On Saturday, 7 October 2017 12:30:16 UTC-4, Robert Dodier wrote: > > Hi Francesco, thanks for your response. The Rubi implementation looks very > interesting, also the paper about matching algorithms. I will definitely > take a close look at that. >
Yes, the MatchPy authors have published a paper on Arxiv. What is your feeling about how the Rubi implementation turned out overall? > Did you run into any roadblocks on the way? Do you think the general > approach can be extended in a straightforward way to all of Rubi? If not, > what are the large-scale changes that you foresee? > Technically we have imported all of RUBI. Unfortunately there's still some work to do, as apparently some rules aren't properly matched. This is probably due to errors in the utility functions implementations and possibly to other factors leading to wrong matchings or processing of data. We haven't had any thorough investigation about whether MatchPy is fully compatible with RUBI, the main possible problem being expressions matched by more than one rules (in MatchPy the matching rule could be different than in Mathematica). > Is it known how the time required to process an expression with this > implementation varies as a function of the number of rules? Is it known > whether the time is linear in the number of rules, or perhaps logarithmic > or something else? > Using MatchPy seems to be slower than using the SymPy integrators. Our major problem now is the loading time for the rules (it's around one hour on my computer). We have talked to the MatchPy developers and they have agreed to add a code generator to compile the rules into a decision tree. Unfortunately the size of the generated file is huge and more work is probably needed on the optimization of the generated code, but I feel like that's the right way to go. Ideally, if MatchPy were able to generate an optimized decision-tree, then it would probably make the integration algorithm faster. I believe that's the way to proceed. -- 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 https://groups.google.com/group/sympy. To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/22cb23a3-539e-4e54-b60b-e61269f0308a%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
