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
|
classdef GrB
%GrB GraphBLAS sparse matrices for Octave/MATLAB.
%
% GraphBLAS is a library for creating graph algorithms based on sparse
% linear algebraic operations over semirings. Visit http://graphblas.org
% for more details and resources. See also the SuiteSparse:GraphBLAS User
% Guide in this package.
%
% The GrB class represents a GraphBLAS sparse matrix. The GrB
% method creates a GraphBLAS sparse matrix from a built-in matrix. Other
% methods also generate GrB matrices. For example:
%
% G = GrB.subassign (C, M, A) ;
%
% constructs a GraphBLAS matrix G, which is the result of C<M>=A in
% GraphBLAS notation (like C(M)=A(M)). The matrices used in any
% GrB.method may be built-in matrices (sparse or full) or GraphBLAS
% matrices (hyper, sparse, bitmap, or full, by row or column), in any
% combination.
%
% --------------------
% The GrB constructor:
% --------------------
%
% The GrB constructor creates a GraphBLAS matrix. The input A may be
% any built-in or GraphBLAS matrix:
%
% C = GrB (A) ; GraphBLAS copy of a matrix A, same type
% C = GrB (m, n) ; m-by-n GraphBLAS double matrix, no entries
% C = GrB (..., type) ; create or typecast to a different type
% C = GrB (..., format) ; create in a specified format
%
% The m and n parameters above are built-in scalars. The type and format
% parameters are strings. The default format is 'by col', to match the
% format used in built-in (see also GrB.format), but many graph
% algorithms are faster if the format is 'by row'. The format can also
% specify the data structure to use (hypersparse, sparse, bitmap, and/or
% full).
%
% The usage C = GrB (m, n, type) is analgous to A = sparse (m, n), which
% creates an empty built-in sparse matrix A. The type parameter is a
% string, which defaults to 'double' if not present.
%
% For the usage C = GrB (A, type), A is either a built-in sparse or full
% matrix, or a GraphBLAS sparse matrix object. C is created as a
% GraphBLAS sparse matrix object that contains a copy of A, typecasted
% to the given type if the type string does not match the type of A.
% If the type string is not present it defaults to 'double'.
%
% --------------------
% Matrix types:
% --------------------
%
% Most of the valid type strings correspond to built-in class of the
% same name (see 'help class'):
%
% 'logical' 8-bit boolean
% 'int8' 8-bit signed integer
% 'int16' 16-bit signed integer
% 'int32' 32-bit signed integer
% 'int64' 64-bit signed integer
% 'uint8' 8-bit unsigned integer
% 'uint16' 16-bit unsigned integer
% 'uint32' 32-bit unsigned integer
% 'uint64' 64-bit unsigned integer
% 'double' 64-bit floating-point (real, not complex)
% 'single' 32-bit floating-point (real, not complex)
% 'single complex' single complex
% 'double complex' double complex (also just 'complex')
%
% In built-in matrices, complex is an attribute, not a class. In GrB
% matrices, 'double complex' and 'single complex' are treated as their
% own data types.
%
% ---------------
% Matrix formats:
% ---------------
%
% The format of a GraphBLAS matrix can have a large impact on
% performance. GraphBLAS matrices can be stored by column or by row.
% The corresponding format string is 'by col' or 'by row', respectively.
% Since the only format for built-in sparse and full matrices is 'by
% col', that is the default format for GraphBLAS matrices via this
% interface to GraphBLAS. However, the default for the C API is 'by
% row' since graph algorithms tend to be faster with that format.
%
% Column vectors are always stored 'by col', and row vectors are always
% stored 'by row'. The format for new matrices propagates from the
% format of their inputs. For example with C=A*B, C takes on the same
% format as A, unless A is a vector, in which case C takes on the format
% of B. If both A and B are vectors, then the format of C is determined
% by the descriptor (if present), or by the default format (see
% GrB.format).
%
% When a GraphBLAS matrix is converted into a built-in sparse or full
% matrix, it is always returned as 'by col'.
%
% The format can also specify the data structure to use. By default
% GraphBLAS selects automatically between hypersparse, sparse, bitmap,
% and full formats. See 'help GrB.format' for details.
%
%--------------------
% Integer operations:
%--------------------
%
% Operations on integer values differ from built-in operations, where
% uint8(255)+1 is 255, since the arithmetic saturates. This is not
% possible in matrix operations such as C=A*B, since saturation of
% integer arithmetic would render most of the monoids useless.
% GraphBLAS instead computes a result modulo the word size, so that
% GrB(uint8(255))+1 is zero. However, new unary and binary operators
% could be added so that element-wise operations saturate. The C
% interface allows for arbitrary creation of user-defined operators, so
% this could be added in the future. See 'help GrB/MATLAB_vs_GrB' for
% more details.
%
%-------------------------------------------------------------------------
% Methods for the GrB class:
%-------------------------------------------------------------------------
%
% C = GrB (...) construct a GraphBLAS matrix
%
% Overloaded operators (all except 'colon'):
%
% C = and (A, B) C = A & B
% C = ctranspose (G) C = G'
% i = end (G, k, ndims) A(1:end,1:end)
% C = eq (A, B) C = A == B
% C = ge (A, B) C = A >= B
% C = gt (A, B) C = A > B
% C = horzcat (A, B) C = [A , B]
% C = ldivide (A, B) C = A .\ B
% C = le (A, B) C = A <= B
% C = lt (A, B) C = A < B
% C = minus (A, B) C = A - B
% C = mldivide (A, B) C = A \ B
% C = mpower (A, B) C = A ^ B
% C = mrdivide (A, B) C = A / B
% C = mtimes (A, B) C = A * B
% C = ne (A, B) C = A ~= B
% C = not (G) C = ~G
% C = or (A, B) C = A | B
% C = plus (A, B) C = A + B
% C = power (A, B) C = A .^ B
% C = rdivide (A, B) C = A ./ B
% I = subsindex (G) X = A (G)
% C = subsasgn (C, S, A) C (I,J) = A or C (M) = A
% C = subsref (A, S) C = A (I,J) or C = A (M)
% C = times (A, B) C = A .* B
% C = transpose (G) C = G.'
% C = uminus (G) C = -G
% C = uplus (G) C = +G
% C = vertcat (A, B) C = [A ; B]
%
% Overloaded functions:
%
% C = abs (G) absolute value
% C = acos (G) inverse cosine
% C = acosh (G) inverse hyperbolic cosine
% C = acot (G) inverse cotangent
% C = acoth (G) inverse hyperbolic cotangent
% C = acsc (G) inverse cosecant
% C = acsch (G) inverse hyperbolic cosecant
% C = all (G, ...) reduce via '&', to vector or scalar
% p = amd (G, ...) approximate minimum degree ordering
% C = angle (G) phase angle of a complex matrix
% C = any (G, ...) reduce via '|', to vector or scalar
% C = asec (G) inverse secant
% C = asech (G) inverse hyperbolic secant
% C = asin (G) inverse sine
% C = asinh (G) inverse hyperbolic sine
% assert (G) generate an error if G is false
% C = atan (G) inverse tangent
% C = atanh (G) inverse hyperbolic tangent
% C = atan2 (A, B) inverse tangent (four-quadrant)
%
% [lo, hi] = bandwidth (G, ...) lower and upper bandwidth of G
% C = bitand (A, B, ...) bitwise and
% C = bitcmp (A, ...) bitwise negation
% C = bitget (A, B, ...) get bits
% C = bitset (A, B, ...) set bits
% C = bitshift (A, B, ...) shift bits
% C = bitor (A, B, ...) bitwise or
% C = bitxor (A, B, ...) bitwise xor
%
% C = cast (G, ...) cast GrB matrix to built-in matrix
% C = cat (dim, ...) contatenate matrices
% C = ceil (G) round towards infinity
% C = cell2mat (A) concatenate a cell array of matrices
% p = colamd (G) column approximate minimum degree ordering
% C = complex (G) cast GrB matrix to built-in sparse complex
% C = conj (G) complex conjugate
% C = cos (G) cosine
% C = cosh (G) hyperbolic cosine
% C = cot (G) cotangent
% C = coth (G) hyperbolic cotangent
% C = csc (G) cosecant
% C = csch (G) hyperbolic cosecant
% C = cbrt (G) cube root
%
% C = diag (A, k) diagonal matrices and diagonals
% DiGraph = digraph (G,...) directed Graph
% disp (A, level) display a built-in or GrB matrix A
% display (G) display a GrB matrix G; same as disp(G,2)
% [...] = dmperm (G) Dulmage-Mendelsohn permutation
% C = double (G) cast GrB matrix to built-in sparse double
%
% [V, ...] = eig (G,...) eigenvalues and eigenvectors
% G = GrB.empty (m, n) empty matrix for the GrB class
% C = eps (G) floating-point spacing
% C = erf (G) error function
% C = erfc (G) complementary error function
% p = etree (G) elimination tree
% C = exp (G) natural exponent
% C = expm1 (G) exp (x) - 1
%
% [I,J,X] = find (G, ...) extract entries from a matrix
% C = fix (G) round towards zero
% C = flip (G, dim) flip the order of entries
% C = floor (G) round towards -infinity
% c = fprintf (...) print to a file or to the Command Window
% C = full (G, ...) adds explicit zeros or id values
%
% C = gamma (G) gamma function
% C = gammaln (G) logarithm of gamma function
% Graph = graph (G, ...) undirected graph
%
% C = hypot (A, B) sqrt of sum of squares
%
% C = imag (G) imaginary part of a complex matrix
% C = int8 (G) cast GrB matrix to built-in full int8
% C = int16 (G) cast GrB matrix to built-in full int16
% C = int32 (G) cast GrB matrix to built-in full int32
% C = int64 (G) cast GrB matrix to built-in full int64
% s = isa (G, classname) check if a GrB matrix is of a specific class
% s = isbanded (G,...) true if G is banded
% s = iscolumn (G) true if n=1, for an m-by-n GrB matrix G
% s = isdiag (G) true if G is diagonal
% s = isempty (G) true if any dimension of G is zero
% s = isequal (A, B) test if equal
% C = isfinite (G) test if finite
% s = isfloat (G) true if GrB matrix is double, single, complex
% s = ishermitian (G) true if G is Hermitian
% C = isinf (G) test if infinite
% s = isinteger (G) true if GrB matrix is int8, int16, ..., uint64
% s = islogical (G) true if GrB matrix is logical
% s = ismatrix (G) true for any GrB matrix G
% C = isnan (G) test if NaN
% s = isnumeric (G) true for any GrB matrix G (even logical)
% s = isreal (G) true if GrB matrix is not complex
% s = isrow (G) true if m=1, for an m-by-n GrB matrix G
% s = isscalar (G) true if G is a 1-by-1 GrB matrix
% s = issparse (G) true for any GrB matrix G
% s = issymmetric (G) true if G is symmetric
% s = istril (G) true if G is lower triangular
% s = istriu (G) true if G is upper triangular
% s = isvector (G) true if m=1 or n=1, for an m-by-n GrB matrix G
%
% C = kron (A, B) Kronecker product
%
% n = length (G) length of a GrB vector
% C = log (G) natural logarithm
% C = log10 (G) base-10 logarithm
% C = log1p (G) log (1+x)
% [F, E] = log2 (G) base-2 logarithm
% C = logical (G) cast GrB matrix to built-in sparse logical
%
% C = mat2cell (A,m,n) break a matrix into a cell array of matrices
% C = max (A,B,option) reduce via max, to vector or scalar
% C = min (A,B,option) reduce via min, to vector or scalar
%
% e = nnz (G) number of entries in a GrB matrix G
% X = nonzeros (G) extract all entries from a GrB matrix
% s = norm (G, kind) norm of a GrB matrix
% C = num2cell (A,dim) convert a matrix into a cell array
% e = numel (G) m*n for an m-by-n GrB matrix G
% e = nzmax (G) number of entries in a GrB matrix G
%
%
% C = pow2 (F, E) base-2 power
% C = prod (G, option) reduce via product, to vector or scalar
%
% C = real (G) real part of a complex matrix
% C = repmat (G, ...) replicate and tile a GraphBLAS matrix
% C = reshape (G, ...) reshape a GraphBLAS matrix
% C = round (G) round towards nearest
%
% C = sec (G) secant
% C = sech (G) hyperbolic secant
% C = sign (G) signum function
% C = sin (G) sine
% C = single (G) cast GrB matrix to built-in full single
% C = sinh (G) hyperbolic sine
% [m,n,t] = size (G,dim) size and type of a GrB matrix
% C = sparse (G) makes a copy of a GrB matrix
% C = spfun (fun, G) evaluate a function on the entries of G
% C = spones (G, type) return pattern of GrB matrix
% C = sprand (...) random GraphBLAS matrix
% C = sprandn (...) random GraphBLAS matrix, normal distribution
% C = sprandsym (...) random symmetric GraphBLAS matrix
% c = sprintf (...) print to a string
% C = sqrt (G) element-wise square root
% C = sum (G, option) reduce via sum, to vector or scalar
% p = symamd (G) approximate minimum degree ordering
% p = symrcm (G) reverse Cuthill-McKee ordering
%
% C = tan (G) tangent
% C = tanh (G) hyperbolic tangent
% L = tril (G, k) lower triangular part of GrB matrix G
% U = triu (G, k) upper triangular part of GrB matrix G
%
% C = uint8 (G) cast GrB matrix to built-in full uint8
% C = uint16 (G) cast GrB matrix to built-in full uint16
% C = uint32 (G) cast GrB matrix to built-in full uint32
% C = uint64 (G) cast GrB matrix to built-in full uint64
%
% C = xor (A, B) exclusive or
%
%-------------------------------------------------------------------------
% Static Methods:
%-------------------------------------------------------------------------
%
% The Static Methods for the GrB class can be used on input matrices of
% any kind: GraphBLAS sparse matrices, built-in sparse matrices, or
% built-in full matrices, in any combination. The output matrix C is a
% GraphBLAS matrix, by default, but can be optionally returned as a
% built-in sparse or full matrix. The static methods divide into three
% categories: those that perform basic functions, graph algorithms, and
% the 12 foundational GraphBLAS operations.
%
%---------------------------
% GraphBLAS basic functions:
%---------------------------
%
% context:
% GrB.clear clear GraphBLAS workspace and settings
% GrB.finalize finish GraphBLAS
% GrB.init initialize GraphBLAS
% t = GrB.threads (t) set/get # of threads to use in GraphBLAS
% c = GrB.chunk (c) set/get chunk size to use in GraphBLAS
% b = GrB.burble (b) set/get burble (diagnostic output)
%
% info:
% GrB.binopinfo (op, type) list properties of a binary operator
% GrB.descriptorinfo (d) list properties of a descriptor
% GrB.monoidinfo (op, type) list properties of a monoid
% GrB.selectopinfo (op, type) list properties of a select operator
% GrB.semiringinfo (s, type) list properties of a semiring
% GrB.unopinfo (op, type) list properties of a unary operator
%
% basic matrices:
% C = GrB.false (...) all-false logical matrix
% C = GrB.true (...) all-true logical matrix
% C = GrB.ones (...) matrix with all ones
% C = GrB.zeros (...) all-zero matrix
%
% operations:
% C = GrB.build (I,J,X,m,n,dup,type,desc) build a GrB matrix from
% list of entries (like C=sparse(I,J,X...))
% [C,I,J] = GrB.compact (A,id) remove empty rows and columns
% c = GrB.entries (A,...) count or query entries in a matrix
% C = GrB.expand (scalar, A) expand a scalar (C = scalar*spones(A))
% [I,J,X] = GrB.extracttuples (A,desc) extract all entries (like 'find')
% C = GrB.eye (m,n,type) identity matrix of any type (like 'speye')
% f = GrB.format (f) set/get matrix format by row or col
% s = GrB.isbyrow (A) true if format f A is 'by row'
% s = GrB.isbycol (A) true if format f A is 'by col'
% s = GrB.isfull (A) true if all entries present
% s = GrB.issigned (type) true if type is signed
% c = GrB.nonz (A,...) count or query nonzeros in a matrix
% s = GrB.normdiff (A,B,kind) norm (A-B,kind)
% C = GrB.offdiag (A) prune diagonal entries
% C = GrB.prune (A, id) prune entries equal to id
% C = GrB.random (...) random GraphBLAS matrix (like 'sprand')
% C = GrB.speye (m,n,type) identity matrix of any type (like 'speye')
% t = GrB.type (A) get the type of a built-in or GrB matrix A
% v = GrB.version string with SuiteSparse:GraphBLAS version
% v = GrB.ver struct with SuiteSparse:GraphBLAS version
%
% load/save:
% C = GrB.load (filename) load a single matrix from a file
% GrB.save (A, filename) save a single matrix to a file
%
%-------------------------------------
% Static Methods for graph algorithms:
%-------------------------------------
%
% [v, parent] = GrB.bfs (A, s, ...) ; breadth-first search
% Y = GrB.dnn (W, bias, Y0) ; deep neural network
% C = GrB.incidence (A, ...) ; incidence matrix
% C = GrB.ktruss (A, k, check) ; k-truss
% L = GrB.laplacian (A, type, check) ; Laplacian graph
% iset = GrB.mis (A, check) ; maximal independent set
% r = GrB.pagerank (A, opts) ; PageRank of a matrix
% s = GrB.tricount (A, check) ; triangle count
%
%-----------------------------------
% Foundational GraphBLAS operations:
%-----------------------------------
%
% GraphBLAS has 12 foundational operations, listed below. All have
% similar parameters. The full set of input parameters is listed in the
% order in which they appear in the GraphBLAS C API, except that for the
% @GrB interface, Cin and C are different matrices. They combine into a
% single input/output matrix in the GraphBLAS C API. In the @GrB
% interface, many of the parameters become optional, and they can appear
% in different order.
%
% GrB.apply apply a unary operator
% GrB.apply2 apply a binary operator
% GrB.assign sparse matrix assignment, such as C(I,J)=A
% GrB.eadd element-wise addition
% GrB.eunion element-wise union
% GrB.emult element-wise multiplication
% GrB.extract extract submatrix, like C=A(I,J)
% GrB.kronecker Kronecker product
% GrB.mxm sparse matrix-matrix multiplication over a semiring
% GrB.reduce reduce a matrix to a scalar
% GrB.select select a subset of entries from a matrix
% GrB.subassign sparse matrix assignment, such as C(I,J)=A
% GrB.trans transpose a matrix
% GrB.vreduce reduce a matrix to a vector
%
% In GraphBLAS notation (with C, Cin arguments for the one matrix
% C), these take the following form:
%
% C<#M,replace> = accum (C, operation (A or A', B or B'))
%
% C is both an input and output matrix. In this interface to
% GraphBLAS, it is split into Cin (the value of C on input) and C
% (the value of C on output). M is the optional mask matrix, and #M is
% either M or ~M depending on whether or not the mask is complemented
% via the desc.mask option. The replace option is determined by
% desc.out; if present, C is cleared after it is used in the accum
% operation but before the final assignment. A and/or B may optionally
% be transposed via the descriptor fields desc.in0 and desc.in1,
% respectively. To select the format of C, use desc.format. See
% GrB.descriptorinfo for more details.
%
% accum is optional; if not is not present, then the operation becomes
% C<...> = operation(A,B). Otherwise, C = C + operation(A,B) is
% computed where '+' is the accum operator. It acts like a sparse
% matrix addition (see GrB.eadd), in terms of the structure of the
% result C, but any binary operator can be used.
%
% The mask M acts like built-in logical indexing. If M(i,j)=1 then
% C(i,j) can be modified; if zero, it cannot be modified by the
% operation.
%
% The full list of parameters is shown below:
%
% C = GrB.apply (Cin, M, accum, op, A, desc)
% C = GrB.apply2 (Cin, M, accum, op, A, B, desc)
% C = GrB.assign (Cin, M, accum, A, I, J, desc)
% C = GrB.eadd (Cin, M, accum, op, A, B, desc)
% C = GrB.eunion (Cin, M, accum, op, A, a, B, b, desc)
% C = GrB.emult (Cin, M, accum, op, A, B, desc)
% C = GrB.extract (Cin, M, accum, A, I, J, desc)
% C = GrB.kronecker (Cin, M, accum, op, A, B, desc)
% C = GrB.mxm (Cin, M, accum, op, A, B, desc)
% C = GrB.reduce (Cin, accum, op, A, desc)
% C = GrB.select (Cin, M, accum, op, A, b, desc)
% C = GrB.subassign (Cin, M, accum, A, I, J, desc)
% C = GrB.trans (Cin, M, accum, A, desc)
% C = GrB.vreduce (Cin, M, accum, op, A, desc)
%
% The parameters divide into 4 classes: matrices, strings, cells, and a
% single optional struct (the descriptor). The order of parameters
% between the matrices, strings, and cell classes is arbitrary. The
% order of parameters within a class is important; for example, if a
% method takes 4 matrix inputs, then they must appear in the order Cin,
% M, A, and then B. However, if a single string appears as a
% parameter, it can appear anywhere within the list of 4 matrices.
%
% (1) Cin, M, A, B are matrices, and a and b are scalars (eunion only).
% If the method takes up to 4 matrices
% (mxm, kronecker, select (with operator requiring a b
% parameter), eadd, emult, apply2), then they appear in this order:
% with 2 matrix inputs: A, B
% with 3 matrix inputs: Cin, A, B
% with 4 matrix inputs: Cin, M, A, B
% For GrB.select, b is a scalar. For GrB.apply2, either A or B
% is a scalar.
%
% If the method takes up to 3 matrices (vreduce, apply, assign,
% subassign, extract, trans, or select without b):
% with 1 matrix input: A
% with 2 matrix inputs: Cin, A
% with 3 matrix inputs: Cin, M, A
% Note that assign and subassign require Cin.
%
% If the method takes up to 2 input matrices (the reduce method):
% with 1 matrix input: A
% with 2 matrix inputs: Cin, A
%
% (2) accum and op are strings. The accum string is always optional.
% If the method has an op parameter, then it is a required input.
%
% If the method has both parameters, and just one string appears,
% it is the op, which is a semiring for mxm, a unary operator for
% apply, a select operator for the select method, and a binary
% operator for all other methods. If 2 strings appear, the first
% one is the accum the second is the op. If the accum appears then
% Cin must also appear as a matrix input.
%
% If the method has no op (assign, subassign, extract, trans), but
% just an accum parameter, then 0 or 1 strings may appear in the
% parameter list. If a string appears, it is the accum.
%
% (3) I and J are cell arrays. For details, see the assign, subassign,
% and extract methods; a short summary appears below. Both are
% optional:
% with no cell inputs: default for I and J
% with 1 cell inputs: I, default for J
% with 2 cell inputs: I, J
%
% Each cell array may appear with 0, 1, 2, or 3 items:
% 0: { } ":" in built-in notation
% 1: { list } a list of integer indices
% 2: { start,fini } start:fini in built-in notation
% 3: { start,inc,fini } start:inc:fini in built-in notation
%
% (4) The descriptor is an optional struct. If present, it must
% appear last, after all other parameters.
%
% Some valid uses are shown below, along with their equivalent in
% GraphBLAS notation. For the first three mxm examples, the four
% matrices C, M, A, and B must appear in that order, and the two
% strings '+' and '+.*' must appear in that order, but the matrices and
% strings may be interleaved arbitrarily.
%
% C = GrB.apply (C, M, '|', '~', A) C<M> |= ~A
% C = GrB.apply ('~', A) C = ~A
%
% C = GrB.assign (C, M, '+', A, I, J) C(I,J)<M> += A
% C = GrB.assign (C, I, J, M, '+', A) C(I,J)<M> += A
%
% C = GrB.assign (C, A, I, J) C(I,J) = A
% C = GrB.assign (C, I, J, A) C(I,J) = A
% C = GrB.assign (C, A) C = A
% C = GrB.assign (C, M, A) C<M> = A
% C = GrB.assign (C, M, '+', A) C<M> += A
% C = GrB.assign (C, '+', A, I) C (I,:) += A
%
% C = GrB.emult (C, M, '+', A, '*', B) C<M> += A.*B
% C = GrB.emult (A, '*', B) C = A.*B
%
% C = GrB.extract (C, M, '+', A, I, J) C<M> += A(I,J)
% C = GrB.extract (A, I, J) C = A(I,J)
% C = GrB.extract (I, J, A) C = A(I,J)
% C = GrB.extract (A) C = A
% C = GrB.extract (C, M, A) C<M> = A
% C = GrB.extract (C, M, '+', A) C<M> += A
% C = GrB.extract (C, '+', A, I) C += A(I,:)
%
% C = GrB.mxm (C, M, '+', '+.*', A, B) C<M> += A*B
% C = GrB.mxm (C, M, '+', A, '+.*', B) C<M> += A*B
% C = GrB.mxm ('+', '+,*', C, M, A, B) C<M> += A*B
%
% C = GrB.mxm ('+.*', A, B) C = A*B
% C = GrB.mxm (A, '+.*', B) C = A*B
% C = GrB.mxm (C, M, A, '+.*', B) C<M> = A*B
%
% c = GrB.reduce (c, '+', 'max', A) c += max (A)
% c = GrB.reduce ('max', A) c = max (A)
% c = GrB.reduce (A, 'max') c = max (A)
% c = GrB.reduce (c, 'max', A) c = max (A)
%
% See also sparse.
%
% SuiteSparse:GraphBLAS, Timothy A. Davis, (c) 2017-2022, All Rights Reserved.
% SPDX-License-Identifier: Apache-2.0
properties (SetAccess = private, GetAccess = private)
% The struct contains the entire opaque content of a GraphBLAS
% GrB_Matrix.
opaque = [ ] ;
end
methods
%---------------------------------------------------------------------
% GrB: GraphBLAS matrix constructor
%---------------------------------------------------------------------
function C = GrB (arg1, arg2, arg3, arg4)
%GRB GraphBLAS constructor: create a GraphBLAS sparse matrix.
%
% C = GrB (A) ; GrB copy of a matrix A, same type and format
%
% C = GrB (A, type) ; GrB typecasted copy of a matrix A, same format
% C = GrB (A, format) ; GrB copy of a matrix A, with given format
% C = GrB (m, n) ; empty m-by-n GrB double matrix, default format
%
% C = GrB (A, type, format) ; GrB copy of A, new type and format
% C = GrB (A, format, type) ; ditto
%
% C = GrB (m,n, type) ; empty m-by-n GrB type matrix, default format
% C = GrB (m,n, format) ; empty m-by-n GrB double matrix, given format
%
% C = GrB (m,n,type,format) ; empty m-by-n matrix, given type & format
% C = GrB (m,n,format,type) ; ditto
%
% See also sparse.
if (nargin == 1)
if (isstruct (arg1))
% C = GrB (A), where the input A is a GraphBLAS struct as
% returned by another GrB* function, but this usage is not
% meant for the end-user. It is only used internally in
% @GrB, to convert a GraphBLAS struct computed by a
% GraphBLAS mexFunction into a GrB matrix object.
C.opaque = arg1 ;
elseif (isobject (arg1))
% arg1 is already a GrB matrix; make a deep copy
C.opaque = gbnew (arg1.opaque) ;
else
% arg1 is a built-in matrix; convert to a GrB matrix
C.opaque = gbnew (arg1) ;
end
else
if (isobject (arg1))
% extract the contents of the GrB object as its opaque
% struct so the gbnew mexFunction can access it.
arg1 = arg1.opaque ;
end
% varargin is more elegant, but it is slower than the switch
switch (nargin)
case 2
C.opaque = gbnew (arg1, arg2) ;
case 3
C.opaque = gbnew (arg1, arg2, arg3) ;
case 4
C.opaque = gbnew (arg1, arg2, arg3, arg4) ;
end
end
end
%---------------------------------------------------------------------
% implicitly-defined methods
%---------------------------------------------------------------------
% The following methods work without any implemention needed here,
% because they are built-in m-files that can operate with GrB
% inputs:
%
% matrix operations: flipdim fliplr flipud cast isrow iscolumn ndims
% sprank etreeplot spy gplot reallog realpow realsqrt
%
% iterative solvers: bicgstabl bicgstab cgs minres gmres bicg pcg
% qmr rjr tfqmr lsqr
%---------------------------------------------------------------------
% FUTURE:: many these could also be overloaded:
%---------------------------------------------------------------------
% methods in the ops folder:
%
% colon idivide ismembertol uniquetol
% m-files: intersect ismember setdiff setxorunion unique
% (if 'sort' is overloaded, and 1D indexing added,
% then all these will work for GrB matrices)
% methods in the datatypes folder:
%
% typecast swapbytes
% methods in the datafun folder:
%
% cummax cummin cumprod cumsum diff histcounts islocalmax
% ismissing issorted maxk mink movmad movmax movmean movmedian
% movmin movprod movstd movsum movvar rmmissing rmoutliers
% sort sortrows standardizeMissing topkrows
%
% m-files: bounds corrcoef cov del2 fillmissing filloutliers
% gradient isoutlier issortedrows mean median mode normalize
% rescale smoothdata std var
% methods the 'double' class that are not yet implemented here:
%
% Possible 'double' functions to overload in the future (note that
% mod and rem are not the same as the ANSI fmod or remainder, but
% the built-in rem is almost the same as the ANSI fmod):
%
% mod rem unwrap sind asind cosd acosd tand
% atand secd asecd cscd acscd cotd acotd atan2d
%
% not needed:
%
% special functions: airy bernoulli besselh besseli besselj
% besselk bessely betainc betaincinv chebyshevT chebyshevU
% coshint cosint dawson dilog dirac ei ellipticCE ellipticCK
% ellipticCPi ellipticE ellipticF ellipticK ellipticNome
% ellipticPi erfcinv erfcx erfi erfinv fresnelc fresnels
% gammainc gammaincinv harmonic igamma jacobiP kummerU laguerreL
% legendreP logint pochhammer psi signIm sinhint sinint ssinint
% whittakerM whittakerW wrightOmega zeta triangularPulse
% rectangularPulse
%
% eigenvalue-related: charpoly euler gegenbauerC hermiteH jordan
% minpoly poly2sym polylog
%
% others: colon factor divisors superiorfloat
% methods in matfun not implemented here:
%
% balance cdf2rdf chol cholupdate condeig condest cond
% decomposition det expm funm gsvd hess inv ldl linsolve logm
% lscov lsqminnorm ltitr lu normest1 normest null ordeig ordqz
% ordschur orth pinv planerot polyeig qrdelete qrinsert qr
% qrupdate qz rank rcond rref rsf2csf schur sqrtm svd sylvester
% trace vecnorm
% methods in sparfun not implemented here:
%
% colperm delsq dissect eigs ichol ilu spalloc spaugment
% spconvert spdiags svds symbfact symmlq unmesh
%
% not needed: treeplot treelayout numgrid nested spparms
% methods in elmat not implemented here:
%
% accumarray blkdiag bsxfun circshift compan gallery
% hadamard hankel hilb inf invhilb ipermute isequaln nan ndgrid
% pascal permute repelem rot90 shiftdim toeplitz vander
% wilkinson
%
% not needed: linspace logspace ind2sub sub2ind meshgrid pi
% freqspace flintmax intmax intmin squeeze realmin realmax i j
% magic rosser
% methods for classes graph and digraph not yet implemented:
%
% addedge addnode bfsearch centrality conncomp dfsearch
% distances findedge findnode isisomorphic isomorphism maxflow
% nearest outedges rmedge rmnode shortestpath shortestpathtree
% simplify
% methods for class graph (not in digraph class) not yet implemented:
%
% bctree biconncomp minspantree neighbors
% methods for class digraph (not in graph class) not yet implemented:
%
% condensation inedges isdag predecessors successors toposort
% transclosure transreduction
% methods in LAGraph: (see the LAGraph/Source folder)
%---------------------------------------------------------------------
% operator overloading
%---------------------------------------------------------------------
C = and (A, B) ; % C = (A & B)
C = ctranspose (A) ; % C = A'
i = end (A, k, ndims) ; % for A (1:end,1:end)
C = eq (A, B) ; % C = (A == B)
C = ge (A, B) ; % C = (A >= B)
C = gt (A, B) ; % C = (A > B)
C = horzcat (varargin) ; % C = [A , B]
C = ldivide (A, B) ; % C = A .\ B
C = le (A, B) ; % C = (A <= B)
C = lt (A, B) ; % C = (A < B)
C = minus (A, B) ; % C = A - B
C = mldivide (A, B) ; % C = A \ B
C = mpower (A, B) ; % C = A^B
C = mrdivide (A, B) ; % C = A / B
C = mtimes (A, B) ; % C = A * B
C = ne (A, B) ; % C = (A ~= B)
C = not (G) ; % C = ~A
C = or (A, B) ; % C = (A | B)
C = plus (A, B) ; % C = A + B
C = power (A, B) ; % C = A .^ B
C = rdivide (A, B) ; % C = A ./ B
I = subsindex (A) ; % for C = X (A), using A as index I
C = subsasgn (C, S, A) ; % C (I,J) = A or C (M) = A
C = subsref (A, S) ; % C = A (I,J) or C = A (M)
C = times (A, B) ; % C = A .* B
C = transpose (G) ; % C = A.'
C = uminus (G) ; % C = -A
C = uplus (G) ; % C = +A
C = vertcat (varargin) ; % C = [A ; B]
%---------------------------------------------------------------------
% Methods that overload built-in functions:
%---------------------------------------------------------------------
% In the list below, G is always a GraphBLAS matrix. The inputs A
% and B can be a mix of GraphBLAS and/or built-in matrices, but at
% least one will be a GraphBLAS matrix because these are all methods
% that are overloaded from the built-in versions. If all inputs are
% built-in matrices, these methods are not used. The output matrix
% (C, L, or U) is always a GraphBLAS matrix. Lower case variables
% i, e, s, and n are scalars. Outputs p, parent, I, J, and X are
% built-in vectors. Graph is a built-in undirected graph. DiGraph
% is a built-in directed digraph.
C = abs (G) ;
C = acos (G) ;
C = acosh (G) ;
C = acot (G) ;
C = acoth (G) ;
C = acsc (G) ;
C = acsch (G) ;
C = all (G, option) ;
p = amd (G, varargin) ;
C = angle (G) ;
C = any (G, option) ;
C = asec (G) ;
C = asech (G) ;
C = asin (G) ;
C = asinh (G) ;
assert (G) ; % test assertion
C = atan (G) ;
C = atanh (G) ;
C = atan2 (A, B) ;
[lo, hi] = bandwidth (G, uplo) ;
C = bitand (A, B, assumedtype) ;
C = bitcmp (A, assumedtype) ;
C = bitget (A, B, assumedtype) ;
C = bitset (A, B, arg3, arg4) ;
C = bitshift (A, B, arg3) ;
C = bitor (A, B, assumedtype) ;
C = bitxor (A, B, assumedtype) ;
% C = cast (G, ...) built-in works as-is
C = cat (dim, varargin) ;
C = ceil (G) ;
[p, varargout] = colamd (G, varargin) ;
C = complex (A, B) ;
C = conj (G) ;
C = cos (G) ;
C = cosh (G) ;
C = cot (G) ;
C = coth (G) ;
C = csc (G) ;
C = csch (G) ;
C = cbrt (G) ;
C = diag (A, k) ;
DiGraph = digraph (G, option) ;
disp (A, level) ;
display (G) ;
[p, varargout] = dmperm (G) ;
C = double (G) ;
[V, varargout] = eig (G, varargin) ; % uses GrB matrices
C = eps (G) ;
C = erf (G) ;
C = erfc (G) ;
[parent, varargout] = etree (G, varargin) ;
C = exp (G) ;
C = expm1 (G) ;
[I,J,X] = find (G, k, search) ;
C = fix (G) ;
C = flip (G, dim) ;
C = floor (G) ;
c = fprintf (varargin) ;
C = full (A, type, identity) ;
C = gamma (G) ;
C = gammaln (G) ;
Graph = graph (G, varargin) ; % uses GrB matrices
C = hypot (A, B) ;
C = imag (G) ;
C = int8 (G) ;
C = int16 (G) ;
C = int32 (G) ;
C = int64 (G) ;
s = isa (G, type) ;
s = isbanded (G, lo, hi) ;
% s = iscolumn (G) built-in works as-is
s = isdiag (G) ;
s = isempty (G) ;
s = isequal (A, B) ;
C = isfinite (G) ;
s = isfloat (G) ;
s = ishermitian (G, option) ;
C = isinf (G) ;
s = isinteger (G) ;
s = islogical (G) ;
s = ismatrix (G) ;
C = isnan (G) ;
s = isnumeric (G) ;
s = isreal (G) ;
% s = isrow (G) built-in works as-is
s = isscalar (G) ;
s = issparse (G) ;
s = issymmetric (G, option) ;
s = istril (G) ;
s = istriu (G) ;
s = isvector (G) ;
C = kron (A, B) ;
n = length (G) ;
C = log (G) ;
C = log10 (G) ;
C = log1p (G) ;
[F, E] = log2 (G) ;
C = logical (G) ;
C = mat2cell (A, m, n) ;
C = max (A, B, option) ;
C = min (A, B, option) ;
e = nnz (G) ;
X = nonzeros (G) ;
s = norm (G, kind) ;
C = num2cell (A, dim) ;
s = numel (G) ;
e = nzmax (G) ;
C = pow2 (A, B) ;
C = prod (G, option) ;
C = real (G) ;
C = repmat (G, m, n) ;
C = reshape (G, m, n, by_col) ;
C = round (G) ;
C = sec (G) ;
C = sech (G) ;
C = sign (G) ;
C = sin (G) ;
C = single (G) ;
C = sinh (G) ;
[m, n, t] = size (G, dim) ;
C = sparse (G) ;
C = spfun (fun, G) ;
C = spones (G, type) ;
C = sprand (arg1, arg2, arg3) ;
C = sprandn (arg1, arg2, arg3) ;
C = sprandsym (arg1, arg2) ;
c = sprintf (varargin) ;
C = sqrt (G) ;
S = struct (G) ;
C = sum (G, option) ;
[p, varargout] = symamd (G, varargin) ;
p = symrcm (G) ;
C = tan (G) ;
C = tanh (G) ;
L = tril (G, k) ;
U = triu (G, k) ;
C = uint8 (G) ;
C = uint16 (G) ;
C = uint32 (G) ;
C = uint64 (G) ;
C = xor (A, B) ;
end
methods (Static)
%---------------------------------------------------------------------
% Static Methods:
%---------------------------------------------------------------------
% All of these are used as GrB.method (...), with the "GrB." prefix.
% The input matrices (A, B, C, M, ...) are of any kind (GraphBLAS,
% built-in sparse, or built-in full). The output matrix C is a
% GraphBLAS matrix.
% Some of the methods listed below are high-level graph algorithms that
% rely on GrB objects internally (bfs, dnn, ktruss, mis, pagerank, and
% tricount), for simplicity and readability. All of the other methods
% extract the opaque content of the GrB objects just once, operate on
% them, and then cast their results back into a GrB object just
% once. This makes for less-readable code, but it avoids the
% performance cost of accessing/modifying a object.
MATLAB_vs_GrB ;
C = apply (Cin, M, accum, op, A, desc) ;
C = apply2 (Cin, M, accum, op, A, B, desc) ;
[x,p] = argmin (A, dim) ;
[C,P] = argsort (A, dim, direction) ;
[x,p] = argmax (A, dim) ;
C = assign (Cin, M, accum, A, I, J, desc) ;
[v, parent] = bfs (A, s, varargin) ; % uses GrB matrices
binopinfo (op, type) ;
C = build (I, J, X, m, n, dup, type, desc) ;
b = burble (b) ;
C = cell2mat (A) ;
c = chunk (c) ;
clear ;
[C, I, J] = compact (A, id) ;
descriptorinfo (d) ;
C = deserialize (blob, mode, arg3) ; % arg3 for testing only
Y = dnn (W, bias, Y0) ; % uses GrB matrices
C = eadd (Cin, M, accum, op, A, B, desc) ;
C = empty (arg1, arg2) ;
C = emult (Cin, M, accum, op, A, B, desc) ;
x = entries (A, arg2, arg3) ;
C = expand (scalar, A, type) ;
C = extract (Cin, M, accum, A, I, J, desc) ;
[I, J, X] = extracttuples (A, desc) ;
C = eunion (Cin, M, accum, op, A, a, B, b, desc) ;
C = eye (m, n, type) ;
finalize ;
[f, s, iso] = format (arg) ;
C = incidence (A, varargin) ;
init ;
s = isbyrow (A) ;
s = isbycol (A) ;
s = isfull (A) ;
s = issigned (arg) ;
C = kronecker (Cin, M, accum, op, A, B, desc) ;
C = ktruss (A, k, check) ; % uses GrB matrices
L = laplacian (A, type, check) ;
C = load (filename) ;
iset = mis (A, check) ; % uses GrB matrices
monoidinfo (monoid, type) ;
C = mxm (Cin, M, accum, semiring, A, B, desc) ;
result = nonz (A, varargin) ;
s = normdiff (A, B, kind) ;
C = offdiag (A) ;
ctype = optype (a, b) ;
[r, stats] = pagerank (A, opts) ; % uses GrB matrices
C = prune (A, identity) ;
C = random (varargin) ;
C = reduce (cin, accum, monoid, A, desc) ;
filename_used = save (C, filename) ;
C = select (Cin, M, accum, selectop, A, b, desc) ;
selectopinfo (op, type) ;
semiringinfo (s, type) ;
blob = serialize (A, method, level) ;
C = speye (m, n, type) ;
C = subassign (Cin, M, accum, A, I, J, desc) ;
nthreads = threads (nthreads) ;
C = trans (Cin, M, accum, A, desc) ;
s = tricount (A, check, d) ; % uses GrB matrices
s = type (A) ;
unopinfo (op, type) ;
v = version ;
v = ver ;
C = vreduce (Cin, M, accum, monoid, A, desc) ;
t = timing (c) ; % timing for diagnositics only, requires -DGB_TIMING
% these were formerly overloaded methods, now Static methods
C = false (varargin) ;
C = true (varargin) ;
C = ones (varargin) ;
C = zeros (varargin) ;
end
end
|