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

# Author: Travis Oliphant 2001 

# Author: Nathan Woods 2013 (nquad &c) 

from __future__ import division, print_function, absolute_import 

 

import sys 

import warnings 

from functools import partial 

 

from . import _quadpack 

import numpy 

from numpy import Inf 

 

__all__ = ['quad', 'dblquad', 'tplquad', 'nquad', 'quad_explain', 

'IntegrationWarning'] 

 

 

error = _quadpack.error 

 

class IntegrationWarning(UserWarning): 

""" 

Warning on issues during integration. 

""" 

pass 

 

 

def quad_explain(output=sys.stdout): 

""" 

Print extra information about integrate.quad() parameters and returns. 

 

Parameters 

---------- 

output : instance with "write" method, optional 

Information about `quad` is passed to ``output.write()``. 

Default is ``sys.stdout``. 

 

Returns 

------- 

None 

 

""" 

output.write(quad.__doc__) 

 

 

def quad(func, a, b, args=(), full_output=0, epsabs=1.49e-8, epsrel=1.49e-8, 

limit=50, points=None, weight=None, wvar=None, wopts=None, maxp1=50, 

limlst=50): 

""" 

Compute a definite integral. 

 

Integrate func from `a` to `b` (possibly infinite interval) using a 

technique from the Fortran library QUADPACK. 

 

Parameters 

---------- 

func : {function, scipy.LowLevelCallable} 

A Python function or method to integrate. If `func` takes many 

arguments, it is integrated along the axis corresponding to the 

first argument. 

 

If the user desires improved integration performance, then `f` may 

be a `scipy.LowLevelCallable` with one of the signatures:: 

 

double func(double x) 

double func(double x, void *user_data) 

double func(int n, double *xx) 

double func(int n, double *xx, void *user_data) 

 

The ``user_data`` is the data contained in the `scipy.LowLevelCallable`. 

In the call forms with ``xx``, ``n`` is the length of the ``xx`` 

array which contains ``xx[0] == x`` and the rest of the items are 

numbers contained in the ``args`` argument of quad. 

 

In addition, certain ctypes call signatures are supported for 

backward compatibility, but those should not be used in new code. 

a : float 

Lower limit of integration (use -numpy.inf for -infinity). 

b : float 

Upper limit of integration (use numpy.inf for +infinity). 

args : tuple, optional 

Extra arguments to pass to `func`. 

full_output : int, optional 

Non-zero to return a dictionary of integration information. 

If non-zero, warning messages are also suppressed and the 

message is appended to the output tuple. 

 

Returns 

------- 

y : float 

The integral of func from `a` to `b`. 

abserr : float 

An estimate of the absolute error in the result. 

infodict : dict 

A dictionary containing additional information. 

Run scipy.integrate.quad_explain() for more information. 

message 

A convergence message. 

explain 

Appended only with 'cos' or 'sin' weighting and infinite 

integration limits, it contains an explanation of the codes in 

infodict['ierlst'] 

 

Other Parameters 

---------------- 

epsabs : float or int, optional 

Absolute error tolerance. 

epsrel : float or int, optional 

Relative error tolerance. 

limit : float or int, optional 

An upper bound on the number of subintervals used in the adaptive 

algorithm. 

points : (sequence of floats,ints), optional 

A sequence of break points in the bounded integration interval 

where local difficulties of the integrand may occur (e.g., 

singularities, discontinuities). The sequence does not have 

to be sorted. 

weight : float or int, optional 

String indicating weighting function. Full explanation for this 

and the remaining arguments can be found below. 

wvar : optional 

Variables for use with weighting functions. 

wopts : optional 

Optional input for reusing Chebyshev moments. 

maxp1 : float or int, optional 

An upper bound on the number of Chebyshev moments. 

limlst : int, optional 

Upper bound on the number of cycles (>=3) for use with a sinusoidal 

weighting and an infinite end-point. 

 

See Also 

-------- 

dblquad : double integral 

tplquad : triple integral 

nquad : n-dimensional integrals (uses `quad` recursively) 

fixed_quad : fixed-order Gaussian quadrature 

quadrature : adaptive Gaussian quadrature 

odeint : ODE integrator 

ode : ODE integrator 

simps : integrator for sampled data 

romb : integrator for sampled data 

scipy.special : for coefficients and roots of orthogonal polynomials 

 

Notes 

----- 

 

**Extra information for quad() inputs and outputs** 

 

If full_output is non-zero, then the third output argument 

(infodict) is a dictionary with entries as tabulated below. For 

infinite limits, the range is transformed to (0,1) and the 

optional outputs are given with respect to this transformed range. 

Let M be the input argument limit and let K be infodict['last']. 

The entries are: 

 

'neval' 

The number of function evaluations. 

'last' 

The number, K, of subintervals produced in the subdivision process. 

'alist' 

A rank-1 array of length M, the first K elements of which are the 

left end points of the subintervals in the partition of the 

integration range. 

'blist' 

A rank-1 array of length M, the first K elements of which are the 

right end points of the subintervals. 

'rlist' 

A rank-1 array of length M, the first K elements of which are the 

integral approximations on the subintervals. 

'elist' 

A rank-1 array of length M, the first K elements of which are the 

moduli of the absolute error estimates on the subintervals. 

'iord' 

A rank-1 integer array of length M, the first L elements of 

which are pointers to the error estimates over the subintervals 

with ``L=K`` if ``K<=M/2+2`` or ``L=M+1-K`` otherwise. Let I be the 

sequence ``infodict['iord']`` and let E be the sequence 

``infodict['elist']``. Then ``E[I[1]], ..., E[I[L]]`` forms a 

decreasing sequence. 

 

If the input argument points is provided (i.e. it is not None), 

the following additional outputs are placed in the output 

dictionary. Assume the points sequence is of length P. 

 

'pts' 

A rank-1 array of length P+2 containing the integration limits 

and the break points of the intervals in ascending order. 

This is an array giving the subintervals over which integration 

will occur. 

'level' 

A rank-1 integer array of length M (=limit), containing the 

subdivision levels of the subintervals, i.e., if (aa,bb) is a 

subinterval of ``(pts[1], pts[2])`` where ``pts[0]`` and ``pts[2]`` 

are adjacent elements of ``infodict['pts']``, then (aa,bb) has level l 

if ``|bb-aa| = |pts[2]-pts[1]| * 2**(-l)``. 

'ndin' 

A rank-1 integer array of length P+2. After the first integration 

over the intervals (pts[1], pts[2]), the error estimates over some 

of the intervals may have been increased artificially in order to 

put their subdivision forward. This array has ones in slots 

corresponding to the subintervals for which this happens. 

 

**Weighting the integrand** 

 

The input variables, *weight* and *wvar*, are used to weight the 

integrand by a select list of functions. Different integration 

methods are used to compute the integral with these weighting 

functions. The possible values of weight and the corresponding 

weighting functions are. 

 

========== =================================== ===================== 

``weight`` Weight function used ``wvar`` 

========== =================================== ===================== 

'cos' cos(w*x) wvar = w 

'sin' sin(w*x) wvar = w 

'alg' g(x) = ((x-a)**alpha)*((b-x)**beta) wvar = (alpha, beta) 

'alg-loga' g(x)*log(x-a) wvar = (alpha, beta) 

'alg-logb' g(x)*log(b-x) wvar = (alpha, beta) 

'alg-log' g(x)*log(x-a)*log(b-x) wvar = (alpha, beta) 

'cauchy' 1/(x-c) wvar = c 

========== =================================== ===================== 

 

wvar holds the parameter w, (alpha, beta), or c depending on the weight 

selected. In these expressions, a and b are the integration limits. 

 

For the 'cos' and 'sin' weighting, additional inputs and outputs are 

available. 

 

For finite integration limits, the integration is performed using a 

Clenshaw-Curtis method which uses Chebyshev moments. For repeated 

calculations, these moments are saved in the output dictionary: 

 

'momcom' 

The maximum level of Chebyshev moments that have been computed, 

i.e., if ``M_c`` is ``infodict['momcom']`` then the moments have been 

computed for intervals of length ``|b-a| * 2**(-l)``, 

``l=0,1,...,M_c``. 

'nnlog' 

A rank-1 integer array of length M(=limit), containing the 

subdivision levels of the subintervals, i.e., an element of this 

array is equal to l if the corresponding subinterval is 

``|b-a|* 2**(-l)``. 

'chebmo' 

A rank-2 array of shape (25, maxp1) containing the computed 

Chebyshev moments. These can be passed on to an integration 

over the same interval by passing this array as the second 

element of the sequence wopts and passing infodict['momcom'] as 

the first element. 

 

If one of the integration limits is infinite, then a Fourier integral is 

computed (assuming w neq 0). If full_output is 1 and a numerical error 

is encountered, besides the error message attached to the output tuple, 

a dictionary is also appended to the output tuple which translates the 

error codes in the array ``info['ierlst']`` to English messages. The 

output information dictionary contains the following entries instead of 

'last', 'alist', 'blist', 'rlist', and 'elist': 

 

'lst' 

The number of subintervals needed for the integration (call it ``K_f``). 

'rslst' 

A rank-1 array of length M_f=limlst, whose first ``K_f`` elements 

contain the integral contribution over the interval 

``(a+(k-1)c, a+kc)`` where ``c = (2*floor(|w|) + 1) * pi / |w|`` 

and ``k=1,2,...,K_f``. 

'erlst' 

A rank-1 array of length ``M_f`` containing the error estimate 

corresponding to the interval in the same position in 

``infodict['rslist']``. 

'ierlst' 

A rank-1 integer array of length ``M_f`` containing an error flag 

corresponding to the interval in the same position in 

``infodict['rslist']``. See the explanation dictionary (last entry 

in the output tuple) for the meaning of the codes. 

 

Examples 

-------- 

Calculate :math:`\\int^4_0 x^2 dx` and compare with an analytic result 

 

>>> from scipy import integrate 

>>> x2 = lambda x: x**2 

>>> integrate.quad(x2, 0, 4) 

(21.333333333333332, 2.3684757858670003e-13) 

>>> print(4**3 / 3.) # analytical result 

21.3333333333 

 

Calculate :math:`\\int^\\infty_0 e^{-x} dx` 

 

>>> invexp = lambda x: np.exp(-x) 

>>> integrate.quad(invexp, 0, np.inf) 

(1.0, 5.842605999138044e-11) 

 

>>> f = lambda x,a : a*x 

>>> y, err = integrate.quad(f, 0, 1, args=(1,)) 

>>> y 

0.5 

>>> y, err = integrate.quad(f, 0, 1, args=(3,)) 

>>> y 

1.5 

 

Calculate :math:`\\int^1_0 x^2 + y^2 dx` with ctypes, holding 

y parameter as 1:: 

 

testlib.c => 

double func(int n, double args[n]){ 

return args[0]*args[0] + args[1]*args[1];} 

compile to library testlib.* 

 

:: 

 

from scipy import integrate 

import ctypes 

lib = ctypes.CDLL('/home/.../testlib.*') #use absolute path 

lib.func.restype = ctypes.c_double 

lib.func.argtypes = (ctypes.c_int,ctypes.c_double) 

integrate.quad(lib.func,0,1,(1)) 

#(1.3333333333333333, 1.4802973661668752e-14) 

print((1.0**3/3.0 + 1.0) - (0.0**3/3.0 + 0.0)) #Analytic result 

# 1.3333333333333333 

 

Be aware that pulse shapes and other sharp features as compared to the 

size of the integration interval may not be integrated correctly using 

this method. A simplified example of this limitation is integrating a 

y-axis reflected step function with many zero values within the integrals 

bounds. 

 

>>> y = lambda x: 1 if x<=0 else 0 

>>> integrate.quad(y, -1, 1) 

(1.0, 1.1102230246251565e-14) 

>>> integrate.quad(y, -1, 100) 

(1.0000000002199108, 1.0189464580163188e-08) 

>>> integrate.quad(y, -1, 10000) 

(0.0, 0.0) 

 

""" 

if not isinstance(args, tuple): 

args = (args,) 

 

# check the limits of integration: \int_a^b, expect a < b 

flip, a, b = b < a, min(a, b), max(a, b) 

 

if weight is None: 

retval = _quad(func, a, b, args, full_output, epsabs, epsrel, limit, 

points) 

else: 

retval = _quad_weight(func, a, b, args, full_output, epsabs, epsrel, 

limlst, limit, maxp1, weight, wvar, wopts) 

 

if flip: 

retval = (-retval[0],) + retval[1:] 

 

ier = retval[-1] 

if ier == 0: 

return retval[:-1] 

 

msgs = {80: "A Python error occurred possibly while calling the function.", 

1: "The maximum number of subdivisions (%d) has been achieved.\n If increasing the limit yields no improvement it is advised to analyze \n the integrand in order to determine the difficulties. If the position of a \n local difficulty can be determined (singularity, discontinuity) one will \n probably gain from splitting up the interval and calling the integrator \n on the subranges. Perhaps a special-purpose integrator should be used." % limit, 

2: "The occurrence of roundoff error is detected, which prevents \n the requested tolerance from being achieved. The error may be \n underestimated.", 

3: "Extremely bad integrand behavior occurs at some points of the\n integration interval.", 

4: "The algorithm does not converge. Roundoff error is detected\n in the extrapolation table. It is assumed that the requested tolerance\n cannot be achieved, and that the returned result (if full_output = 1) is \n the best which can be obtained.", 

5: "The integral is probably divergent, or slowly convergent.", 

6: "The input is invalid.", 

7: "Abnormal termination of the routine. The estimates for result\n and error are less reliable. It is assumed that the requested accuracy\n has not been achieved.", 

'unknown': "Unknown error."} 

 

if weight in ['cos','sin'] and (b == Inf or a == -Inf): 

msgs[1] = "The maximum number of cycles allowed has been achieved., e.e.\n of subintervals (a+(k-1)c, a+kc) where c = (2*int(abs(omega)+1))\n *pi/abs(omega), for k = 1, 2, ..., lst. One can allow more cycles by increasing the value of limlst. Look at info['ierlst'] with full_output=1." 

msgs[4] = "The extrapolation table constructed for convergence acceleration\n of the series formed by the integral contributions over the cycles, \n does not converge to within the requested accuracy. Look at \n info['ierlst'] with full_output=1." 

msgs[7] = "Bad integrand behavior occurs within one or more of the cycles.\n Location and type of the difficulty involved can be determined from \n the vector info['ierlist'] obtained with full_output=1." 

explain = {1: "The maximum number of subdivisions (= limit) has been \n achieved on this cycle.", 

2: "The occurrence of roundoff error is detected and prevents\n the tolerance imposed on this cycle from being achieved.", 

3: "Extremely bad integrand behavior occurs at some points of\n this cycle.", 

4: "The integral over this cycle does not converge (to within the required accuracy) due to roundoff in the extrapolation procedure invoked on this cycle. It is assumed that the result on this interval is the best which can be obtained.", 

5: "The integral over this cycle is probably divergent or slowly convergent."} 

 

try: 

msg = msgs[ier] 

except KeyError: 

msg = msgs['unknown'] 

 

if ier in [1,2,3,4,5,7]: 

if full_output: 

if weight in ['cos', 'sin'] and (b == Inf or a == Inf): 

return retval[:-1] + (msg, explain) 

else: 

return retval[:-1] + (msg,) 

else: 

warnings.warn(msg, IntegrationWarning) 

return retval[:-1] 

 

elif ier == 6: # Forensic decision tree when QUADPACK throws ier=6 

if epsabs <= 0: # Small error tolerance - applies to all methods 

if epsrel < max(50 * sys.float_info.epsilon, 5e-29): 

msg = ("If 'errabs'<=0, 'epsrel' must be greater than both" 

" 5e-29 and 50*(machine epsilon).") 

elif weight in ['sin', 'cos'] and (abs(a) + abs(b) == Inf): 

msg = ("Sine or cosine weighted intergals with infinite domain" 

" must have 'epsabs'>0.") 

 

elif weight is None: 

if points is None: # QAGSE/QAGIE 

msg = ("Invalid 'limit' argument. There must be" 

" at least one subinterval") 

else: # QAGPE 

if not (min(a, b) <= min(points) <= max(points) <= max(a, b)): 

msg = ("All break points in 'points' must lie within the" 

" integration limits.") 

elif len(points) >= limit: 

msg = ("Number of break points ({:d})" 

" must be less than subinterval" 

" limit ({:d})").format(len(points), limit) 

 

else: 

if maxp1 < 1: 

msg = "Chebyshev moment limit maxp1 must be >=1." 

 

elif weight in ('cos', 'sin') and abs(a+b) == Inf: # QAWFE 

msg = "Cycle limit limlst must be >=3." 

 

elif weight.startswith('alg'): # QAWSE 

if min(wvar) < -1: 

msg = "wvar parameters (alpha, beta) must both be >= -1." 

if b < a: 

msg = "Integration limits a, b must satistfy a<b." 

 

elif weight == 'cauchy' and wvar in (a, b): 

msg = ("Parameter 'wvar' must not equal" 

" integration limits 'a' or 'b'.") 

 

raise ValueError(msg) 

 

 

def _quad(func,a,b,args,full_output,epsabs,epsrel,limit,points): 

infbounds = 0 

if (b != Inf and a != -Inf): 

pass # standard integration 

elif (b == Inf and a != -Inf): 

infbounds = 1 

bound = a 

elif (b == Inf and a == -Inf): 

infbounds = 2 

bound = 0 # ignored 

elif (b != Inf and a == -Inf): 

infbounds = -1 

bound = b 

else: 

raise RuntimeError("Infinity comparisons don't work for you.") 

 

if points is None: 

if infbounds == 0: 

return _quadpack._qagse(func,a,b,args,full_output,epsabs,epsrel,limit) 

else: 

return _quadpack._qagie(func,bound,infbounds,args,full_output,epsabs,epsrel,limit) 

else: 

if infbounds != 0: 

raise ValueError("Infinity inputs cannot be used with break points.") 

else: 

#Duplicates force function evaluation at sinular points 

the_points = numpy.unique(points) 

the_points = the_points[a < the_points] 

the_points = the_points[the_points < b] 

the_points = numpy.concatenate((the_points, (0., 0.))) 

return _quadpack._qagpe(func,a,b,the_points,args,full_output,epsabs,epsrel,limit) 

 

 

def _quad_weight(func,a,b,args,full_output,epsabs,epsrel,limlst,limit,maxp1,weight,wvar,wopts): 

if weight not in ['cos','sin','alg','alg-loga','alg-logb','alg-log','cauchy']: 

raise ValueError("%s not a recognized weighting function." % weight) 

 

strdict = {'cos':1,'sin':2,'alg':1,'alg-loga':2,'alg-logb':3,'alg-log':4} 

 

if weight in ['cos','sin']: 

integr = strdict[weight] 

if (b != Inf and a != -Inf): # finite limits 

if wopts is None: # no precomputed chebyshev moments 

return _quadpack._qawoe(func, a, b, wvar, integr, args, full_output, 

epsabs, epsrel, limit, maxp1,1) 

else: # precomputed chebyshev moments 

momcom = wopts[0] 

chebcom = wopts[1] 

return _quadpack._qawoe(func, a, b, wvar, integr, args, full_output, 

epsabs, epsrel, limit, maxp1, 2, momcom, chebcom) 

 

elif (b == Inf and a != -Inf): 

return _quadpack._qawfe(func, a, wvar, integr, args, full_output, 

epsabs,limlst,limit,maxp1) 

elif (b != Inf and a == -Inf): # remap function and interval 

if weight == 'cos': 

def thefunc(x,*myargs): 

y = -x 

func = myargs[0] 

myargs = (y,) + myargs[1:] 

return func(*myargs) 

else: 

def thefunc(x,*myargs): 

y = -x 

func = myargs[0] 

myargs = (y,) + myargs[1:] 

return -func(*myargs) 

args = (func,) + args 

return _quadpack._qawfe(thefunc, -b, wvar, integr, args, 

full_output, epsabs, limlst, limit, maxp1) 

else: 

raise ValueError("Cannot integrate with this weight from -Inf to +Inf.") 

else: 

if a in [-Inf,Inf] or b in [-Inf,Inf]: 

raise ValueError("Cannot integrate with this weight over an infinite interval.") 

 

if weight.startswith('alg'): 

integr = strdict[weight] 

return _quadpack._qawse(func, a, b, wvar, integr, args, 

full_output, epsabs, epsrel, limit) 

else: # weight == 'cauchy' 

return _quadpack._qawce(func, a, b, wvar, args, full_output, 

epsabs, epsrel, limit) 

 

 

def dblquad(func, a, b, gfun, hfun, args=(), epsabs=1.49e-8, epsrel=1.49e-8): 

""" 

Compute a double integral. 

 

Return the double (definite) integral of ``func(y, x)`` from ``x = a..b`` 

and ``y = gfun(x)..hfun(x)``. 

 

Parameters 

---------- 

func : callable 

A Python function or method of at least two variables: y must be the 

first argument and x the second argument. 

a, b : float 

The limits of integration in x: `a` < `b` 

gfun : callable or float 

The lower boundary curve in y which is a function taking a single 

floating point argument (x) and returning a floating point result 

or a float indicating a constant boundary curve. 

hfun : callable or float 

The upper boundary curve in y (same requirements as `gfun`). 

args : sequence, optional 

Extra arguments to pass to `func`. 

epsabs : float, optional 

Absolute tolerance passed directly to the inner 1-D quadrature 

integration. Default is 1.49e-8. 

epsrel : float, optional 

Relative tolerance of the inner 1-D integrals. Default is 1.49e-8. 

 

Returns 

------- 

y : float 

The resultant integral. 

abserr : float 

An estimate of the error. 

 

See also 

-------- 

quad : single integral 

tplquad : triple integral 

nquad : N-dimensional integrals 

fixed_quad : fixed-order Gaussian quadrature 

quadrature : adaptive Gaussian quadrature 

odeint : ODE integrator 

ode : ODE integrator 

simps : integrator for sampled data 

romb : integrator for sampled data 

scipy.special : for coefficients and roots of orthogonal polynomials 

 

Examples 

-------- 

 

Compute the double integral of ``x * y**2`` over the box 

``x`` ranging from 0 to 2 and ``y`` ranging from 0 to 1. 

 

>>> from scipy import integrate 

>>> f = lambda y, x: x*y**2 

>>> integrate.dblquad(f, 0, 2, lambda x: 0, lambda x: 1) 

(0.6666666666666667, 7.401486830834377e-15) 

 

""" 

 

def temp_ranges(*args): 

return [gfun(args[0]) if callable(gfun) else gfun, 

hfun(args[0]) if callable(hfun) else hfun] 

 

return nquad(func, [temp_ranges, [a, b]], args=args, 

opts={"epsabs": epsabs, "epsrel": epsrel}) 

 

 

def tplquad(func, a, b, gfun, hfun, qfun, rfun, args=(), epsabs=1.49e-8, 

epsrel=1.49e-8): 

""" 

Compute a triple (definite) integral. 

 

Return the triple integral of ``func(z, y, x)`` from ``x = a..b``, 

``y = gfun(x)..hfun(x)``, and ``z = qfun(x,y)..rfun(x,y)``. 

 

Parameters 

---------- 

func : function 

A Python function or method of at least three variables in the 

order (z, y, x). 

a, b : float 

The limits of integration in x: `a` < `b` 

gfun : function or float 

The lower boundary curve in y which is a function taking a single 

floating point argument (x) and returning a floating point result 

or a float indicating a constant boundary curve. 

hfun : function or float 

The upper boundary curve in y (same requirements as `gfun`). 

qfun : function or float 

The lower boundary surface in z. It must be a function that takes 

two floats in the order (x, y) and returns a float or a float 

indicating a constant boundary surface. 

rfun : function or float 

The upper boundary surface in z. (Same requirements as `qfun`.) 

args : tuple, optional 

Extra arguments to pass to `func`. 

epsabs : float, optional 

Absolute tolerance passed directly to the innermost 1-D quadrature 

integration. Default is 1.49e-8. 

epsrel : float, optional 

Relative tolerance of the innermost 1-D integrals. Default is 1.49e-8. 

 

Returns 

------- 

y : float 

The resultant integral. 

abserr : float 

An estimate of the error. 

 

See Also 

-------- 

quad: Adaptive quadrature using QUADPACK 

quadrature: Adaptive Gaussian quadrature 

fixed_quad: Fixed-order Gaussian quadrature 

dblquad: Double integrals 

nquad : N-dimensional integrals 

romb: Integrators for sampled data 

simps: Integrators for sampled data 

ode: ODE integrators 

odeint: ODE integrators 

scipy.special: For coefficients and roots of orthogonal polynomials 

 

Examples 

-------- 

 

Compute the triple integral of ``x * y * z``, over ``x`` ranging  

from 1 to 2, ``y`` ranging from 2 to 3, ``z`` ranging from 0 to 1. 

 

>>> from scipy import integrate 

>>> f = lambda z, y, x: x*y*z 

>>> integrate.tplquad(f, 1, 2, lambda x: 2, lambda x: 3, 

... lambda x, y: 0, lambda x, y: 1) 

(1.8750000000000002, 3.324644794257407e-14) 

 

 

""" 

# f(z, y, x) 

# qfun/rfun (x, y) 

# gfun/hfun(x) 

# nquad will hand (y, x, t0, ...) to ranges0 

# nquad will hand (x, t0, ...) to ranges1 

# Stupid different API... 

 

def ranges0(*args): 

return [qfun(args[1], args[0]) if callable(qfun) else qfun, 

rfun(args[1], args[0]) if callable(rfun) else rfun] 

 

def ranges1(*args): 

return [gfun(args[0]) if callable(gfun) else gfun, 

hfun(args[0]) if callable(hfun) else hfun] 

 

ranges = [ranges0, ranges1, [a, b]] 

return nquad(func, ranges, args=args, 

opts={"epsabs": epsabs, "epsrel": epsrel}) 

 

 

def nquad(func, ranges, args=None, opts=None, full_output=False): 

""" 

Integration over multiple variables. 

 

Wraps `quad` to enable integration over multiple variables. 

Various options allow improved integration of discontinuous functions, as 

well as the use of weighted integration, and generally finer control of the 

integration process. 

 

Parameters 

---------- 

func : {callable, scipy.LowLevelCallable} 

The function to be integrated. Has arguments of ``x0, ... xn``, 

``t0, tm``, where integration is carried out over ``x0, ... xn``, which 

must be floats. Function signature should be 

``func(x0, x1, ..., xn, t0, t1, ..., tm)``. Integration is carried out 

in order. That is, integration over ``x0`` is the innermost integral, 

and ``xn`` is the outermost. 

 

If the user desires improved integration performance, then `f` may 

be a `scipy.LowLevelCallable` with one of the signatures:: 

 

double func(int n, double *xx) 

double func(int n, double *xx, void *user_data) 

 

where ``n`` is the number of extra parameters and args is an array 

of doubles of the additional parameters, the ``xx`` array contains the 

coordinates. The ``user_data`` is the data contained in the 

`scipy.LowLevelCallable`. 

ranges : iterable object 

Each element of ranges may be either a sequence of 2 numbers, or else 

a callable that returns such a sequence. ``ranges[0]`` corresponds to 

integration over x0, and so on. If an element of ranges is a callable, 

then it will be called with all of the integration arguments available, 

as well as any parametric arguments. e.g. if 

``func = f(x0, x1, x2, t0, t1)``, then ``ranges[0]`` may be defined as 

either ``(a, b)`` or else as ``(a, b) = range0(x1, x2, t0, t1)``. 

args : iterable object, optional 

Additional arguments ``t0, ..., tn``, required by `func`, `ranges`, and 

``opts``. 

opts : iterable object or dict, optional 

Options to be passed to `quad`. May be empty, a dict, or 

a sequence of dicts or functions that return a dict. If empty, the 

default options from scipy.integrate.quad are used. If a dict, the same 

options are used for all levels of integraion. If a sequence, then each 

element of the sequence corresponds to a particular integration. e.g. 

opts[0] corresponds to integration over x0, and so on. If a callable, 

the signature must be the same as for ``ranges``. The available 

options together with their default values are: 

 

- epsabs = 1.49e-08 

- epsrel = 1.49e-08 

- limit = 50 

- points = None 

- weight = None 

- wvar = None 

- wopts = None 

 

For more information on these options, see `quad` and `quad_explain`. 

 

full_output : bool, optional 

Partial implementation of ``full_output`` from scipy.integrate.quad. 

The number of integrand function evaluations ``neval`` can be obtained 

by setting ``full_output=True`` when calling nquad. 

 

Returns 

------- 

result : float 

The result of the integration. 

abserr : float 

The maximum of the estimates of the absolute error in the various 

integration results. 

out_dict : dict, optional 

A dict containing additional information on the integration. 

 

See Also 

-------- 

quad : 1-dimensional numerical integration 

dblquad, tplquad : double and triple integrals 

fixed_quad : fixed-order Gaussian quadrature 

quadrature : adaptive Gaussian quadrature 

 

Examples 

-------- 

>>> from scipy import integrate 

>>> func = lambda x0,x1,x2,x3 : x0**2 + x1*x2 - x3**3 + np.sin(x0) + ( 

... 1 if (x0-.2*x3-.5-.25*x1>0) else 0) 

>>> points = [[lambda x1,x2,x3 : 0.2*x3 + 0.5 + 0.25*x1], [], [], []] 

>>> def opts0(*args, **kwargs): 

... return {'points':[0.2*args[2] + 0.5 + 0.25*args[0]]} 

>>> integrate.nquad(func, [[0,1], [-1,1], [.13,.8], [-.15,1]], 

... opts=[opts0,{},{},{}], full_output=True) 

(1.5267454070738633, 2.9437360001402324e-14, {'neval': 388962}) 

 

>>> scale = .1 

>>> def func2(x0, x1, x2, x3, t0, t1): 

... return x0*x1*x3**2 + np.sin(x2) + 1 + (1 if x0+t1*x1-t0>0 else 0) 

>>> def lim0(x1, x2, x3, t0, t1): 

... return [scale * (x1**2 + x2 + np.cos(x3)*t0*t1 + 1) - 1, 

... scale * (x1**2 + x2 + np.cos(x3)*t0*t1 + 1) + 1] 

>>> def lim1(x2, x3, t0, t1): 

... return [scale * (t0*x2 + t1*x3) - 1, 

... scale * (t0*x2 + t1*x3) + 1] 

>>> def lim2(x3, t0, t1): 

... return [scale * (x3 + t0**2*t1**3) - 1, 

... scale * (x3 + t0**2*t1**3) + 1] 

>>> def lim3(t0, t1): 

... return [scale * (t0+t1) - 1, scale * (t0+t1) + 1] 

>>> def opts0(x1, x2, x3, t0, t1): 

... return {'points' : [t0 - t1*x1]} 

>>> def opts1(x2, x3, t0, t1): 

... return {} 

>>> def opts2(x3, t0, t1): 

... return {} 

>>> def opts3(t0, t1): 

... return {} 

>>> integrate.nquad(func2, [lim0, lim1, lim2, lim3], args=(0,0), 

... opts=[opts0, opts1, opts2, opts3]) 

(25.066666666666666, 2.7829590483937256e-13) 

 

""" 

depth = len(ranges) 

ranges = [rng if callable(rng) else _RangeFunc(rng) for rng in ranges] 

if args is None: 

args = () 

if opts is None: 

opts = [dict([])] * depth 

 

if isinstance(opts, dict): 

opts = [_OptFunc(opts)] * depth 

else: 

opts = [opt if callable(opt) else _OptFunc(opt) for opt in opts] 

return _NQuad(func, ranges, opts, full_output).integrate(*args) 

 

 

class _RangeFunc(object): 

def __init__(self, range_): 

self.range_ = range_ 

 

def __call__(self, *args): 

"""Return stored value. 

 

*args needed because range_ can be float or func, and is called with 

variable number of parameters. 

""" 

return self.range_ 

 

 

class _OptFunc(object): 

def __init__(self, opt): 

self.opt = opt 

 

def __call__(self, *args): 

"""Return stored dict.""" 

return self.opt 

 

 

class _NQuad(object): 

def __init__(self, func, ranges, opts, full_output): 

self.abserr = 0 

self.func = func 

self.ranges = ranges 

self.opts = opts 

self.maxdepth = len(ranges) 

self.full_output = full_output 

if self.full_output: 

self.out_dict = {'neval': 0} 

 

def integrate(self, *args, **kwargs): 

depth = kwargs.pop('depth', 0) 

if kwargs: 

raise ValueError('unexpected kwargs') 

 

# Get the integration range and options for this depth. 

ind = -(depth + 1) 

fn_range = self.ranges[ind] 

low, high = fn_range(*args) 

fn_opt = self.opts[ind] 

opt = dict(fn_opt(*args)) 

 

if 'points' in opt: 

opt['points'] = [x for x in opt['points'] if low <= x <= high] 

if depth + 1 == self.maxdepth: 

f = self.func 

else: 

f = partial(self.integrate, depth=depth+1) 

quad_r = quad(f, low, high, args=args, full_output=self.full_output, 

**opt) 

value = quad_r[0] 

abserr = quad_r[1] 

if self.full_output: 

infodict = quad_r[2] 

# The 'neval' parameter in full_output returns the total 

# number of times the integrand function was evaluated. 

# Therefore, only the innermost integration loop counts. 

if depth + 1 == self.maxdepth: 

self.out_dict['neval'] += infodict['neval'] 

self.abserr = max(self.abserr, abserr) 

if depth > 0: 

return value 

else: 

# Final result of n-D integration with error 

if self.full_output: 

return value, self.abserr, self.out_dict 

else: 

return value, self.abserr