It may be late to reply on this topic. but i just want to add that all functionality of sympy works with evaluate=True on add/mul . If you are dealing with evaluate=False ,there are lots of chances that you will face max recursion errors or some unwanted results . I have faced similar issues while working on my project ( http://tuttr.com) and ended up writing overrides for these classes. if you want I can share the possible issues/constraints in this regard
--- Thanks, Ghanshyam Agrawal Contact No : +919717167192 Core Team Member :tutrr.com On Thu, Mar 14, 2013 at 4:46 AM, Aaron Meurer <asmeu...@gmail.com> wrote: > The idea from https://code.google.com/p/sympy/issues/detail?id=3613 > might also interest you. It's not simple to implement, but I think it > wouldn't be too hard. > > Aaron Meurer > > On Wed, Mar 13, 2013 at 1:40 PM, Duane Nykamp <dqnyk...@comcast.net> > wrote: > > It was an interesting exercise trying to get an UnevaluatedAdd/Mul/Pow > > implemented. It ended up being trickier because the arguments weren't > sympy > > objects, so the latex and str print code didn't work outright. What > ended > > up defeating me, though, was the term ordering. Since I would use the > > unevaluated expressions for intermediate steps in a solution, the folks > who > > would benefit from seeing the intermediate steps would presumably be > those > > for whom term reordering would confuse them. The term reordering ended > up > > getting deeper into the fundamentals of sympy than I could understand. > > > > I recognize that this wouldn't be a standard use of sympy, so not a high > > priority. But, if someday, you do get around to rework the > evaluate=False > > functionality, that would be great. > > > > Duane > > > > > > > > On Tuesday, March 12, 2013 12:30:59 AM UTC-5, Aaron Meurer wrote: > >> > >> On Mar 11, 2013, at 10:58 PM, Duane Nykamp <dqny...@comcast.net> wrote: > >> > >> Getting closer. With > >> > >> class UnevaluatedAdd(Expr): > >> def _sympystr(self, printer): > >> return "%s + %s" % (printer._print(self.args[0]), > >> printer._print(self.args[1])) > >> > >> parse_expr("x+1").replace(Add,UnevaluatedAdd).subs(x,-1) > >> > >> returns 1+-1, for obvious reasons. So I made it even uglier > >> > >> class UnevaluatedAdd(Expr): > >> def _sympystr(self, printer): > >> if(self.args[1] < 0): > >> return "%s %s" % (printer._print(self.args[0]), > >> printer._print(self.args[1])) > >> else: > >> return "%s + %s" % (printer._print(self.args[0]), > >> printer._print(self.args[1])) > >> > >> > >> Add takes an arbitrary number of arguments. So you'll need to be a > little > >> more general than that. You can probably just set _sympystr equal to > >> StrPrinter._print_Add (modulo some argument shuffling). Or if you find > it > >> needs to be modified just copy the source of StrPrinter._print_Add and > >> modify it (it's BSD licensed for a reason). > >> > >> > >> I don't know if this is a recommended way to do it, but it seems to > work. > >> > >> Interestingly, I noticed with playing around that > >> Add(x,1,evaluate=False).subs(x,-1) > >> returns 0. I don't know if that is intended. > >> > >> > >> Yes, evaluate=False is somewhat of a hack. Many operations will undo it, > >> because they rebuild expressions. That's actually what the other > discussion > >> going on in this thread is all about: a better replacement for > >> evaluate=False. > >> > >> Aaron Meurer > >> > >> > >> Duane > >> > >> > >> On Monday, March 11, 2013 11:25:31 PM UTC-5, Aaron Meurer wrote: > >>> > >>> On Mon, Mar 11, 2013 at 9:38 PM, Duane Nykamp <dqny...@comcast.net> > >>> wrote: > >>> > Thanks for your responses. > >>> > > >>> > I was pretty sure parse_expr was secure, given how it tokenizes > before > >>> > it > >>> > uses an eval. > >>> > > >>> > The UnevaluatedLog example does exactly what I want for log. > However, > >>> > I > >>> > cannot figure out how to extend it to an Add or Mul that I get from a > >>> > parsed > >>> > input. In many of the examples, one calculates a function, then > >>> > evaluates > >>> > it at particular value. For example, if f(x)=3+x, then I'd like to > >>> > show > >>> > f(5)=3+5 = 8 and f(-5) = 3 - 5 = -2 (where, in the second case, it > does > >>> > some > >>> > simplification to avoid 3+-5). Given your evaluate hint, I could > do > >>> > this > >>> > with Add(-3,-2, evaluate=False). But, there doesn't seem to be away > >>> > to do > >>> > this with the parser. Unfortunately, I can't even do > >>> > parse_expr("Add(-3,-2, > >>> > evaluate=False)"), as parse_expr doesn't allow keywords. If I > >>> > understand > >>> > the conversation right, this isn't possible with the current version. > >>> > >>> That bug has been fixed in the git version. > >>> > >>> I guess what you also could do for that case is create an > >>> UnevaluatedAdd (in this case, you should just subclass Expr rather > >>> than Function), and replace the Add with that. Then replace the x > >>> with the number in that. The logic for this might be a little more > >>> complicated, but it's doable. > >>> > >>> Aaron Meurer > >>> > >>> > > >>> > Thanks for the tips. > >>> > > >>> > > >>> > Duane > >>> > > >>> > > >>> > -- > >>> > 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 sympy+un...@googlegroups.com. > >>> > To post to this group, send email to sy...@googlegroups.com. > >>> > Visit this group at http://groups.google.com/group/sympy?hl=en. > >>> > 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 sympy+un...@googlegroups.com. > >> To post to this group, send email to sy...@googlegroups.com. > >> Visit this group at http://groups.google.com/group/sympy?hl=en. > >> 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 sympy+unsubscr...@googlegroups.com. > > To post to this group, send email to sympy@googlegroups.com. > > Visit this group at http://groups.google.com/group/sympy?hl=en. > > 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 sympy+unsubscr...@googlegroups.com. > To post to this group, send email to sympy@googlegroups.com. > Visit this group at http://groups.google.com/group/sympy?hl=en. > 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 sympy+unsubscr...@googlegroups.com. To post to this group, send email to sympy@googlegroups.com. Visit this group at http://groups.google.com/group/sympy?hl=en-US. For more options, visit https://groups.google.com/groups/opt_out.
