/usr/include/dune/functions/functionspacebases/bsplinebasis.hh is in libdune-functions-dev 2.5.0-1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
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 | #ifndef DUNE_FUNCTIONS_FUNCTIONSPACEBASES_BSPLINEBASIS_HH
#define DUNE_FUNCTIONS_FUNCTIONSPACEBASES_BSPLINEBASIS_HH
/** \file
* \brief The B-spline global function space basis
*/
#include <numeric>
/** \todo Don't use this matrix */
#include <dune/common/dynmatrix.hh>
#include <dune/localfunctions/common/localbasis.hh>
#include <dune/common/diagonalmatrix.hh>
#include <dune/localfunctions/common/localkey.hh>
#include <dune/localfunctions/common/localfiniteelementtraits.hh>
#include <dune/geometry/type.hh>
#include <dune/functions/functionspacebases/nodes.hh>
namespace Dune
{
namespace Functions {
// A maze of dependencies between the different parts of this. We need lots of forward declarations
template<typename GV, typename R>
class BSplineLocalFiniteElement;
template<typename GV, class MI>
class BSplineNodeFactory;
/** \brief LocalBasis class in the sense of dune-localfunctions, presenting the restriction
* of a B-spline patch to a knot span
*
* \ingroup FunctionSpaceBasesImplementations
*
* \tparam GV Grid view that the basis is defined on
* \tparam R Number type used for spline function values
*/
template<class GV, class R>
class BSplineLocalBasis
{
friend class BSplineLocalFiniteElement<GV,R>;
typedef typename GV::ctype D;
enum {dim = GV::dimension};
public:
//! \brief export type traits for function signature
typedef LocalBasisTraits<D,dim,Dune::FieldVector<D,dim>,R,1,Dune::FieldVector<R,1>,
Dune::FieldMatrix<R,1,dim>, 2> Traits;
/** \brief Constructor with a given B-spline patch
*
* The patch object does all the work.
*/
BSplineLocalBasis(const BSplineNodeFactory<GV,FlatMultiIndex<std::size_t>>& nodeFactory,
const BSplineLocalFiniteElement<GV,R>& lFE)
: nodeFactory_(nodeFactory),
lFE_(lFE)
{}
/** \brief Evaluate all shape functions
* \param in Coordinates where to evaluate the functions, in local coordinates of the current knot span
*/
void evaluateFunction (const FieldVector<D,dim>& in,
std::vector<FieldVector<R,1> >& out) const
{
FieldVector<D,dim> globalIn = offset_;
scaling_.umv(in,globalIn);
nodeFactory_.evaluateFunction(globalIn, out, lFE_.currentKnotSpan_);
}
/** \brief Evaluate Jacobian of all shape functions
* \param in Coordinates where to evaluate the Jacobian, in local coordinates of the current knot span
*/
void evaluateJacobian (const FieldVector<D,dim>& in,
std::vector<FieldMatrix<D,1,dim> >& out) const
{
FieldVector<D,dim> globalIn = offset_;
scaling_.umv(in,globalIn);
nodeFactory_.evaluateJacobian(globalIn, out, lFE_.currentKnotSpan_);
for (size_t i=0; i<out.size(); i++)
for (int j=0; j<dim; j++)
out[i][0][j] *= scaling_[j][j];
}
//! \brief Evaluate all shape functions and derivatives of any order
template<size_t k>
inline void evaluate (const typename Dune::array<int,k>& directions,
const typename Traits::DomainType& in,
std::vector<typename Traits::RangeType>& out) const
{
switch(k)
{
case 0:
evaluateFunction(in, out);
break;
case 1:
{
FieldVector<D,dim> globalIn = offset_;
scaling_.umv(in,globalIn);
nodeFactory_.evaluate(directions, globalIn, out, lFE_.currentKnotSpan_);
for (size_t i=0; i<out.size(); i++)
out[i][0] *= scaling_[directions[0]][directions[0]];
break;
}
case 2:
{
FieldVector<D,dim> globalIn = offset_;
scaling_.umv(in,globalIn);
nodeFactory_.evaluate(directions, globalIn, out, lFE_.currentKnotSpan_);
for (size_t i=0; i<out.size(); i++)
out[i][0] *= scaling_[directions[0]][directions[0]]*scaling_[directions[1]][directions[1]];
break;
}
default:
DUNE_THROW(NotImplemented, "B-Spline derivatives of order " << k << " not implemented yet!");
}
}
/** \brief Polynomial order of the shape functions
*
* Unfortunately, the general interface of the LocalBasis class mandates that the 'order' method
* takes no arguments, and returns a single integer. It therefore cannot reflect that fact that
* a B-spline basis function can easily have different orders in the different coordinate directions.
* We therefore take the conservative choice and return the maximum over the orders of all directions.
*/
unsigned int order () const
{
return *std::max_element(nodeFactory_.order_.begin(), nodeFactory_.order_.end());
}
/** \brief Return the number of basis functions on the current knot span
*/
std::size_t size() const
{
return lFE_.size();
}
private:
const BSplineNodeFactory<GV,FlatMultiIndex<std::size_t>>& nodeFactory_;
const BSplineLocalFiniteElement<GV,R>& lFE_;
// Coordinates in a single knot span differ from coordinates on the B-spline patch
// by an affine transformation. This transformation is stored in offset_ and scaling_.
FieldVector<D,dim> offset_;
DiagonalMatrix<D,dim> scaling_;
};
/** \brief Attaches a shape function to an entity
*
* \ingroup FunctionSpaceBasesImplementations
*
* The attachment uses the same order as for Qk elements. This does *not* provide sufficient information
* to compute global indices for the shape functions. However, it does allow to find all degrees of freedom
* that belong to the grid boundary, if the knot vector is open.
*
* \note Currently only implemented for 1d and 2d grids. For higher dimensions you can still use
* the BSplineBasis, but you won't be able to find the degrees of freedom on the grid boundary.
*
* \tparam dim Dimension of the reference cube
*/
template<int dim>
class BSplineLocalCoefficients
{
// Return i as a d-digit number in the (k+1)-nary system
std::array<unsigned int,dim> multiindex (unsigned int i) const
{
std::array<unsigned int,dim> alpha;
for (int j=0; j<dim; j++)
{
alpha[j] = i % sizes_[j];
i = i/sizes_[j];
}
return alpha;
}
/** \brief Set the 'subentity' field for each dof for a 1d element */
void setup1d(std::vector<unsigned int>& subEntity)
{
if (sizes_[0]==1)
{
subEntity[0] = 0;
return;
}
/* edge and vertex numbering
0----0----1
*/
unsigned lastIndex=0;
subEntity[lastIndex++] = 0; // corner 0
for (unsigned i = 0; i < sizes_[0] - 2; ++i)
subEntity[lastIndex++] = 0; // inner dofs of element (0)
subEntity[lastIndex++] = 1; // corner 1
assert(size()==lastIndex);
}
void setup2d(std::vector<unsigned int>& subEntity)
{
unsigned lastIndex=0;
// LocalKey: entity number , entity codim, dof indices within each entity
/* edge and vertex numbering
2----3----3
| |
| |
0 1
| |
| |
0----2----1
*/
// lower edge (2)
subEntity[lastIndex++] = 0; // corner 0
for (unsigned i = 0; i < sizes_[0]-2; ++i)
subEntity[lastIndex++] = 2; // inner dofs of lower edge (2)
subEntity[lastIndex++] = 1; // corner 1
// iterate from bottom to top over inner edge dofs
for (unsigned e = 0; e < sizes_[1]-2; ++e)
{
subEntity[lastIndex++] = 0; // left edge (0)
for (unsigned i = 0; i < sizes_[0]-2; ++i)
subEntity[lastIndex++] = 0; // face dofs
subEntity[lastIndex++] = 1; // right edge (1)
}
// upper edge (3)
subEntity[lastIndex++] = 2; // corner 2
for (unsigned i = 0; i < sizes_[0]-2; ++i)
subEntity[lastIndex++] = 3; // inner dofs of upper edge (3)
subEntity[lastIndex++] = 3; // corner 3
assert(size()==lastIndex);
}
public:
void init(const std::array<unsigned,dim>& sizes)
{
sizes_ = sizes;
li_.resize(size());
// Set up array of codimension-per-dof-number
std::vector<unsigned int> codim(li_.size());
for (std::size_t i=0; i<codim.size(); i++)
{
codim[i] = 0;
// Codimension gets increased by 1 for each coordinate direction
// where dof is on boundary
std::array<unsigned int,dim> mIdx = multiindex(i);
for (int j=0; j<dim; j++)
if (mIdx[j]==0 or mIdx[j]==sizes[j]-1)
codim[i]++;
}
// Set up index vector (the index of the dof in the set of dofs of a given subentity)
// Algorithm: the 'index' has the same ordering as the dof number 'i'.
// To make it consecutive we interpret 'i' in the (k+1)-adic system, omit all digits
// that correspond to axes where the dof is on the element boundary, and transform the
// rest to the (k-1)-adic system.
std::vector<unsigned int> index(size());
for (std::size_t i=0; i<index.size(); i++)
{
index[i] = 0;
std::array<unsigned int,dim> mIdx = multiindex(i);
for (int j=dim-1; j>=0; j--)
if (mIdx[j]>0 and mIdx[j]<sizes[j]-1)
index[i] = (sizes[j]-1)*index[i] + (mIdx[j]-1);
}
// Set up entity and dof numbers for each (supported) dimension separately
std::vector<unsigned int> subEntity(li_.size());
if (subEntity.size() > 0)
{
if (dim==1) {
setup1d(subEntity);
} else if (dim==2 and sizes_[0]>1 and sizes_[1]>1) {
setup2d(subEntity);
}
}
for (size_t i=0; i<li_.size(); i++)
li_[i] = LocalKey(subEntity[i], codim[i], index[i]);
}
//! number of coefficients
std::size_t size () const
{
return std::accumulate(sizes_.begin(), sizes_.end(), 1, std::multiplies<unsigned int>());
}
//! get i'th index
const LocalKey& localKey (std::size_t i) const
{
return li_[i];
}
private:
// Number of shape functions on this element per coordinate direction
std::array<unsigned, dim> sizes_;
std::vector<LocalKey> li_;
};
/** \brief Local interpolation in the sense of dune-localfunctions, for the B-spline basis on tensor-product grids
*
* \ingroup FunctionSpaceBasesImplementations
*/
template<int dim, class LB>
class BSplineLocalInterpolation
{
public:
//! \brief Local interpolation of a function
template<typename F, typename C>
void interpolate (const F& f, std::vector<C>& out) const
{
DUNE_THROW(NotImplemented, "BSplineLocalInterpolation::interpolate");
}
};
/** \brief LocalFiniteElement in the sense of dune-localfunctions, for the B-spline basis on tensor-product grids
*
* \ingroup FunctionSpaceBasesImplementations
*
* This class ties together the implementation classes BSplineLocalBasis, BSplineLocalCoefficients, and BSplineLocalInterpolation
*
* \tparam D Number type used for domain coordinates
* \tparam R Number type used for spline function values
* \tparam dim Dimension of the patch
*/
template<class GV, class R>
class BSplineLocalFiniteElement
{
typedef typename GV::ctype D;
enum {dim = GV::dimension};
friend class BSplineLocalBasis<GV,R>;
public:
/** \brief Export various types related to this LocalFiniteElement
*/
typedef LocalFiniteElementTraits<BSplineLocalBasis<GV,R>,
BSplineLocalCoefficients<dim>,
BSplineLocalInterpolation<dim,BSplineLocalBasis<GV,R> > > Traits;
/** \brief Constructor with a given B-spline basis
*/
BSplineLocalFiniteElement(const BSplineNodeFactory<GV,FlatMultiIndex<std::size_t>>& nodeFactory)
: nodeFactory_(nodeFactory),
localBasis_(nodeFactory,*this)
{}
/** \brief Bind LocalFiniteElement to a specific knot span of the spline patch
*
* Elements are the non-empty knot spans, here we do the renumbering
*
* \param ijk Integer coordinates in the tensor product patch
*/
void bind(const std::array<uint,dim>& elementIdx)
{
/* \todo In the long run we need to precompute a table for this */
for (size_t i=0; i<elementIdx.size(); i++)
{
currentKnotSpan_[i] = 0;
// Skip over degenerate knot spans
while (nodeFactory_.knotVectors_[i][currentKnotSpan_[i]+1] < nodeFactory_.knotVectors_[i][currentKnotSpan_[i]]+1e-8)
currentKnotSpan_[i]++;
for (size_t j=0; j<elementIdx[i]; j++)
{
currentKnotSpan_[i]++;
// Skip over degenerate knot spans
while (nodeFactory_.knotVectors_[i][currentKnotSpan_[i]+1] < nodeFactory_.knotVectors_[i][currentKnotSpan_[i]]+1e-8)
currentKnotSpan_[i]++;
}
// Compute the geometric transformation from knotspan-local to global coordinates
localBasis_.offset_[i] = nodeFactory_.knotVectors_[i][currentKnotSpan_[i]];
localBasis_.scaling_[i][i] = nodeFactory_.knotVectors_[i][currentKnotSpan_[i]+1] - nodeFactory_.knotVectors_[i][currentKnotSpan_[i]];
}
// Set up the LocalCoefficients object
std::array<unsigned int, dim> sizes;
for (size_t i=0; i<dim; i++)
sizes[i] = size(i);
localCoefficients_.init(sizes);
}
/** \brief Hand out a LocalBasis object */
const BSplineLocalBasis<GV,R>& localBasis() const
{
return localBasis_;
}
/** \brief Hand out a LocalCoefficients object */
const BSplineLocalCoefficients<dim>& localCoefficients() const
{
return localCoefficients_;
}
/** \brief Hand out a LocalInterpolation object */
const BSplineLocalInterpolation<dim,BSplineLocalBasis<GV,R> >& localInterpolation() const
{
return localInterpolation_;
}
/** \brief Number of shape functions in this finite element */
uint size () const
{
std::size_t r = 1;
for (int i=0; i<dim; i++)
r *= size(i);
return r;
}
/** \brief Return the reference element that the local finite element is defined on (here, a hypercube)
*/
GeometryType type () const
{
return GeometryType(GeometryType::cube,dim);
}
//private:
/** \brief Number of degrees of freedom for one coordinate direction */
unsigned int size(int i) const
{
const auto& order = nodeFactory_.order_;
unsigned int r = order[i]+1; // The 'normal' value
if (currentKnotSpan_[i]<order[i]) // Less near the left end of the knot vector
r -= (order[i] - currentKnotSpan_[i]);
if ( order[i] > (nodeFactory_.knotVectors_[i].size() - currentKnotSpan_[i] - 2) )
r -= order[i] - (nodeFactory_.knotVectors_[i].size() - currentKnotSpan_[i] - 2);
return r;
}
const BSplineNodeFactory<GV,FlatMultiIndex<std::size_t>>& nodeFactory_;
BSplineLocalBasis<GV,R> localBasis_;
BSplineLocalCoefficients<dim> localCoefficients_;
BSplineLocalInterpolation<dim,BSplineLocalBasis<GV,R> > localInterpolation_;
// The knot span we are bound to
std::array<uint,dim> currentKnotSpan_;
};
template<typename GV, typename TP>
class BSplineNode;
template<typename GV, class MI, class TP>
class BSplineNodeIndexSet;
/** \brief Node factory for B-spline basis
*
* \ingroup FunctionSpaceBasesImplementations
*
* \tparam GV The GridView that the space is defined on
* \tparam MI Type to be used for multi-indices
*
* The BSplineNodeFactory can be used to embed a BSplineBasis
* in a larger basis for the construction of product spaces.
*/
template<typename GV, class MI>
class BSplineNodeFactory
{
static const int dim = GV::dimension;
/** \brief Simple dim-dimensional multi-index class */
class MultiDigitCounter
{
public:
/** \brief Constructs a new multi-index, and sets all digits to zero
* \param limits Number of different digit values for each digit, i.e., digit i counts from 0 to limits[i]-1
*/
MultiDigitCounter(const std::array<unsigned int,dim>& limits)
: limits_(limits)
{
std::fill(counter_.begin(), counter_.end(), 0);
}
/** \brief Increment the multi-index */
MultiDigitCounter& operator++()
{
for (int i=0; i<dim; i++)
{
++counter_[i];
// no overflow?
if (counter_[i] < limits_[i])
break;
counter_[i] = 0;
}
return *this;
}
/** \brief Access the i-th digit of the multi-index */
const unsigned int& operator[](int i) const
{
return counter_[i];
}
/** \brief How many times can you increment this multi-index before it overflows? */
unsigned int cycle() const
{
unsigned int r = 1;
for (int i=0; i<dim; i++)
r *= limits_[i];
return r;
}
private:
/** \brief The number of different digit values for each place */
const std::array<unsigned int,dim> limits_;
/** \brief The values of the multi-index. Each array entry is one digit */
std::array<unsigned int,dim> counter_;
};
public:
/** \brief The grid view that the FE space is defined on */
using GridView = GV;
using size_type = std::size_t;
template<class TP>
using Node = BSplineNode<GV, TP>;
template<class TP>
using IndexSet = BSplineNodeIndexSet<GV, MI, TP>;
/** \brief Type used for global numbering of the basis vectors */
using MultiIndex = MI;
using SizePrefix = Dune::ReservedVector<size_type, 1>;
// Type used for function values
using R = double;
/** \brief Construct a B-spline basis for a given grid view and set of knot vectors
*
* The grid *must* match the knot vectors, i.e.:
* - The grid must be structured and Cartesian, and have cube elements only
* - The number of elements in each direction must match the number of knot spans in that direction
* - In fact, the element spacing in any direction must match the knot spacing in that direction
* (disregarding knot multiplicities)
* - When ordering the grid elements according to their indices, the resulting order must
* be lexicographical, with the x-index increasing fastest.
*
* Unfortunately, not all of these conditions can be checked for automatically.
*
* \param knotVector A single knot vector, which will be used for all coordinate directions
* \param order B-spline order, will be used for all coordinate directions
* \param makeOpen If this is true, then knots are prepended and appended to the knot vector to make the knot vector 'open'.
* i.e., start and end with 'order+1' identical knots. Basis functions from such knot vectors are interpolatory at
* the end of the parameter interval.
*/
BSplineNodeFactory(const GridView& gridView,
const std::vector<double>& knotVector,
unsigned int order,
bool makeOpen = true)
: gridView_(gridView)
{
// \todo Detection of duplicate knots
std::fill(elements_.begin(), elements_.end(), knotVector.size()-1);
// Mediocre sanity check: we don't know the number of grid elements in each direction.
// but at least we know the total number of elements.
assert( std::accumulate(elements_.begin(), elements_.end(), 1, std::multiplies<uint>()) == gridView_.size(0) );
for (int i=0; i<dim; i++)
{
// Prepend the correct number of additional knots to open the knot vector
//! \todo maybe test whether the knot vector is already open?
if (makeOpen)
for (unsigned int j=0; j<order; j++)
knotVectors_[i].push_back(knotVector[0]);
knotVectors_[i].insert(knotVectors_[i].end(), knotVector.begin(), knotVector.end());
if (makeOpen)
for (unsigned int j=0; j<order; j++)
knotVectors_[i].push_back(knotVector.back());
}
std::fill(order_.begin(), order_.end(), order);
}
/** \brief Construct a B-spline basis for a given grid view with uniform knot vectors
*
* The grid *must* match the knot vectors, i.e.:
* - The grid must be structured and Cartesian, and have cube elements only
* - Bounding box and number of elements of the grid must match the corresponding arguments
* given to this constructor.
* - The element spacing must be uniform
* - When ordering the grid elements according to their indices, the resulting order must
* be lexicographical, with the x-index increasing fastest.
*
* Unfortunately, not all of these conditions can be checked for automatically.
*
* \param gridView The grid we are defining the basis on
* \param lowerLeft Lower left corner of the structured grid
* \param upperRight Upper right corner of the structured grid
* \param elements Number of elements in each coordinate direction
* \param order B-spline order, will be used for all coordinate directions
* \param makeOpen If this is true, then knots are prepended and appended to the knot vector to make the knot vector 'open'.
* i.e., start and end with 'order+1' identical knots. Basis functions from such knot vectors are interpolatory at
* the end of the parameter interval.
*/
BSplineNodeFactory(const GridView& gridView,
const FieldVector<double,dim>& lowerLeft,
const FieldVector<double,dim>& upperRight,
const array<unsigned int,dim>& elements,
unsigned int order,
bool makeOpen = true)
: elements_(elements),
gridView_(gridView)
{
// Mediocre sanity check: we don't know the number of grid elements in each direction.
// but at least we know the total number of elements.
assert( std::accumulate(elements_.begin(), elements_.end(), 1, std::multiplies<uint>()) == gridView_.size(0) );
for (int i=0; i<dim; i++)
{
// Prepend the correct number of additional knots to open the knot vector
//! \todo maybe test whether the knot vector is already open?
if (makeOpen)
for (unsigned int j=0; j<order; j++)
knotVectors_[i].push_back(lowerLeft[i]);
// Construct the actual knot vector
for (size_t j=0; j<elements[i]+1; j++)
knotVectors_[i].push_back(lowerLeft[i] + j*(upperRight[i]-lowerLeft[i]) / elements[i]);
if (makeOpen)
for (unsigned int j=0; j<order; j++)
knotVectors_[i].push_back(upperRight[i]);
}
std::fill(order_.begin(), order_.end(), order);
}
//! Initialize the global indices
void initializeIndices()
{}
//! Obtain the grid view that the basis is defined on
const GridView& gridView() const
{
return gridView_;
}
//! Update the stored grid view, to be called if the grid has changed
void update(const GridView& gv)
{
gridView_ = gv;
}
/**
* \brief Create tree node with given root tree path
*
* \tparam TP Type of root tree path
* \param tp Root tree path
*
* By passing a non-trivial root tree path this can be used
* to create a node suitable for being placed in a tree at
* the position specified by the root tree path.
*/
template<class TP>
Node<TP> node(const TP& tp) const
{
return Node<TP>{tp,this};
}
/**
* \brief Create tree node index set with given root tree path
*
* \tparam TP Type of root tree path
* \param tp Root tree path
*
* Create an index set suitable for the tree node obtained
* by node(tp).
*/
template<class TP>
IndexSet<TP> indexSet() const
{
return IndexSet<TP>{*this};
}
//! Return number of possible values for next position in multi index
size_type size(const SizePrefix prefix) const
{
if (prefix.size() == 0)
return size();
assert(false);
}
//! Get the total dimension of the space spanned by this basis
size_type dimension() const
{
return size();
}
//! Get the maximal number of DOFs associated to node for any element
size_type maxNodeSize() const
{
size_type result = 1;
for (int i=0; i<dim; i++)
result *= order_[i]+1;
return result;
}
//! \brief Total number of B-spline basis functions
unsigned int size () const
{
unsigned int result = 1;
for (size_t i=0; i<dim; i++)
result *= size(i);
return result;
}
//! \brief Number of shape functions in one direction
unsigned int size (size_t d) const
{
return knotVectors_[d].size() - order_[d] - 1;
}
/** \brief Evaluate all B-spline basis functions at a given point
*/
void evaluateFunction (const FieldVector<typename GV::ctype,dim>& in,
std::vector<FieldVector<R,1> >& out,
const std::array<uint,dim>& currentKnotSpan) const
{
// Evaluate
Dune::array<std::vector<R>, dim> oneDValues;
for (size_t i=0; i<dim; i++)
evaluateFunction(in[i], oneDValues[i], knotVectors_[i], order_[i], currentKnotSpan[i]);
std::array<unsigned int, dim> limits;
for (int i=0; i<dim; i++)
limits[i] = oneDValues[i].size();
MultiDigitCounter ijkCounter(limits);
out.resize(ijkCounter.cycle());
for (size_t i=0; i<out.size(); i++, ++ijkCounter)
{
out[i] = R(1.0);
for (size_t j=0; j<dim; j++)
out[i] *= oneDValues[j][ijkCounter[j]];
}
}
/** \brief Evaluate Jacobian of all B-spline basis functions
*
* In theory this is easy: just look up the formula in a B-spline text of your choice.
* The challenge is compute only the values needed for the current knot span.
*/
void evaluateJacobian (const FieldVector<typename GV::ctype,dim>& in,
std::vector<FieldMatrix<R,1,dim> >& out,
const std::array<uint,dim>& currentKnotSpan) const
{
// How many shape functions to we have in each coordinate direction?
std::array<unsigned int, dim> limits;
for (int i=0; i<dim; i++)
{
limits[i] = order_[i]+1; // The 'standard' value away from the boundaries of the knot vector
if (currentKnotSpan[i]<order_[i])
limits[i] -= (order_[i] - currentKnotSpan[i]);
if ( order_[i] > (knotVectors_[i].size() - currentKnotSpan[i] - 2) )
limits[i] -= order_[i] - (knotVectors_[i].size() - currentKnotSpan[i] - 2);
}
// The lowest knot spans that we need values from
std::array<unsigned int, dim> offset;
for (int i=0; i<dim; i++)
offset[i] = std::max((int)(currentKnotSpan[i] - order_[i]),0);
// Evaluate 1d function values (needed for the product rule)
Dune::array<std::vector<R>, dim> oneDValues;
// Evaluate 1d function values of one order lower (needed for the derivative formula)
Dune::array<std::vector<R>, dim> lowOrderOneDValues;
Dune::array<DynamicMatrix<R>, dim> values;
for (size_t i=0; i<dim; i++)
{
evaluateFunctionFull(in[i], values[i], knotVectors_[i], order_[i], currentKnotSpan[i]);
oneDValues[i].resize(knotVectors_[i].size()-order_[i]-1);
for (size_t j=0; j<oneDValues[i].size(); j++)
oneDValues[i][j] = values[i][order_[i]][j];
if (order_[i]!=0)
{
lowOrderOneDValues[i].resize(knotVectors_[i].size()-(order_[i]-1)-1);
for (size_t j=0; j<lowOrderOneDValues[i].size(); j++)
lowOrderOneDValues[i][j] = values[i][order_[i]-1][j];
}
}
// Evaluate 1d function derivatives
Dune::array<std::vector<R>, dim> oneDDerivatives;
for (size_t i=0; i<dim; i++)
{
oneDDerivatives[i].resize(limits[i]);
if (order_[i]==0) // order-zero functions are piecewise constant, hence all derivatives are zero
std::fill(oneDDerivatives[i].begin(), oneDDerivatives[i].end(), R(0.0));
else
{
for (size_t j=offset[i]; j<offset[i]+limits[i]; j++)
{
R derivativeAddend1 = lowOrderOneDValues[i][j] / (knotVectors_[i][j+order_[i]]-knotVectors_[i][j]);
R derivativeAddend2 = lowOrderOneDValues[i][j+1] / (knotVectors_[i][j+order_[i]+1]-knotVectors_[i][j+1]);
// The two previous terms may evaluate as 0/0. This is to be interpreted as 0.
if (std::isnan(derivativeAddend1))
derivativeAddend1 = 0;
if (std::isnan(derivativeAddend2))
derivativeAddend2 = 0;
oneDDerivatives[i][j-offset[i]] = order_[i] * ( derivativeAddend1 - derivativeAddend2 );
}
}
}
// Working towards computing only the parts that we really need:
// Let's copy them out into a separate array
Dune::array<std::vector<R>, dim> oneDValuesShort;
for (int i=0; i<dim; i++)
{
oneDValuesShort[i].resize(limits[i]);
for (size_t j=0; j<limits[i]; j++)
oneDValuesShort[i][j] = oneDValues[i][offset[i] + j];
}
// Set up a multi-index to go from consecutive indices to integer coordinates
MultiDigitCounter ijkCounter(limits);
out.resize(ijkCounter.cycle());
// Complete Jacobian is given by the product rule
for (size_t i=0; i<out.size(); i++, ++ijkCounter)
for (int j=0; j<dim; j++)
{
out[i][0][j] = 1.0;
for (int k=0; k<dim; k++)
out[i][0][j] *= (j==k) ? oneDDerivatives[k][ijkCounter[k]]
: oneDValuesShort[k][ijkCounter[k]];
}
}
//! \brief Evaluate Derivatives of all B-spline basis functions
template <size_type k>
void evaluate(const typename std::array<int,k>& directions,
const FieldVector<typename GV::ctype,dim>& in,
std::vector<FieldVector<R,1> >& out,
const std::array<uint,dim>& currentKnotSpan) const
{
if (k != 1 && k != 2)
DUNE_THROW(RangeError, "Differentiation order greater than 2 is not supported!");
// Evaluate 1d function values (needed for the product rule)
std::array<std::vector<R>, dim> oneDValues;
std::array<std::vector<R>, dim> oneDDerivatives;
std::array<std::vector<R>, dim> oneDSecondDerivatives;
// Evaluate 1d function derivatives
if (k==1)
for (size_t i=0; i<dim; i++)
evaluateAll(in[i], oneDValues[i], true, oneDDerivatives[i], false, oneDSecondDerivatives[i], knotVectors_[i], order_[i], currentKnotSpan[i]);
else
for (size_t i=0; i<dim; i++)
evaluateAll(in[i], oneDValues[i], true, oneDDerivatives[i], true, oneDSecondDerivatives[i], knotVectors_[i], order_[i], currentKnotSpan[i]);
// The lowest knot spans that we need values from
std::array<unsigned int, dim> offset;
for (int i=0; i<dim; i++)
offset[i] = std::max((int)(currentKnotSpan[i] - order_[i]),0);
// Set up a multi-index to go from consecutive indices to integer coordinates
std::array<unsigned int, dim> limits;
for (int i=0; i<dim; i++)
{
// In a proper implementation, the following line would do
//limits[i] = oneDValues[i].size();
limits[i] = order_[i]+1; // The 'standard' value away from the boundaries of the knot vector
if (currentKnotSpan[i]<order_[i])
limits[i] -= (order_[i] - currentKnotSpan[i]);
if ( order_[i] > (knotVectors_[i].size() - currentKnotSpan[i] - 2) )
limits[i] -= order_[i] - (knotVectors_[i].size() - currentKnotSpan[i] - 2);
}
// Working towards computing only the parts that we really need:
// Let's copy them out into a separate array
std::array<std::vector<R>, dim> oneDValuesShort;
for (int i=0; i<dim; i++)
{
oneDValuesShort[i].resize(limits[i]);
for (size_t j=0; j<limits[i]; j++)
oneDValuesShort[i][j] = oneDValues[i][offset[i] + j];
}
MultiDigitCounter ijkCounter(limits);
out.resize(ijkCounter.cycle());
if (k == 1)
{
// Complete Jacobian is given by the product rule
for (size_t i=0; i<out.size(); i++, ++ijkCounter)
{
out[i][0] = 1.0;
for (int l=0; l<dim; l++)
out[i][0] *= (directions[0]==l) ? oneDDerivatives[l][ijkCounter[l]]
: oneDValuesShort[l][ijkCounter[l]];
}
}
if (k == 2)
{
// Complete derivation by deriving the tensor product
for (size_t i=0; i<out.size(); i++, ++ijkCounter)
{
out[i][0] = 1.0;
for (int j=0; j<dim; j++)
{
if (directions[0] != directions[1]) //derivation in two different variables
if (directions[0] == j || directions[1] == j) //the spline has to be derived (once) in this direction
out[i][0] *= oneDDerivatives[j][ijkCounter[j]];
else //no derivation in this direction
out[i][0] *= oneDValuesShort[j][ijkCounter[j]];
else //spline is derived two times in the same direction
if (directions[0] == j) //the spline is derived two times in this direction
out[i][0] *= oneDSecondDerivatives[j][ijkCounter[j]];
else //no derivation in this direction
out[i][0] *= oneDValuesShort[j][ijkCounter[j]];
}
}
}
}
/** \brief Compute integer element coordinates from the element index
* \warning This method makes strong assumptions about the grid, namely that it is
* structured, and that indices are given in a x-fastest fashion.
*/
static std::array<unsigned int,dim> getIJK(typename GridView::IndexSet::IndexType idx, std::array<unsigned int,dim> elements)
{
std::array<uint,dim> result;
for (int i=0; i<dim; i++)
{
result[i] = idx%elements[i];
idx /= elements[i];
}
return result;
}
/** \brief Evaluate all one-dimensional B-spline functions for a given coordinate direction
*
* This implementations was based on the explanations in the book of
* Cottrell, Hughes, Bazilevs, "Isogeometric Analysis"
*
* \param in Scalar(!) coordinate where to evaluate the functions
* \param [out] out Vector containing the values of all B-spline functions at 'in'
*/
static void evaluateFunction (const typename GV::ctype& in, std::vector<R>& out,
const std::vector<R>& knotVector,
unsigned int order,
unsigned int currentKnotSpan)
{
std::size_t outSize = order+1; // The 'standard' value away from the boundaries of the knot vector
if (currentKnotSpan<order) // Less near the left end of the knot vector
outSize -= (order - currentKnotSpan);
if ( order > (knotVector.size() - currentKnotSpan - 2) )
outSize -= order - (knotVector.size() - currentKnotSpan - 2);
out.resize(outSize);
// It's not really a matrix that is needed here, a plain 2d array would do
DynamicMatrix<R> N(order+1, knotVector.size()-1);
// The text books on splines use the following geometric condition here to fill the array N
// (see for example Cottrell, Hughes, Bazilevs, Formula (2.1). However, this condition
// only works if splines are never evaluated exactly on the knots.
//
// for (size_t i=0; i<knotVector.size()-1; i++)
// N[0][i] = (knotVector[i] <= in) and (in < knotVector[i+1]);
for (size_t i=0; i<knotVector.size()-1; i++)
N[0][i] = (i == currentKnotSpan);
for (size_t r=1; r<=order; r++)
for (size_t i=0; i<knotVector.size()-r-1; i++)
{
R factor1 = ((knotVector[i+r] - knotVector[i]) > 1e-10)
? (in - knotVector[i]) / (knotVector[i+r] - knotVector[i])
: 0;
R factor2 = ((knotVector[i+r+1] - knotVector[i+1]) > 1e-10)
? (knotVector[i+r+1] - in) / (knotVector[i+r+1] - knotVector[i+1])
: 0;
N[r][i] = factor1 * N[r-1][i] + factor2 * N[r-1][i+1];
}
/** \todo We only hand out function values for those basis functions whose support overlaps
* the current knot span. However, in the preceding loop we still computed _all_ values_.
* This won't scale.
*/
for (size_t i=0; i<out.size(); i++) {
out[i] = N[order][std::max((int)(currentKnotSpan - order),0) + i];
}
}
/** \brief Evaluate all one-dimensional B-spline functions for a given coordinate direction
*
* This implementations was based on the explanations in the book of
* Cottrell, Hughes, Bazilevs, "Isogeometric Analysis"
*
* \todo This method is a hack! I computes the derivatives of ALL B-splines, even the ones that
* are zero on the current knot span. I need it as an intermediate step to get the derivatives
* working. It will/must be removed as soon as possible.
*
* \param in Scalar(!) coordinate where to evaluate the functions
* \param [out] out Vector containing the values of all B-spline functions at 'in'
*/
static void evaluateFunctionFull(const typename GV::ctype& in,
DynamicMatrix<R>& out,
const std::vector<R>& knotVector,
unsigned int order,
unsigned int currentKnotSpan)
{
// It's not really a matrix that is needed here, a plain 2d array would do
DynamicMatrix<R>& N = out;
N.resize(order+1, knotVector.size()-1);
// The text books on splines use the following geometric condition here to fill the array N
// (see for example Cottrell, Hughes, Bazilevs, Formula (2.1). However, this condition
// only works if splines are never evaluated exactly on the knots.
//
// for (size_t i=0; i<knotVector.size()-1; i++)
// N[0][i] = (knotVector[i] <= in) and (in < knotVector[i+1]);
for (size_t i=0; i<knotVector.size()-1; i++)
N[0][i] = (i == currentKnotSpan);
for (size_t r=1; r<=order; r++)
for (size_t i=0; i<knotVector.size()-r-1; i++)
{
R factor1 = ((knotVector[i+r] - knotVector[i]) > 1e-10)
? (in - knotVector[i]) / (knotVector[i+r] - knotVector[i])
: 0;
R factor2 = ((knotVector[i+r+1] - knotVector[i+1]) > 1e-10)
? (knotVector[i+r+1] - in) / (knotVector[i+r+1] - knotVector[i+1])
: 0;
N[r][i] = factor1 * N[r-1][i] + factor2 * N[r-1][i+1];
}
}
/** \brief Evaluate the second derivatives of all one-dimensional B-spline functions for a given coordinate direction
*
* \param in Scalar(!) coordinate where to evaluate the functions
* \param enableEvaluations switches calculation of desired derivatives on
* \param [out] out Vector containing the values of all B-spline derivatives at 'in'
* \param [out] outJac Vector containing the first derivatives of all B-spline derivatives at 'in' (only if calculation was switched on by enableEvaluations)
* \param [out] outHess Vector containing the second derivatives of all B-spline derivatives at 'in' (only if calculation was switched on by enableEvaluations)
*/
static void evaluateAll(const typename GV::ctype& in,
std::vector<R>& out,
bool evaluateJacobian, std::vector<R>& outJac,
bool evaluateHessian, std::vector<R>& outHess,
const std::vector<R>& knotVector,
unsigned int order,
unsigned int currentKnotSpan)
{
// How many shape functions to we have in each coordinate direction?
unsigned int limit;
limit = order+1; // The 'standard' value away from the boundaries of the knot vector
if (currentKnotSpan<order)
limit -= (order - currentKnotSpan);
if ( order > (knotVector.size() - currentKnotSpan - 2) )
limit -= order - (knotVector.size() - currentKnotSpan - 2);
// The lowest knot spans that we need values from
unsigned int offset;
offset = std::max((int)(currentKnotSpan - order),0);
// Evaluate 1d function values (needed for the product rule)
DynamicMatrix<R> values;
evaluateFunctionFull(in, values, knotVector, order, currentKnotSpan);
out.resize(knotVector.size()-order-1);
for (size_t j=0; j<out.size(); j++)
out[j] = values[order][j];
// Evaluate 1d function values of one order lower (needed for the derivative formula)
std::vector<R> lowOrderOneDValues;
if (order!=0)
{
lowOrderOneDValues.resize(knotVector.size()-(order-1)-1);
for (size_t j=0; j<lowOrderOneDValues.size(); j++)
lowOrderOneDValues[j] = values[order-1][j];
}
// Evaluate 1d function values of two order lower (needed for the (second) derivative formula)
std::vector<R> lowOrderTwoDValues;
if (order>1 && evaluateHessian)
{
lowOrderTwoDValues.resize(knotVector.size()-(order-2)-1);
for (size_t j=0; j<lowOrderTwoDValues.size(); j++)
lowOrderTwoDValues[j] = values[order-2][j];
}
// Evaluate 1d function derivatives
if (evaluateJacobian)
{
outJac.resize(limit);
if (order==0) // order-zero functions are piecewise constant, hence all derivatives are zero
std::fill(outJac.begin(), outJac.end(), R(0.0));
else
{
for (size_t j=offset; j<offset+limit; j++)
{
R derivativeAddend1 = lowOrderOneDValues[j] / (knotVector[j+order]-knotVector[j]);
R derivativeAddend2 = lowOrderOneDValues[j+1] / (knotVector[j+order+1]-knotVector[j+1]);
// The two previous terms may evaluate as 0/0. This is to be interpreted as 0.
if (std::isnan(derivativeAddend1))
derivativeAddend1 = 0;
if (std::isnan(derivativeAddend2))
derivativeAddend2 = 0;
outJac[j-offset] = order * ( derivativeAddend1 - derivativeAddend2 );
}
}
}
// Evaluate 1d function second derivatives
if (evaluateHessian)
{
outHess.resize(limit);
if (order<2) // order-zero functions are piecewise constant, hence all derivatives are zero
std::fill(outHess.begin(), outHess.end(), R(0.0));
else
{
for (size_t j=offset; j<offset+limit; j++)
{
assert(j+2 < lowOrderTwoDValues.size());
R derivativeAddend1 = lowOrderTwoDValues[j] / (knotVector[j+order]-knotVector[j]) / (knotVector[j+order-1]-knotVector[j]);
R derivativeAddend2 = lowOrderTwoDValues[j+1] / (knotVector[j+order]-knotVector[j]) / (knotVector[j+order]-knotVector[j+1]);
R derivativeAddend3 = lowOrderTwoDValues[j+1] / (knotVector[j+order+1]-knotVector[j+1]) / (knotVector[j+order]-knotVector[j+1]);
R derivativeAddend4 = lowOrderTwoDValues[j+2] / (knotVector[j+order+1]-knotVector[j+1]) / (knotVector[j+1+order]-knotVector[j+2]);
// The two previous terms may evaluate as 0/0. This is to be interpreted as 0.
if (std::isnan(derivativeAddend1))
derivativeAddend1 = 0;
if (std::isnan(derivativeAddend2))
derivativeAddend2 = 0;
if (std::isnan(derivativeAddend3))
derivativeAddend3 = 0;
if (std::isnan(derivativeAddend4))
derivativeAddend4 = 0;
outHess[j-offset] = order * (order-1) * ( derivativeAddend1 - derivativeAddend2 -derivativeAddend3 + derivativeAddend4 );
}
}
}
}
/** \brief Order of the B-spline for each space dimension */
array<unsigned int, dim> order_;
/** \brief The knot vectors, one for each space dimension */
array<std::vector<double>, dim> knotVectors_;
/** \brief Number of grid elements in the different coordinate directions */
std::array<uint,dim> elements_;
GridView gridView_;
};
template<typename GV, typename TP>
class BSplineNode :
public LeafBasisNode<std::size_t, TP>
{
static const int dim = GV::dimension;
using Base = LeafBasisNode<std::size_t,TP>;
public:
using size_type = std::size_t;
using TreePath = TP;
using Element = typename GV::template Codim<0>::Entity;
using FiniteElement = BSplineLocalFiniteElement<GV,double>;
BSplineNode(const TreePath& treePath, const BSplineNodeFactory<GV, FlatMultiIndex<std::size_t>>* nodeFactory) :
Base(treePath),
nodeFactory_(nodeFactory),
finiteElement_(*nodeFactory)
{}
//! Return current element, throw if unbound
const Element& element() const
{
return element_;
}
/** \brief Return the LocalFiniteElement for the element we are bound to
*
* The LocalFiniteElement implements the corresponding interfaces of the dune-localfunctions module
*/
const FiniteElement& finiteElement() const
{
return finiteElement_;
}
//! Bind to element.
void bind(const Element& e)
{
element_ = e;
auto elementIndex = nodeFactory_->gridView().indexSet().index(e);
finiteElement_.bind(nodeFactory_->getIJK(elementIndex,nodeFactory_->elements_));
this->setSize(finiteElement_.size());
}
protected:
const BSplineNodeFactory<GV, FlatMultiIndex<std::size_t>>* nodeFactory_;
FiniteElement finiteElement_;
Element element_;
};
template<typename GV, class MI, class TP>
class BSplineNodeIndexSet
{
enum {dim = GV::dimension};
public:
using size_type = std::size_t;
/** \brief Type used for global numbering of the basis vectors */
using MultiIndex = MI;
using NodeFactory = BSplineNodeFactory<GV, MI>;
using Node = typename NodeFactory::template Node<TP>;
BSplineNodeIndexSet(const NodeFactory& nodeFactory) :
nodeFactory_(&nodeFactory)
{}
/** \brief Bind the view to a grid element
*
* Having to bind the view to an element before being able to actually access any of its data members
* offers to centralize some expensive setup code in the 'bind' method, which can save a lot of run-time.
*/
void bind(const Node& node)
{
node_ = &node;
// Local degrees of freedom are arranged in a lattice.
// We need the lattice dimensions to be able to compute lattice coordinates from a local index
for (int i=0; i<dim; i++)
localSizes_[i] = node_->finiteElement().size(i);
}
/** \brief Unbind the view
*/
void unbind()
{
node_ = nullptr;
}
/** \brief Size of subtree rooted in this node (element-local)
*/
size_type size() const
{
return node_->finiteElement().size();
}
//! Maps from subtree index set [0..size-1] to a globally unique multi index in global basis
MultiIndex index(size_type i) const
{
std::array<unsigned int,dim> localIJK = nodeFactory_->getIJK(i, localSizes_);
const auto currentKnotSpan = node_->finiteElement().currentKnotSpan_;
const auto order = nodeFactory_->order_;
std::array<unsigned int,dim> globalIJK;
for (int i=0; i<dim; i++)
globalIJK[i] = std::max((int)currentKnotSpan[i] - (int)order[i], 0) + localIJK[i]; // needs to be a signed type!
// Make one global flat index from the globalIJK tuple
size_type globalIdx = globalIJK[dim-1];
for (int i=dim-2; i>=0; i--)
globalIdx = globalIdx * nodeFactory_->size(i) + globalIJK[i];
return { globalIdx };
}
protected:
const NodeFactory* nodeFactory_;
const Node* node_;
std::array<unsigned int, dim> localSizes_;
};
// *****************************************************************************
// This is the actual global basis implementation based on the reusable parts.
// *****************************************************************************
/** \brief A global B-spline basis
*
* \ingroup FunctionSpaceBasesImplementations
*
* \tparam GV The GridView that the space is defined on
*/
template<typename GV>
using BSplineBasis = DefaultGlobalBasis<BSplineNodeFactory<GV, FlatMultiIndex<std::size_t>> >;
} // namespace Functions
} // namespace Dune
#endif // DUNE_FUNCTIONS_FUNCTIONSPACEBASES_BSPLINEBASIS_HH
|