Thanks Aaron, Harsh and Matthew. Can one treat SetExprs as ordinary Symbols, and are they differentiable in particular? My problem arose when I tried to take a derivative of an expression, but Sympy couldn't deduce the sign of the expression log((N-1)+exp(beta*epsilon))/N) because of the problem I mentioned above.
thanks again, Pat. On Sat, Feb 8, 2014 at 3:09 AM, Matthew Rocklin <[email protected]> wrote: > Recent work on sets provides mechanisms to handle this problem > > In [2]: from sympy.sets.setexpr import SetExpr > > In [3]: epsilon = SetExpr(Interval(0, oo)) > > In [4]: beta = SetExpr(Interval(0, oo)) > > In [5]: exp(epsilon * beta) - 1 > Out[5]: [0, ∞) > > This is from my setexpr development branch > https://github.com/sympy/sympy/pull/2721 . It is not yet in master > > Harsh has also done some work on this issue. In particular > https://github.com/sympy/sympy/pull/2682 > > > On Fri, Feb 7, 2014 at 7:50 PM, Harsh Gupta <[email protected]>wrote: > >> Thanks for posting this on the mailing list. >> >> >> There's also a bug in the new `solve_univariate_inequalities` >> >> ``` >> >> In [2]: x = Symbol('x', positive=True) >> >> >> In [3]: y = Symbol('y') >> >> >> In [4]: solve_univariate_inequality(exp(x) > 1, x) >> >> Out[4]: False >> >> >> In [5]: solve_univariate_inequality(exp(y) > 1, y) >> >> Out[5]: y > 0 >> ``` >> >> This was caused by me by not taking care of the assumptions on the >> variables. I'll fix it asap. >> >> On 8 February 2014 07:28, Aaron Meurer <[email protected]> wrote: >> > No, unfortunately, non-trivial relational assumptions like this are >> > largely not implemented. As you may or may not know, the assumptions >> > system in SymPy is currently in a bit of a mess. There are two >> > assumptions systems, the "old" one, which is the one you used, and the >> > "new" one, which uses ask() and Q. >> > >> > Neither can handle this, or really any kind of non-trivial >> > Add.is_positive type assumption. It's not even clear to me at the >> > moment how one would go about implementing such things. >> > >> > Aaron Meurer >> > >> > On Fri, Feb 7, 2014 at 11:47 AM, Patrick O'Neill <[email protected]> wrote: >> >> Howdy folks, >> >> >> >> I'm a very new to sympy, and am stumped by the following error. >> Considering >> >> the following code snippet: >> >> >> >> from sympy import * >> >> print sympy.__version__ >> >> #'0.7.4.1-git' >> >> epsilon,beta = var("epsilon,beta",positive=True) >> >> (exp(epsilon*beta) - 1).is_positive # is None >> >> >> >> Sympy knows that beta and epsilon are positive, but seems not to know >> that >> >> the exponential function takes positive quantities to quantities >> greater >> >> than one. Is there a workaround for this issue, or is there something >> >> fundamental to sympy's internals that makes this sort of deduction >> >> impossible? >> >> >> >> I checked the mailing list, stack overflow and the issue tracker, and >> didn't >> >> see anything pertaining to this question, but I admit I might not know >> what >> >> I'm looking for. >> >> >> >> Thanks in advance! >> >> >> >> Cheers, >> >> Pat. >> >> >> >> >> >> On Fri, Feb 7, 2014 at 12:45 PM, Patrick O'Neill <[email protected]> >> wrote: >> >>> >> >>> Howdy folks, >> >>> >> >>> I'm a very new to sympy, and am stumped by the following error. >> >>> Considering the following code snippet: >> >>> >> >>> from sympy import * >> >>> print sympy.__version__ >> >>> #'0.7.4.1-git' >> >>> epsilon,beta = var("epsilon,beta",positive=True) >> >>> (exp(epsilon*beta) - 1).is_positive # is None >> >>> >> >>> Sympy knows that beta and epsilon are positive, but seems not to know >> that >> >>> the exponential function takes positive quantities to quantities >> greater >> >>> than one. Is there a workaround for this issue, or is there something >> >>> fundamental to sympy's internals that makes this sort of deduction >> >>> impossible? >> >>> >> >>> I checked the mailing list, stack overflow and the issue tracker, and >> >>> didn't see anything pertaining to this question, but I admit I might >> not >> >>> know what I'm looking for. >> >>> >> >>> Thanks in advance! >> >>> >> >>> Cheers, >> >>> Pat. >> >> >> >> >> >> -- >> >> You received this message because you are subscribed to the Google >> Groups >> >> "sympy" group. >> >> To unsubscribe from this group and stop receiving emails from it, send >> an >> >> email to [email protected]. >> >> To post to this group, send email to [email protected]. >> >> Visit this group at http://groups.google.com/group/sympy. >> >> For more options, visit https://groups.google.com/groups/opt_out. >> > >> > -- >> > You received this message because you are subscribed to the Google >> Groups "sympy" group. >> > To unsubscribe from this group and stop receiving emails from it, send >> an email to [email protected]. >> > To post to this group, send email to [email protected]. >> > Visit this group at http://groups.google.com/group/sympy. >> > For more options, visit https://groups.google.com/groups/opt_out. >> >> >> >> -- >> Harsh >> >> -- >> You received this message because you are subscribed to the Google Groups >> "sympy" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> To post to this group, send email to [email protected]. >> Visit this group at http://groups.google.com/group/sympy. >> For more options, visit https://groups.google.com/groups/opt_out. >> > > -- > You received this message because you are subscribed to the Google Groups > "sympy" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at http://groups.google.com/group/sympy. > For more options, visit https://groups.google.com/groups/opt_out. > -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/sympy. For more options, visit https://groups.google.com/groups/opt_out.
