.2572053268208538 

 

""" 

tck = fitpack.splantider(self._eval_args, n) 

return UnivariateSpline._from_tck(tck, self.ext) 

 

 

class InterpolatedUnivariateSpline(UnivariateSpline): 

""" 

One-dimensional interpolating spline for a given set of data points. 

 

Fits a spline y = spl(x) of degree `k` to the provided `x`, `y` data. Spline 

function passes through all provided points. Equivalent to 

`UnivariateSpline` with s=0. 

 

Parameters 

---------- 

x : (N,) array_like 

Input dimension of data points -- must be increasing 

y : (N,) array_like 

input dimension of data points 

w : (N,) array_like, optional 

Weights for spline fitting. Must be positive. If None (default), 

weights are all equal. 

bbox : (2,) array_like, optional 

2-sequence specifying the boundary of the approximation interval. If 

None (default), ``bbox=[x[0], x[-1]]``. 

k : int, optional 

Degree of the smoothing spline. Must be 1 <= `k` <= 5. 

ext : int or str, optional 

Controls the extrapolation mode for elements 

not in the interval defined by the knot sequence. 

 

* if ext=0 or 'extrapolate', return the extrapolated value. 

* if ext=1 or 'zeros', return 0 

* if ext=2 or 'raise', raise a ValueError 

* if ext=3 of 'const', return the boundary value. 

 

The default value is 0. 

 

check_finite : bool, optional 

Whether to check that the input arrays contain only finite numbers. 

Disabling may give a performance gain, but may result in problems 

(crashes, non-termination or non-sensical results) if the inputs 

do contain infinities or NaNs. 

Default is False. 

 

See Also 

-------- 

UnivariateSpline : Superclass -- allows knots to be selected by a 

smoothing condition 

LSQUnivariateSpline : spline for which knots are user-selected 

splrep : An older, non object-oriented wrapping of FITPACK 

splev, sproot, splint, spalde 

BivariateSpline : A similar class for two-dimensional spline interpolation 

 

Notes 

----- 

The number of data points must be larger than the spline degree `k`. 

 

Examples 

-------- 

>>> import matplotlib.pyplot as plt 

>>> from scipy.interpolate import InterpolatedUnivariateSpline 

>>> x = np.linspace(-3, 3, 50) 

>>> y = np.exp(-x**2) + 0.1 * np.random.randn(50) 

>>> spl = InterpolatedUnivariateSpline(x, y) 

>>> plt.plot(x, y, 'ro', ms=5) 

>>> xs = np.linspace(-3, 3, 1000) 

>>> plt.plot(xs, spl(xs), 'g', lw=3, alpha=0.7) 

>>> plt.show() 

 

Notice that the ``spl(x)`` interpolates `y`: 

 

>>> spl.get_residual() 

0.0 

 

""" 

def __init__(self, x, y, w=None, bbox=[None]*2, k=3, 

ext=0, check_finite=False): 

 

if check_finite: 

w_finite = np.isfinite(w).all() if w is not None else True 

if (not np.isfinite(x).all() or not np.isfinite(y).all() or 

not w_finite): 

raise ValueError("Input must not contain NaNs or infs.") 

if not all(diff(x) > 0.0): 

raise ValueError('x must be strictly increasing') 

 

# _data == x,y,w,xb,xe,k,s,n,t,c,fp,fpint,nrdata,ier 

self._data = dfitpack.fpcurf0(x,y,k,w=w, 

xb=bbox[0],xe=bbox[1],s=0) 

self._reset_class() 

 

try: 

self.ext = _extrap_modes[ext] 

except KeyError: 

raise ValueError("Unknown extrapolation mode %s." % ext) 

 

 

_fpchec_error_string = """The input parameters have been rejected by fpchec. \ 

This means that at least one of the following conditions is violated: 

 

1) k+1 <= n-k-1 <= m 

2) t(1) <= t(2) <= ... <= t(k+1) 

t(n-k) <= t(n-k+1) <= ... <= t(n) 

3) t(k+1) < t(k+2) < ... < t(n-k) 

4) t(k+1) <= x(i) <= t(n-k) 

5) The conditions specified by Schoenberg and Whitney must hold 

for at least one subset of data points, i.e., there must be a 

subset of data points y(j) such that 

t(j) < y(j) < t(j+k+1), j=1,2,...,n-k-1 

""" 

 

 

class LSQUnivariateSpline(UnivariateSpline): 

""" 

One-dimensional spline with explicit internal knots. 

 

Fits a spline y = spl(x) of degree `k` to the provided `x`, `y` data. `t` 

specifies the internal knots of the spline 

 

Parameters 

---------- 

x : (N,) array_like 

Input dimension of data points -- must be increasing 

y : (N,) array_like 

Input dimension of data points 

t : (M,) array_like 

interior knots of the spline. Must be in ascending order and:: 

 

bbox[0] < t[0] < ... < t[-1] < bbox[-1] 

 

w : (N,) array_like, optional 

weights for spline fitting. Must be positive. If None (default), 

weights are all equal. 

bbox : (2,) array_like, optional 

2-sequence specifying the boundary of the approximation interval. If 

None (default), ``bbox = [x[0], x[-1]]``. 

k : int, optional 

Degree of the smoothing spline. Must be 1 <= `k` <= 5. 

Default is k=3, a cubic spline. 

ext : int or str, optional 

Controls the extrapolation mode for elements 

not in the interval defined by the knot sequence. 

 

* if ext=0 or 'extrapolate', return the extrapolated value. 

* if ext=1 or 'zeros', return 0 

* if ext=2 or 'raise', raise a ValueError 

* if ext=3 of 'const', return the boundary value. 

 

The default value is 0. 

 

check_finite : bool, optional 

Whether to check that the input arrays contain only finite numbers. 

Disabling may give a performance gain, but may result in problems 

(crashes, non-termination or non-sensical results) if the inputs 

do contain infinities or NaNs. 

Default is False. 

 

Raises 

------ 

ValueError 

If the interior knots do not satisfy the Schoenberg-Whitney conditions 

 

See Also 

-------- 

UnivariateSpline : Superclass -- knots are specified by setting a 

smoothing condition 

InterpolatedUnivariateSpline : spline passing through all points 

splrep : An older, non object-oriented wrapping of FITPACK 

splev, sproot, splint, spalde 

BivariateSpline : A similar class for two-dimensional spline interpolation 

 

Notes 

----- 

The number of data points must be larger than the spline degree `k`. 

 

Knots `t` must satisfy the Schoenberg-Whitney conditions, 

i.e., there must be a subset of data points ``x[j]`` such that 

``t[j] < x[j] < t[j+k+1]``, for ``j=0, 1,...,n-k-2``. 

 

Examples 

-------- 

>>> from scipy.interpolate import LSQUnivariateSpline, UnivariateSpline 

>>> import matplotlib.pyplot as plt 

>>> x = np.linspace(-3, 3, 50) 

>>> y = np.exp(-x**2) + 0.1 * np.random.randn(50) 

 

Fit a smoothing spline with a pre-defined internal knots: 

 

>>> t = [-1, 0, 1] 

>>> spl = LSQUnivariateSpline(x, y, t) 

 

>>> xs = np.linspace(-3, 3, 1000) 

>>> plt.plot(x, y, 'ro', ms=5) 

>>> plt.plot(xs, spl(xs), 'g-', lw=3) 

>>> plt.show() 

 

Check the knot vector: 

 

>>> spl.get_knots() 

array([-3., -1., 0., 1., 3.]) 

 

Constructing lsq spline using the knots from another spline: 

 

>>> x = np.arange(10) 

>>> s = UnivariateSpline(x, x, s=0) 

>>> s.get_knots() 

array([ 0., 2., 3., 4., 5., 6., 7., 9.]) 

>>> knt = s.get_knots() 

>>> s1 = LSQUnivariateSpline(x, x, knt[1:-1]) # Chop 1st and last knot 

>>> s1.get_knots() 

array([ 0., 2., 3., 4., 5., 6., 7., 9.]) 

 

""" 

 

def __init__(self, x, y, t, w=None, bbox=[None]*2, k=3, 

ext=0, check_finite=False): 

 

if check_finite: 

w_finite = np.isfinite(w).all() if w is not None else True 

if (not np.isfinite(x).all() or not np.isfinite(y).all() or 

not w_finite or not np.isfinite(t).all()): 

raise ValueError("Input(s) must not contain NaNs or infs.") 

if not all(diff(x) > 0.0): 

raise ValueError('x must be strictly increasing') 

 

# _data == x,y,w,xb,xe,k,s,n,t,c,fp,fpint,nrdata,ier 

xb = bbox[0] 

xe = bbox[1] 

if xb is None: 

xb = x[0] 

if xe is None: 

xe = x[-1] 

t = concatenate(([xb]*(k+1), t, [xe]*(k+1))) 

n = len(t) 

if not alltrue(t[k+1:n-k]-t[k:n-k-1] > 0, axis=0): 

raise ValueError('Interior knots t must satisfy ' 

'Schoenberg-Whitney conditions') 

if not dfitpack.fpchec(x, t, k) == 0: 

raise ValueError(_fpchec_error_string) 

data = dfitpack.fpcurfm1(x, y, k, t, w=w, xb=xb, xe=xe) 

self._data = data[:-3] + (None, None, data[-1]) 

self._reset_class() 

 

try: 

self.ext = _extrap_modes[ext] 

except KeyError: 

raise ValueError("Unknown extrapolation mode %s." % ext) 

 

 

################ Bivariate spline #################### 

 

class _BivariateSplineBase(object): 

""" Base class for Bivariate spline s(x,y) interpolation on the rectangle 

[xb,xe] x [yb, ye] calculated from a given set of data points 

(x,y,z). 

 

See Also 

-------- 

bisplrep, bisplev : an older wrapping of FITPACK 

BivariateSpline : 

implementation of bivariate spline interpolation on a plane grid 

SphereBivariateSpline : 

implementation of bivariate spline interpolation on a spherical grid 

""" 

 

def get_residual(self): 

""" Return weighted sum of squared residuals of the spline 

approximation: sum ((w[i]*(z[i]-s(x[i],y[i])))**2,axis=0) 

""" 

return self.fp 

 

def get_knots(self): 

""" Return a tuple (tx,ty) where tx,ty contain knots positions 

of the spline with respect to x-, y-variable, respectively. 

The position of interior and additional knots are given as 

t[k+1:-k-1] and t[:k+1]=b, t[-k-1:]=e, respectively. 

""" 

return self.tck[:2] 

 

def get_coeffs(self): 

""" Return spline coefficients.""" 

return self.tck[2] 

 

def __call__(self, x, y, dx=0, dy=0, grid=True): 

""" 

Evaluate the spline or its derivatives at given positions. 

 

Parameters 

---------- 

x, y : array_like 

Input coordinates. 

 

If `grid` is False, evaluate the spline at points ``(x[i], 

y[i]), i=0, ..., len(x)-1``. Standard Numpy broadcasting 

is obeyed. 

 

If `grid` is True: evaluate spline at the grid points 

defined by the coordinate arrays x, y. The arrays must be 

sorted to increasing order. 

 

Note that the axis ordering is inverted relative to 

the output of meshgrid. 

dx : int 

Order of x-derivative 

 

.. versionadded:: 0.14.0 

dy : int 

Order of y-derivative 

 

.. versionadded:: 0.14.0 

grid : bool 

Whether to evaluate the results on a grid spanned by the 

input arrays, or at points specified by the input arrays. 

 

.. versionadded:: 0.14.0 

 

""" 

x = np.asarray(x) 

y = np.asarray(y) 

 

tx, ty, c = self.tck[:3] 

kx, ky = self.degrees 

if grid: 

if x.size == 0 or y.size == 0: 

return np.zeros((x.size, y.size), dtype=self.tck[2].dtype) 

 

if dx or dy: 

z,ier = dfitpack.parder(tx,ty,c,kx,ky,dx,dy,x,y) 

if not ier == 0: 

raise ValueError("Error code returned by parder: %s" % ier) 

else: 

z,ier = dfitpack.bispev(tx,ty,c,kx,ky,x,y) 

if not ier == 0: 

raise ValueError("Error code returned by bispev: %s" % ier) 

else: 

# standard Numpy broadcasting 

if x.shape != y.shape: 

x, y = np.broadcast_arrays(x, y) 

 

shape = x.shape 

x = x.ravel() 

y = y.ravel() 

 

if x.size == 0 or y.size == 0: 

return np.zeros(shape, dtype=self.tck[2].dtype) 

 

if dx or dy: 

z,ier = dfitpack.pardeu(tx,ty,c,kx,ky,dx,dy,x,y) 

if not ier == 0: 

raise ValueError("Error code returned by pardeu: %s" % ier) 

else: 

z,ier = dfitpack.bispeu(tx,ty,c,kx,ky,x,y) 

if not ier == 0: 

raise ValueError("Error code returned by bispeu: %s" % ier) 

 

z = z.reshape(shape) 

return z 

 

 

_surfit_messages = {1:""" 

The required storage space exceeds the available storage space: nxest 

or nyest too small, or s too small. 

The weighted least-squares spline corresponds to the current set of 

knots.""", 

2:""" 

A theoretically impossible result was found during the iteration 

process for finding a smoothing spline with fp = s: s too small or 

badly chosen eps. 

Weighted sum of squared residuals does not satisfy abs(fp-s)/s < tol.""", 

3:""" 

the maximal number of iterations maxit (set to 20 by the program) 

allowed for finding a smoothing spline with fp=s has been reached: 

s too small. 

Weighted sum of squared residuals does not satisfy abs(fp-s)/s < tol.""", 

4:""" 

No more knots can be added because the number of b-spline coefficients 

(nx-kx-1)*(ny-ky-1) already exceeds the number of data points m: 

either s or m too small. 

The weighted least-squares spline corresponds to the current set of 

knots.""", 

5:""" 

No more knots can be added because the additional knot would (quasi) 

coincide with an old one: s too small or too large a weight to an 

inaccurate data point. 

The weighted least-squares spline corresponds to the current set of 

knots.""", 

10:""" 

Error on entry, no approximation returned. The following conditions 

must hold: 

xb<=x[i]<=xe, yb<=y[i]<=ye, w[i]>0, i=0..m-1 

If iopt==-1, then 

xb<tx[kx+1]<tx[kx+2]<...<tx[nx-kx-2]<xe 

yb<ty[ky+1]<ty[ky+2]<...<ty[ny-ky-2]<ye""", 

-3:""" 

The coefficients of the spline returned have been computed as the 

minimal norm least-squares solution of a (numerically) rank deficient 

system (deficiency=%i). If deficiency is large, the results may be 

inaccurate. Deficiency may strongly depend on the value of eps.""" 

} 

 

 

class BivariateSpline(_BivariateSplineBase): 

""" 

Base class for bivariate splines. 

 

This describes a spline ``s(x, y)`` of degrees ``kx`` and ``ky`` on 

the rectangle ``[xb, xe] * [yb, ye]`` calculated from a given set 

of data points ``(x, y, z)``. 

 

This class is meant to be subclassed, not instantiated directly. 

To construct these splines, call either `SmoothBivariateSpline` or 

`LSQBivariateSpline`. 

 

See Also 

-------- 

UnivariateSpline : a similar class for univariate spline interpolation 

SmoothBivariateSpline : 

to create a BivariateSpline through the given points 

LSQBivariateSpline : 

to create a BivariateSpline using weighted least-squares fitting 

SphereBivariateSpline : 

bivariate spline interpolation in spherical cooridinates 

bisplrep : older wrapping of FITPACK 

bisplev : older wrapping of FITPACK 

 

""" 

 

@classmethod 

def _from_tck(cls, tck): 

"""Construct a spline object from given tck and degree""" 

self = cls.__new__(cls) 

if len(tck) != 5: 

raise ValueError("tck should be a 5 element tuple of tx, ty, c, kx, ky") 

self.tck = tck[:3] 

self.degrees = tck[3:] 

return self 

 

def ev(self, xi, yi, dx=0, dy=0): 

""" 

Evaluate the spline at points 

 

Returns the interpolated value at ``(xi[i], yi[i]), 

i=0,...,len(xi)-1``. 

 

Parameters 

---------- 

xi, yi : array_like 

Input coordinates. Standard Numpy broadcasting is obeyed. 

dx : int, optional 

Order of x-derivative 

 

.. versionadded:: 0.14.0 

dy : int, optional 

Order of y-derivative 

 

.. versionadded:: 0.14.0 

""" 

return self.__call__(xi, yi, dx=dx, dy=dy, grid=False) 

 

def integral(self, xa, xb, ya, yb): 

""" 

Evaluate the integral of the spline over area [xa,xb] x [ya,yb]. 

 

Parameters 

---------- 

xa, xb : float 

The end-points of the x integration interval. 

ya, yb : float 

The end-points of the y integration interval. 

 

Returns 

------- 

integ : float 

The value of the resulting integral. 

 

""" 

tx,ty,c = self.tck[:3] 

kx,ky = self.degrees 

return dfitpack.dblint(tx,ty,c,kx,ky,xa,xb,ya,yb) 

 

 

class SmoothBivariateSpline(BivariateSpline): 

""" 

Smooth bivariate spline approximation. 

 

Parameters 

---------- 

x, y, z : array_like 

1-D sequences of data points (order is not important). 

w : array_like, optional 

Positive 1-D sequence of weights, of same length as `x`, `y` and `z`. 

bbox : array_like, optional 

Sequence of length 4 specifying the boundary of the rectangular 

approximation domain. By default, 

``bbox=[min(x,tx),max(x,tx), min(y,ty),max(y,ty)]``. 

kx, ky : ints, optional 

Degrees of the bivariate spline. Default is 3. 

s : float, optional 

Positive smoothing factor defined for estimation condition: 

``sum((w[i]*(z[i]-s(x[i], y[i])))**2, axis=0) <= s`` 

Default ``s=len(w)`` which should be a good value if ``1/w[i]`` is an 

estimate of the standard deviation of ``z[i]``. 

eps : float, optional 

A threshold for determining the effective rank of an over-determined 

linear system of equations. `eps` should have a value between 0 and 1, 

the default is 1e-16. 

 

See Also 

-------- 

bisplrep : an older wrapping of FITPACK 

bisplev : an older wrapping of FITPACK 

UnivariateSpline : a similar class for univariate spline interpolation 

LSQUnivariateSpline : to create a BivariateSpline using weighted 

 

Notes 

----- 

The length of `x`, `y` and `z` should be at least ``(kx+1) * (ky+1)``. 

 

""" 

 

def __init__(self, x, y, z, w=None, bbox=[None] * 4, kx=3, ky=3, s=None, 

eps=None): 

xb,xe,yb,ye = bbox 

nx,tx,ny,ty,c,fp,wrk1,ier = dfitpack.surfit_smth(x,y,z,w, 

xb,xe,yb,ye, 

kx,ky,s=s, 

eps=eps,lwrk2=1) 

if ier > 10: # lwrk2 was to small, re-run 

nx,tx,ny,ty,c,fp,wrk1,ier = dfitpack.surfit_smth(x,y,z,w, 

xb,xe,yb,ye, 

kx,ky,s=s, 

eps=eps,lwrk2=ier) 

if ier in [0,-1,-2]: # normal return 

pass 

else: 

message = _surfit_messages.get(ier,'ier=%s' % (ier)) 

warnings.warn(message) 

 

self.fp = fp 

self.tck = tx[:nx],ty[:ny],c[:(nx-kx-1)*(ny-ky-1)] 

self.degrees = kx,ky 

 

 

class LSQBivariateSpline(BivariateSpline): 

""" 

Weighted least-squares bivariate spline approximation. 

 

Parameters 

---------- 

x, y, z : array_like 

1-D sequences of data points (order is not important). 

tx, ty : array_like 

Strictly ordered 1-D sequences of knots coordinates. 

w : array_like, optional 

Positive 1-D array of weights, of the same length as `x`, `y` and `z`. 

bbox : (4,) array_like, optional 

Sequence of length 4 specifying the boundary of the rectangular 

approximation domain. By default, 

``bbox=[min(x,tx),max(x,tx), min(y,ty),max(y,ty)]``. 

kx, ky : ints, optional 

Degrees of the bivariate spline. Default is 3. 

eps : float, optional 

A threshold for determining the effective rank of an over-determined 

linear system of equations. `eps` should have a value between 0 and 1, 

the default is 1e-16. 

 

See Also 

-------- 

bisplrep : an older wrapping of FITPACK 

bisplev : an older wrapping of FITPACK 

UnivariateSpline : a similar class for univariate spline interpolation 

SmoothBivariateSpline : create a smoothing BivariateSpline 

 

Notes 

----- 

The length of `x`, `y` and `z` should be at least ``(kx+1) * (ky+1)``. 

 

""" 

 

def __init__(self, x, y, z, tx, ty, w=None, bbox=[None]*4, kx=3, ky=3, 

eps=None): 

nx = 2*kx+2+len(tx) 

ny = 2*ky+2+len(ty) 

tx1 = zeros((nx,),float) 

ty1 = zeros((ny,),float) 

tx1[kx+1:nx-kx-1] = tx 

ty1[ky+1:ny-ky-1] = ty 

 

xb,xe,yb,ye = bbox 

tx1,ty1,c,fp,ier = dfitpack.surfit_lsq(x,y,z,tx1,ty1,w, 

xb,xe,yb,ye, 

kx,ky,eps,lwrk2=1) 

if ier > 10: 

tx1,ty1,c,fp,ier = dfitpack.surfit_lsq(x,y,z,tx1,ty1,w, 

xb,xe,yb,ye, 

kx,ky,eps,lwrk2=ier) 

if ier in [0,-1,-2]: # normal return 

pass 

else: 

if ier < -2: 

deficiency = (nx-kx-1)*(ny-ky-1)+ier 

message = _surfit_messages.get(-3) % (deficiency) 

else: 

message = _surfit_messages.get(ier, 'ier=%s' % (ier)) 

warnings.warn(message) 

self.fp = fp 

self.tck = tx1, ty1, c 

self.degrees = kx, ky 

 

 

class RectBivariateSpline(BivariateSpline): 

""" 

Bivariate spline approximation over a rectangular mesh. 

 

Can be used for both smoothing and interpolating data. 

 

Parameters 

---------- 

x,y : array_like 

1-D arrays of coordinates in strictly ascending order. 

z : array_like 

2-D array of data with shape (x.size,y.size). 

bbox : array_like, optional 

Sequence of length 4 specifying the boundary of the rectangular 

approximation domain. By default, 

``bbox=[min(x,tx),max(x,tx), min(y,ty),max(y,ty)]``. 

kx, ky : ints, optional 

Degrees of the bivariate spline. Default is 3. 

s : float, optional 

Positive smoothing factor defined for estimation condition: 

``sum((w[i]*(z[i]-s(x[i], y[i])))**2, axis=0) <= s`` 

Default is ``s=0``, which is for interpolation. 

 

See Also 

-------- 

SmoothBivariateSpline : a smoothing bivariate spline for scattered data 

bisplrep : an older wrapping of FITPACK 

bisplev : an older wrapping of FITPACK 

UnivariateSpline : a similar class for univariate spline interpolation 

 

""" 

 

def __init__(self, x, y, z, bbox=[None] * 4, kx=3, ky=3, s=0): 

x, y = ravel(x), ravel(y) 

if not all(diff(x) > 0.0): 

raise ValueError('x must be strictly increasing') 

if not all(diff(y) > 0.0): 

raise ValueError('y must be strictly increasing') 

if not ((x.min() == x[0]) and (x.max() == x[-1])): 

raise ValueError('x must be strictly ascending') 

if not ((y.min() == y[0]) and (y.max() == y[-1])): 

raise ValueError('y must be strictly ascending') 

if not x.size == z.shape[0]: 

raise ValueError('x dimension of z must have same number of ' 

'elements as x') 

if not y.size == z.shape[1]: 

raise ValueError('y dimension of z must have same number of ' 

'elements as y') 

z = ravel(z) 

xb, xe, yb, ye = bbox 

nx, tx, ny, ty, c, fp, ier = dfitpack.regrid_smth(x, y, z, xb, xe, yb, 

ye, kx, ky, s) 

 

if ier not in [0, -1, -2]: 

msg = _surfit_messages.get(ier, 'ier=%s' % (ier)) 

raise ValueError(msg) 

 

self.fp = fp 

self.tck = tx[:nx], ty[:ny], c[:(nx - kx - 1) * (ny - ky - 1)] 

self.degrees = kx, ky 

 

 

_spherefit_messages = _surfit_messages.copy() 

_spherefit_messages[10] = """ 

ERROR. On entry, the input data are controlled on validity. The following 

restrictions must be satisfied: 

-1<=iopt<=1, m>=2, ntest>=8 ,npest >=8, 0<eps<1, 

0<=teta(i)<=pi, 0<=phi(i)<=2*pi, w(i)>0, i=1,...,m 

lwrk1 >= 185+52*v+10*u+14*u*v+8*(u-1)*v**2+8*m 

kwrk >= m+(ntest-7)*(npest-7) 

if iopt=-1: 8<=nt<=ntest , 9<=np<=npest 

0<tt(5)<tt(6)<...<tt(nt-4)<pi 

0<tp(5)<tp(6)<...<tp(np-4)<2*pi 

if iopt>=0: s>=0 

if one of these conditions is found to be violated,control 

is immediately repassed to the calling program. in that 

case there is no approximation returned.""" 

_spherefit_messages[-3] = """ 

WARNING. The coefficients of the spline returned have been computed as the 

minimal norm least-squares solution of a (numerically) rank 

deficient system (deficiency=%i, rank=%i). Especially if the rank 

deficiency, which is computed by 6+(nt-8)*(np-7)+ier, is large, 

the results may be inaccurate. They could also seriously depend on 

the value of eps.""" 

 

 

class SphereBivariateSpline(_BivariateSplineBase): 

""" 

Bivariate spline s(x,y) of degrees 3 on a sphere, calculated from a 

given set of data points (theta,phi,r). 

 

.. versionadded:: 0.11.0 

 

See Also 

-------- 

bisplrep, bisplev : an older wrapping of FITPACK 

UnivariateSpline : a similar class for univariate spline interpolation 

SmoothUnivariateSpline : 

to create a BivariateSpline through the given points 

LSQUnivariateSpline : 

to create a BivariateSpline using weighted least-squares fitting 

""" 

 

def __call__(self, theta, phi, dtheta=0, dphi=0, grid=True): 

""" 

Evaluate the spline or its derivatives at given positions. 

 

Parameters 

---------- 

theta, phi : array_like 

Input coordinates. 

 

If `grid` is False, evaluate the spline at points 

``(theta[i], phi[i]), i=0, ..., len(x)-1``. Standard 

Numpy broadcasting is obeyed. 

 

If `grid` is True: evaluate spline at the grid points 

defined by the coordinate arrays theta, phi. The arrays 

must be sorted to increasing order. 

dtheta : int, optional 

Order of theta-derivative 

 

.. versionadded:: 0.14.0 

dphi : int 

Order of phi-derivative 

 

.. versionadded:: 0.14.0 

grid : bool 

Whether to evaluate the results on a grid spanned by the 

input arrays, or at points specified by the input arrays. 

 

.. versionadded:: 0.14.0 

 

""" 

theta = np.asarray(theta) 

phi = np.asarray(phi) 

 

if theta.size > 0 and (theta.min() < 0. or theta.max() > np.pi): 

raise ValueError("requested theta out of bounds.") 

if phi.size > 0 and (phi.min() < 0. or phi.max() > 2. * np.pi): 

raise ValueError("requested phi out of bounds.") 

 

return _BivariateSplineBase.__call__(self, theta, phi, 

dx=dtheta, dy=dphi, grid=grid) 

 

def ev(self, theta, phi, dtheta=0, dphi=0): 

""" 

Evaluate the spline at points 

 

Returns the interpolated value at ``(theta[i], phi[i]), 

i=0,...,len(theta)-1``. 

 

Parameters 

---------- 

theta, phi : array_like 

Input coordinates. Standard Numpy broadcasting is obeyed. 

dtheta : int, optional 

Order of theta-derivative 

 

.. versionadded:: 0.14.0 

dphi : int, optional 

Order of phi-derivative 

 

.. versionadded:: 0.14.0 

""" 

return self.__call__(theta, phi, dtheta=dtheta, dphi=dphi, grid=False) 

 

 

class SmoothSphereBivariateSpline(SphereBivariateSpline): 

""" 

Smooth bivariate spline approximation in spherical coordinates. 

 

.. versionadded:: 0.11.0 

 

Parameters 

---------- 

theta, phi, r : array_like 

1-D sequences of data points (order is not important). Coordinates 

must be given in radians. Theta must lie within the interval (0, pi), 

and phi must lie within the interval (0, 2pi). 

w : array_like, optional 

Positive 1-D sequence of weights. 

s : float, optional 

Positive smoothing factor defined for estimation condition: 

``sum((w(i)*(r(i) - s(theta(i), phi(i))))**2, axis=0) <= s`` 

Default ``s=len(w)`` which should be a good value if 1/w[i] is an 

estimate of the standard deviation of r[i]. 

eps : float, optional 

A threshold for determining the effective rank of an over-determined 

linear system of equations. `eps` should have a value between 0 and 1, 

the default is 1e-16. 

 

Notes 

----- 

For more information, see the FITPACK_ site about this function. 

 

.. _FITPACK: http://www.netlib.org/dierckx/sphere.f 

 

Examples 

-------- 

Suppose we have global data on a coarse grid (the input data does not 

have to be on a grid): 

 

>>> theta = np.linspace(0., np.pi, 7) 

>>> phi = np.linspace(0., 2*np.pi, 9) 

>>> data = np.empty((theta.shape[0], phi.shape[0])) 

>>> data[:,0], data[0,:], data[-1,:] = 0., 0., 0. 

>>> data[1:-1,1], data[1:-1,-1] = 1., 1. 

>>> data[1,1:-1], data[-2,1:-1] = 1., 1. 

>>> data[2:-2,2], data[2:-2,-2] = 2., 2. 

>>> data[2,2:-2], data[-3,2:-2] = 2., 2. 

>>> data[3,3:-2] = 3. 

>>> data = np.roll(data, 4, 1) 

 

We need to set up the interpolator object 

 

>>> lats, lons = np.meshgrid(theta, phi) 

>>> from scipy.interpolate import SmoothSphereBivariateSpline 

>>> lut = SmoothSphereBivariateSpline(lats.ravel(), lons.ravel(), 

... data.T.ravel(), s=3.5) 

 

As a first test, we'll see what the algorithm returns when run on the 

input coordinates 

 

>>> data_orig = lut(theta, phi) 

 

Finally we interpolate the data to a finer grid 

 

>>> fine_lats = np.linspace(0., np.pi, 70) 

>>> fine_lons = np.linspace(0., 2 * np.pi, 90) 

 

>>> data_smth = lut(fine_lats, fine_lons) 

 

>>> import matplotlib.pyplot as plt 

>>> fig = plt.figure() 

>>> ax1 = fig.add_subplot(131) 

>>> ax1.imshow(data, interpolation='nearest') 

>>> ax2 = fig.add_subplot(132) 

>>> ax2.imshow(data_orig, interpolation='nearest') 

>>> ax3 = fig.add_subplot(133) 

>>> ax3.imshow(data_smth, interpolation='nearest') 

>>> plt.show() 

 

""" 

 

def __init__(self, theta, phi, r, w=None, s=0., eps=1E-16): 

if np.issubclass_(w, float): 

w = ones(len(theta)) * w 

nt_, tt_, np_, tp_, c, fp, ier = dfitpack.spherfit_smth(theta, phi, 

r, w=w, s=s, 

eps=eps) 

if ier not in [0, -1, -2]: 

message = _spherefit_messages.get(ier, 'ier=%s' % (ier)) 

raise ValueError(message) 

 

self.fp = fp 

self.tck = tt_[:nt_], tp_[:np_], c[:(nt_ - 4) * (np_ - 4)] 

self.degrees = (3, 3) 

 

 

class LSQSphereBivariateSpline(SphereBivariateSpline): 

""" 

Weighted least-squares bivariate spline approximation in spherical 

coordinates. 

 

.. versionadded:: 0.11.0 

 

Parameters 

---------- 

theta, phi, r : array_like 

1-D sequences of data points (order is not important). Coordinates 

must be given in radians. Theta must lie within the interval (0, pi), 

and phi must lie within the interval (0, 2pi). 

tt, tp : array_like 

Strictly ordered 1-D sequences of knots coordinates. 

Coordinates must satisfy ``0 < tt[i] < pi``, ``0 < tp[i] < 2*pi``. 

w : array_like, optional 

Positive 1-D sequence of weights, of the same length as `theta`, `phi` 

and `r`. 

eps : float, optional 

A threshold for determining the effective rank of an over-determined 

linear system of equations. `eps` should have a value between 0 and 1, 

the default is 1e-16. 

 

Notes 

----- 

For more information, see the FITPACK_ site about this function. 

 

.. _FITPACK: http://www.netlib.org/dierckx/sphere.f 

 

Examples 

-------- 

Suppose we have global data on a coarse grid (the input data does not 

have to be on a grid): 

 

>>> theta = np.linspace(0., np.pi, 7) 

>>> phi = np.linspace(0., 2*np.pi, 9) 

>>> data = np.empty((theta.shape[0], phi.shape[0])) 

>>> data[:,0], data[0,:], data[-1,:] = 0., 0., 0. 

>>> data[1:-1,1], data[1:-1,-1] = 1., 1. 

>>> data[1,1:-1], data[-2,1:-1] = 1., 1. 

>>> data[2:-2,2], data[2:-2,-2] = 2., 2. 

>>> data[2,2:-2], data[-3,2:-2] = 2., 2. 

>>> data[3,3:-2] = 3. 

>>> data = np.roll(data, 4, 1) 

 

We need to set up the interpolator object. Here, we must also specify the 

coordinates of the knots to use. 

 

>>> lats, lons = np.meshgrid(theta, phi) 

>>> knotst, knotsp = theta.copy(), phi.copy() 

>>> knotst[0] += .0001 

>>> knotst[-1] -= .0001 

>>> knotsp[0] += .0001 

>>> knotsp[-1] -= .0001 

>>> from scipy.interpolate import LSQSphereBivariateSpline 

>>> lut = LSQSphereBivariateSpline(lats.ravel(), lons.ravel(), 

... data.T.ravel(), knotst, knotsp) 

 

As a first test, we'll see what the algorithm returns when run on the 

input coordinates 

 

>>> data_orig = lut(theta, phi) 

 

Finally we interpolate the data to a finer grid 

 

>>> fine_lats = np.linspace(0., np.pi, 70) 

>>> fine_lons = np.linspace(0., 2*np.pi, 90) 

 

>>> data_lsq = lut(fine_lats, fine_lons) 

 

>>> import matplotlib.pyplot as plt 

>>> fig = plt.figure() 

>>> ax1 = fig.add_subplot(131) 

>>> ax1.imshow(data, interpolation='nearest') 

>>> ax2 = fig.add_subplot(132) 

>>> ax2.imshow(data_orig, interpolation='nearest') 

>>> ax3 = fig.add_subplot(133) 

>>> ax3.imshow(data_lsq, interpolation='nearest') 

>>> plt.show() 

 

""" 

 

def __init__(self, theta, phi, r, tt, tp, w=None, eps=1E-16): 

if np.issubclass_(w, float): 

w = ones(len(theta)) * w 

nt_, np_ = 8 + len(tt), 8 + len(tp) 

tt_, tp_ = zeros((nt_,), float), zeros((np_,), float) 

tt_[4:-4], tp_[4:-4] = tt, tp 

tt_[-4:], tp_[-4:] = np.pi, 2. * np.pi 

tt_, tp_, c, fp, ier = dfitpack.spherfit_lsq(theta, phi, r, tt_, tp_, 

w=w, eps=eps) 

if ier < -2: 

deficiency = 6 + (nt_ - 8) * (np_ - 7) + ier 

message = _spherefit_messages.get(-3) % (deficiency, -ier) 

warnings.warn(message) 

elif ier not in [0, -1, -2]: 

message = _spherefit_messages.get(ier, 'ier=%s' % (ier)) 

raise ValueError(message) 

 

self.fp = fp 

self.tck = tt_, tp_, c 

self.degrees = (3, 3) 

 

 

_spfit_messages = _surfit_messages.copy() 

_spfit_messages[10] = """ 

ERROR: on entry, the input data are controlled on validity 

the following restrictions must be satisfied. 

-1<=iopt(1)<=1, 0<=iopt(2)<=1, 0<=iopt(3)<=1, 

-1<=ider(1)<=1, 0<=ider(2)<=1, ider(2)=0 if iopt(2)=0. 

-1<=ider(3)<=1, 0<=ider(4)<=1, ider(4)=0 if iopt(3)=0. 

mu >= mumin (see above), mv >= 4, nuest >=8, nvest >= 8, 

kwrk>=5+mu+mv+nuest+nvest, 

lwrk >= 12+nuest*(mv+nvest+3)+nvest*24+4*mu+8*mv+max(nuest,mv+nvest) 

0< u(i-1)<u(i)< pi,i=2,..,mu, 

-pi<=v(1)< pi, v(1)<v(i-1)<v(i)<v(1)+2*pi, i=3,...,mv 

if iopt(1)=-1: 8<=nu<=min(nuest,mu+6+iopt(2)+iopt(3)) 

0<tu(5)<tu(6)<...<tu(nu-4)< pi 

8<=nv<=min(nvest,mv+7) 

v(1)<tv(5)<tv(6)<...<tv(nv-4)<v(1)+2*pi 

the schoenberg-whitney conditions, i.e. there must be 

subset of grid co-ordinates uu(p) and vv(q) such that 

tu(p) < uu(p) < tu(p+4) ,p=1,...,nu-4 

(iopt(2)=1 and iopt(3)=1 also count for a uu-value 

tv(q) < vv(q) < tv(q+4) ,q=1,...,nv-4 

(vv(q) is either a value v(j) or v(j)+2*pi) 

if iopt(1)>=0: s>=0 

if s=0: nuest>=mu+6+iopt(2)+iopt(3), nvest>=mv+7 

if one of these conditions is found to be violated,control is 

immediately repassed to the calling program. in that case there is no 

approximation returned.""" 

 

 

class RectSphereBivariateSpline(SphereBivariateSpline): 

""" 

Bivariate spline approximation over a rectangular mesh on a sphere. 

 

Can be used for smoothing data. 

 

.. versionadded:: 0.11.0 

 

Parameters 

---------- 

u : array_like 

1-D array of latitude coordinates in strictly ascending order. 

Coordinates must be given in radians and lie within the interval 

(0, pi). 

v : array_like 

1-D array of longitude coordinates in strictly ascending order. 

Coordinates must be given in radians. First element (v[0]) must lie 

within the interval [-pi, pi). Last element (v[-1]) must satisfy 

v[-1] <= v[0] + 2*pi. 

r : array_like 

2-D array of data with shape ``(u.size, v.size)``. 

s : float, optional 

Positive smoothing factor defined for estimation condition 

(``s=0`` is for interpolation). 

pole_continuity : bool or (bool, bool), optional 

Order of continuity at the poles ``u=0`` (``pole_continuity[0]``) and 

``u=pi`` (``pole_continuity[1]``). The order of continuity at the pole 

will be 1 or 0 when this is True or False, respectively. 

Defaults to False. 

pole_values : float or (float, float), optional 

Data values at the poles ``u=0`` and ``u=pi``. Either the whole 

parameter or each individual element can be None. Defaults to None. 

pole_exact : bool or (bool, bool), optional 

Data value exactness at the poles ``u=0`` and ``u=pi``. If True, the 

value is considered to be the right function value, and it will be 

fitted exactly. If False, the value will be considered to be a data 

value just like the other data values. Defaults to False. 

pole_flat : bool or (bool, bool), optional 

For the poles at ``u=0`` and ``u=pi``, specify whether or not the 

approximation has vanishing derivatives. Defaults to False. 

 

See Also 

-------- 

RectBivariateSpline : bivariate spline approximation over a rectangular 

mesh 

 

Notes 

----- 

Currently, only the smoothing spline approximation (``iopt[0] = 0`` and 

``iopt[0] = 1`` in the FITPACK routine) is supported. The exact 

least-squares spline approximation is not implemented yet. 

 

When actually performing the interpolation, the requested `v` values must 

lie within the same length 2pi interval that the original `v` values were 

chosen from. 

 

For more information, see the FITPACK_ site about this function. 

 

.. _FITPACK: http://www.netlib.org/dierckx/spgrid.f 

 

Examples 

-------- 

Suppose we have global data on a coarse grid 

 

>>> lats = np.linspace(10, 170, 9) * np.pi / 180. 

>>> lons = np.linspace(0, 350, 18) * np.pi / 180. 

>>> data = np.dot(np.atleast_2d(90. - np.linspace(-80., 80., 18)).T, 

... np.atleast_2d(180. - np.abs(np.linspace(0., 350., 9)))).T 

 

We want to interpolate it to a global one-degree grid 

 

>>> new_lats = np.linspace(1, 180, 180) * np.pi / 180 

>>> new_lons = np.linspace(1, 360, 360) * np.pi / 180 

>>> new_lats, new_lons = np.meshgrid(new_lats, new_lons) 

 

We need to set up the interpolator object 

 

>>> from scipy.interpolate import RectSphereBivariateSpline 

>>> lut = RectSphereBivariateSpline(lats, lons, data) 

 

Finally we interpolate the data. The `RectSphereBivariateSpline` object 

only takes 1-D arrays as input, therefore we need to do some reshaping. 

 

>>> data_interp = lut.ev(new_lats.ravel(), 

... new_lons.ravel()).reshape((360, 180)).T 

 

Looking at the original and the interpolated data, one can see that the 

interpolant reproduces the original data very well: 

 

>>> import matplotlib.pyplot as plt 

>>> fig = plt.figure() 

>>> ax1 = fig.add_subplot(211) 

>>> ax1.imshow(data, interpolation='nearest') 

>>> ax2 = fig.add_subplot(212) 

>>> ax2.imshow(data_interp, interpolation='nearest') 

>>> plt.show() 

 

Choosing the optimal value of ``s`` can be a delicate task. Recommended 

values for ``s`` depend on the accuracy of the data values. If the user 

has an idea of the statistical errors on the data, she can also find a 

proper estimate for ``s``. By assuming that, if she specifies the 

right ``s``, the interpolator will use a spline ``f(u,v)`` which exactly 

reproduces the function underlying the data, she can evaluate 

``sum((r(i,j)-s(u(i),v(j)))**2)`` to find a good estimate for this ``s``. 

For example, if she knows that the statistical errors on her 

``r(i,j)``-values are not greater than 0.1, she may expect that a good 

``s`` should have a value not larger than ``u.size * v.size * (0.1)**2``. 

 

If nothing is known about the statistical error in ``r(i,j)``, ``s`` must 

be determined by trial and error. The best is then to start with a very 

large value of ``s`` (to determine the least-squares polynomial and the 

corresponding upper bound ``fp0`` for ``s``) and then to progressively 

decrease the value of ``s`` (say by a factor 10 in the beginning, i.e. 

``s = fp0 / 10, fp0 / 100, ...`` and more carefully as the approximation 

shows more detail) to obtain closer fits. 

 

The interpolation results for different values of ``s`` give some insight 

into this process: 

 

>>> fig2 = plt.figure() 

>>> s = [3e9, 2e9, 1e9, 1e8] 

>>> for ii in range(len(s)): 

... lut = RectSphereBivariateSpline(lats, lons, data, s=s[ii]) 

... data_interp = lut.ev(new_lats.ravel(), 

... new_lons.ravel()).reshape((360, 180)).T 

... ax = fig2.add_subplot(2, 2, ii+1) 

... ax.imshow(data_interp, interpolation='nearest') 

... ax.set_title("s = %g" % s[ii]) 

>>> plt.show() 

 

""" 

 

def __init__(self, u, v, r, s=0., pole_continuity=False, pole_values=None, 

pole_exact=False, pole_flat=False): 

iopt = np.array([0, 0, 0], dtype=int) 

ider = np.array([-1, 0, -1, 0], dtype=int) 

if pole_values is None: 

pole_values = (None, None) 

elif isinstance(pole_values, (float, np.float32, np.float64)): 

pole_values = (pole_values, pole_values) 

if isinstance(pole_continuity, bool): 

pole_continuity = (pole_continuity, pole_continuity) 

if isinstance(pole_exact, bool): 

pole_exact = (pole_exact, pole_exact) 

if isinstance(pole_flat, bool): 

pole_flat = (pole_flat, pole_flat) 

 

r0, r1 = pole_values 

iopt[1:] = pole_continuity 

if r0 is None: 

ider[0] = -1 

else: 

ider[0] = pole_exact[0] 

 

if r1 is None: 

ider[2] = -1 

else: 

ider[2] = pole_exact[1] 

 

ider[1], ider[3] = pole_flat 

 

u, v = np.ravel(u), np.ravel(v) 

if not np.all(np.diff(u) > 0.0): 

raise ValueError('u must be strictly increasing') 

if not np.all(np.diff(v) > 0.0): 

raise ValueError('v must be strictly increasing') 

 

if not u.size == r.shape[0]: 

raise ValueError('u dimension of r must have same number of ' 

'elements as u') 

if not v.size == r.shape[1]: 

raise ValueError('v dimension of r must have same number of ' 

'elements as v') 

 

if pole_continuity[1] is False and pole_flat[1] is True: 

raise ValueError('if pole_continuity is False, so must be ' 

'pole_flat') 

if pole_continuity[0] is False and pole_flat[0] is True: 

raise ValueError('if pole_continuity is False, so must be ' 

'pole_flat') 

 

r = np.ravel(r) 

nu, tu, nv, tv, c, fp, ier = dfitpack.regrid_smth_spher(iopt, ider, 

u.copy(), v.copy(), r.copy(), r0, r1, s) 

 

if ier not in [0, -1, -2]: 

msg = _spfit_messages.get(ier, 'ier=%s' % (ier)) 

raise ValueError(msg) 

 

self.fp = fp 

self.tck = tu[:nu], tv[:nv], c[:(nu - 4) * (nv-4)] 

self.degrees = (3, 3)