On Monday, October 21, 2013 9:24:15 PM UTC+2, Aaron Meurer wrote: > > > I don't know. You'll have to try it. Also try using xreplace and > sympy.core.rules.Transform. I'm becoming convinced that those two make > complex replacement more elegant and readable than any other combination of > functions currently in SymPy. > > Aaron Meurer >
Well, the problem with that is that tensor expressions are somewhat complex, for example consider: pattern: A(m1, m2)*B(-m2) on the expr: C(d0)*A(d1, d2)*D(d3)*B(-d2)*A(d4, d5)*B(-d4) supposing that A has no symmetries on its indices, and that only components have to be matched exactly, the pattern should match A(d1, d2)*B(-d2) only. Another problem is that indices are not associated with components anymore when they're inside a tensor expression. They are not part of the component args tree. I don't know how to match A(d1, d2) with a tree-search, as its pattern in the args tree won't be found anymore in a complex tensor expression, even if it exists. I'm more and more convinced that the tensor module needs a pattern matching and replacement algorithm just for itself. -- 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.
