well there are all sorts of tools to draw equations... I kind of remember something in Excel that even tried to estimate curves... the issue being it spouted out hugely complex equations
i think those that need to do this kind of thing (e.g. scientists trying to describe stuff) tend to be well versed in maths, so they know the kind of functions they want to use.... say like a sound designer knows how to combined oscillators/filters...
that said, there are some really powerful mathematical tools out there, like R so perhaps they can do stuff... but im not a mathematician... so dont know that 'landscape' at all.
that also might prove problematic, equations are not generally described in C++ code, so there is then some translation to be done... also if it an intensive function, on axoloti it going to need optimising and that is non-trivial.
another question to ask yourself is why?
obviously a sampled waveform is fairly accurate...
so the benefit of a function is to be able to 'mess' with it real time (thats why physical modelling is so useful , imo) ... but perhaps you can get some of those benefits by doing re-synthesis and then additive synthesis.
(you'll still be limited to how many sine waves you can generate, but its more manageable as a problem)
will it get you where you want, probably not but it might be a fun journey...
that said, seems like granular and fm is the fashionable thing at the moment, perhaps with this resurgence of digital, additive will make a come back.
btw: if you want to play with re-synthesis/additive l a bit, there are VSTs doing this, I quite enjoyed Icarus (tone2) ... but im sure there are lots of others.
also I learnt quite a bit, with an old windows app called AudioTerm... though its got quite a 'challenging' user interface