- GiNaC-list - ginac.de

Re: [GiNaC-list] comparing symbols
by Chris Bouchard 25 Jul '08

25 Jul '08

Hi, >I don't think this situation is possible (modulo compiler bugs and >things like that). > >$ cat test.cc >#include <ginac/ginac.h> >#include <iostream> >using namespace GiNaC; >using namespace std; > >int main(int argc, char** argv) >{ > lst x; > x = ex(0), ex(1); > if (is_a<symbol>(x.op(0))) { > cerr << "Huh, 0 is a symbol?" << endl; > return 1; > } > return 0; >} >$ g++ test.cc -lginac >$ ./a.out >$ echo $? >0 Hmmm... that seems to make sense. For sake of clarity I've pasted code fragment below followed by associated output. Code fragment ------------- ... namespace xloops { static ex r_sum(const ex& t, const lst& b, const lst& x, const ex& eps, int order) { // class of arguments cout << "t is class mul?: " << is_a<mul>(t) << "\n"; cout << "b is class list?: " << is_a<lst>(b)<< "\n"; cout << "b.op(0) is class add?: " << is_a<add>(b.op(0)) << "\n"; cout << "b.op(1) is class numeric?: " << is_a<numeric>(b.op(1)) << "\n"; cout << "x is class list?: " << is_a<lst>(x) << "\n"; cout << "x.op(0) is class symbol?: " << is_a<symbol>(x.op(0)) << "\n"; cout << "x.op(0) is class mul?: " << is_a<mul>(x.op(0)) << "\n"; cout << "x.op(0) is class numeric?: " << is_a<numeric>(x.op(0)) << "\n"; cout << "x.op(0) is class add?: " << is_a<add>(x.op(0)) << "\n"; cout << "x.op(1) is class symbol?: " << is_a<symbol>(x.op(1)) << "\n"; cout << "x.op(1) is class mul?: " << is_a<mul>(x.op(1)) << "\n"; cout << "x.op(1) is class numeric?: " << is_a<numeric>(x.op(1)) << "\n"; cout << "x.op(1) is class add?: " << is_a<add>(x.op(1)) << "\n"; cout << "eps is class symbol?: " << is_a<symbol>(eps) << "\n"; // value of arguments cout << "t = " << t << "\n"; cout << "b = " << b << "\n"; cout << "x = " << x << "\n"; cout << "eps = " << eps << "\n"; ... } Output ------ t is class mul?: 1 b is class list?: 1 b.op(0) is class add?: 1 b.op(1) is class numeric?: 1 x is class list?: 1 x.op(0) is class symbol?: 1 x.op(0) is class mul?: 0 x.op(0) is class numeric?: 0 x.op(0) is class add?: 0 x.op(1) is class symbol?: 1 x.op(1) is class mul?: 0 x.op(1) is class numeric?: 0 x.op(1) is class add?: 0 eps is class symbol?: 1 t = -eps b = {-1/2+eps,1} x = {0,1} eps = eps Is there any other information I could provide to shed light on this? Sincerely, Chris Bouchard

2 1

Re: [GiNaC-list] comparing symbols
by Chris Bouchard 24 Jul '08

24 Jul '08

Hello, >Is it this what you want: > > symbol x; > ex e1 = beta(x, lst(0,1)); > ex e2 = beta(x, lst(4,1)); > > if ( e1 == beta(x, lst(0,1)) ) cout << "e1 BINGO" << endl; > if ( e2 == beta(x, lst(0,1)) ) cout << "e2 BINGO" << endl; > >Otherwise I do not completely understand your problem yet. Could you >clarify it more, then? The x-loops function I'm working with is r_sum. It's definition is (from r.cpp): static ex r_sum(const ex& t, const lst& b, const lst& x, const ex& eps, int order) and I'm trying to flag the occurrence of a certain set of argument values. The boolean operation "is_a<symbol>(x.op(i));" returns 1 for i=0,1 so I think I'm working with a list of symbols, x. Additionally, "cout << x;" yields {0,1} making me think that the symbols have been assigned numerical value. >That is what confuses me. One cannot assign numerical values to symbols >in GiNaC. One can replace symbols by numbers, of course. Perhaps I'm not correctly using/interpreting is_a<class>(ex) and/or cout<<x? Sincerely, Chris Bouchard

2 1

comparing symbols
by Chris Bouchard 24 Jul '08

24 Jul '08

GiNaC discussion list, I'd like to discern when arguments of a pre-defined function (from another library, xloops) attain a certain value. One of the arguments is type: lst x(symbol("a"), symbol("b")); cout << x yields {0,1} and these are the values I'm looking for. I'm unable to define a relational that accurately tests the values attained by a & b. I tried: if (a==0 && ...) if (a.is_equal(0) && ...) if (a.is_zero() && ...) if (match(a,0,lst placeholder) && ...) with no success. "Probably because a is type symbol", I reasoned. So I defined lst y(symbol("c"), symbol("d")); and used in-place modification y[0] = 0; y[1] = 1; to set symbols c and d to 0 and 1, resp. I then tried the conditionals if (x==y && ...) if (x.is_equal(y) && ...) if (match(x,y,lst placeholder) && ...) with no success. How does one set up a conditional for symbols that have been assigned numeric values? Any ideas? Thanks in advance! Sincerely, Chris Bouchard

2 1

ginac-config woe
by naruto canada 23 Jul '08

23 Jul '08

hi Is it possible to resume ginac-config support? I can't get xloops to "configure". I've even tried the cvs version of xloops hoping that it has "kept up" with ginac, but nope. my pkg-config don't support --cppflags either. (which xloops requires) I am trying older versions of GiNac now.

2 2

Re: [GiNaC-list] modify source, run make, then ...
by Chris Bouchard 11 Jul '08

11 Jul '08

Hi, >> If I modify a .cpp source file (say to change a thrown >> error message to pinpoint the location in the code where >> the error is being generated) and then run 'make' within >> the subdirectory containing the .cpp source file to >> generate an updated .o file, do I have to do anything >> else to incorporate my modified error message? > Could you please be more specific? What is exactly > that .cpp file? Is it a part of your program, or GiNaC (or > any other shared library)? The file, r.cpp, is part of another shared library, xloops. >> I think not > It depends. (Sorry, I need to know more details to give a > useful answer). By modifying r.cpp, running 'make' and then running 'make install' the changes made to r.cpp were incorporated (the error message thrown was the modified message). After stumbling upon the solution it appears my original question should have been posted to a beginner's gcc msg board and not GiNaC. Sorry for the clutter. Sincerely, Chris Bouchard

2 1

modify source, run make, then ...
by Chris Bouchard 11 Jul '08

11 Jul '08

GiNaC discussion list, If I modify a .cpp source file (say to change a thrown error message to pinpoint the location in the code where the error is being generated) and then run 'make' within the subdirectory containing the .cpp source file to generate an updated .o file, do I have to do anything else to incorporate my modified error message? I think not but am not getting the updated error message after doing this and I can't find any other instances of the error in the source files. Thanks for your help! Chris Bouchard

2 1

Re: [GiNaC-list] modify source, run make, then ...
by Chris Bouchard 11 Jul '08

11 Jul '08

'make install'... sorry ---- Original message ---- >Date: Thu, 10 Jul 2008 20:39:40 -0500 (CDT) >From: Chris Bouchard <cbouchrd(a)uiuc.edu> >Subject: [GiNaC-list] modify source, run make, then ... >To: ginac-list(a)ginac.de > >GiNaC discussion list, > >If I modify a .cpp source file (say to change a thrown error message to pinpoint the location in the code where the error is being generated) and then run 'make' within the subdirectory containing the .cpp source file to generate an updated .o file, do I have to do anything else to incorporate my modified error message? > >I think not but am not getting the updated error message after doing this and I can't find any other instances of the error in the source files. > >Thanks for your help! >Chris Bouchard >_______________________________________________ >GiNaC-list mailing list >GiNaC-list(a)ginac.de >https://www.cebix.net/mailman/listinfo/ginac-list

1 0

Algebraic sqrt()
by Timo Schlüßler 16 Jun '08

16 Jun '08

Hello, is there a way to compute a sqrt() algebraic? E.g. that sqrt(2) * sqrt(2) gives exact 2 and not 2.0000001 or whatever else. Or that sqrt(45) * sqrt(3) gives sqrt(135) and not the rounded value. thanks in advance, timo

2 2

Re: [GiNaC-list] advice on a GiNaC based project.
by Bhalchandra Thatte 23 May '08

23 May '08

Hello Alex, Thanks for the script. I will take a look at it soon. At some point the graph theoretic part of my project will involve graph isomorphism, and generating all unlabeled graphs from some classes of graphs, generating random objects from various classes, ... and computing various polynomial invariants. Also, at the moment all I know is C (and some C++), so I will probably go ahead and give a try to C+++GiNaC. But I am definitely interested in looking at your script since there is also another open source project in python (called SAGE - software for algebra and geometry experimentation), and some of what I want to do may be possible in SAGE. Thanks to all of you who responded promptly. Regards, Bhalchandra Thatte Alfréd Rényi Institute of Mathematics Budapest ----- Original Message ---- From: alexander baker <baker.alexander(a)gmail.com> To: GiNaC discussion list <ginac-list(a)ginac.de> Cc: Bhalchandra Thatte <bdthatte(a)yahoo.com> Sent: Friday, May 23, 2008 11:12:35 AM Subject: Re: [GiNaC-list] advice on a GiNaC based project. I agree with Alexei, here is a similar example that I put together some time ago using Python, perhaps a high level script would be a better tool for the job! Enjoy. Alex http://www.alexfb.com/cgi-bin/twiki/view/PtPhysics/Adjacency-Matrix-EigenVe… 2008/5/23 Alexei Sheplyakov <varg(a)theor.jinr.ru>: Hello! On Fri, May 23, 2008 at 12:47:55AM -0700, Bhalchandra Thatte wrote: > I wish to write a few programs that do something like this: > take a graph or a tree > each vertex or edge of the graph has a variable associated with it. > walk on the graph or find certain subgraphs, and compute certain > multivariable polynomials in terms of the variables associated > with vertices or edges. I think boost.graph (for graph manipulations) and GiNaC (for symbolic computation) is a reasonable choice. > later possibly do numerical optimization of the polynomials using > an optimization library. Be forewarned: GiNaC does not provide any kind of numerical optimization. > I thought GiNaC would be ideal for this type of applications because > C++ will give me complete control over the data structures and > the efficiency. I suggest to think twice if you really need a low level programming language (such as C++) at all. Best regards, Alexei -- All science is either physics or stamp collecting. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) iQIVAwUBSDaEzyXMWnuYxRVqAQL7XxAAok3ywm5yWWvGwSOI/XMCIE3Y7j23Q9td DH8TEX5hIt1kzASA33iaDsWVkvBPRyWPGTF6AJLOvnyOLX1LHcB0Q8F+ojfKmVIw o2raKOltLYO4oDB8K/edUtfLlUrLv7r688t6qdBBzDmxfbMQ/4/W32hioMvzr9yh isWgXIjBV/FV86D2hVF/q7PvbloulHVyxH3TwINkmyItaygfToaS5OyDJVdBeKbv SJ0gbqEKNkyWZ/4ZIZeJ0ckkOhP/fIDGO04YqVKrLUwy7Q1UZuuti3i92Vosr+Yb m7x7v5yZk0J/3W/S1x+ZOl/1tt8UVacAc9maIlQXM9kH0B8wG7vQCHHNKZ6XzHtN krUxNxw7ciWz4WY0tXg2vzePzMCzVN3nfaWLJ9Fznt3VPPDYuFfWQmsS7zPXXOS9 nHkmM9pSZ4wR8nLDZAmIH36J9QMmImx9d/T91na3tAhKm1/4Q/3uO1SLHahIgCcj dYkA8F1GWR082XlBqnWoISwlJjv1ZqNFsjISllixYk24+MvtODHohDEJnEUsAoWg jdol4XxpZS3UAGa+Cy7xpzC1iOpksEmG/3zQaI9QUI8GXdnXfM1noeAGDnH/JeF8 aEsu5Z6Wva3iZYBFj3SM/mnFqleumt6Jf4YBMDYiohQG5RoQCtQ4k/DjSED1JZiP gx4tfNdbA9A= =HgPN -----END PGP SIGNATURE----- _______________________________________________ GiNaC-list mailing list GiNaC-list(a)ginac.de https://www.cebix.net/mailman/listinfo/ginac-list -- Mobile: 07788 872118 Blog: www.alexfb.com -- All science is either physics or stamp collecting.

2 1

advice on a GiNaC based project.
by Bhalchandra Thatte 23 May '08

23 May '08

Hi Everybody, I have just started using GiNaC (actually yet to be able to compile my first program), and I am also not an expert on C++. So I thought I would ask for a little advice here. I wish to write a few programs that do something like this: take a graph or a tree each vertex or edge of the graph has a variable associated with it. walk on the graph or find certain subgraphs, and compute certain multivariable polynomials in terms of the variables associated with vertices or edges. later possibly do numerical optimization of the polynomials using an optimization library. I thought GiNaC would be ideal for this type of applications because C++ will give me complete control over the data structures and the efficiency. But could someone comment if GiNaC is not the right choice? Thank you, Bhalchandra Thatte Alfréd Rényi Institute of Mathematics Budapest

4 3