How to use Base.rationalize on whole symbolic expression?

New to Julia
,
3

Chris’ answer implies that it should work.

But… I am not in any way an expert on this but Symbolics always fascinates me, so I tried to find a way to do this and I thought it should work somehow recursively. This is the solution I created. It’s not complete and generic enough to work on all expressions, but it can easily be expanded (so I hope).

And maybe someone is motivated to show us a more Symbolics-Way of doing it.

using Symbolics

sr = @rule ( +(~~x) ) => ~~x
mr = @rule ( *(~~x) ) => ~~x
pr = @rule ( ^(~x,~y) ) => [~x,~y]

function rationalize!(a)
	for index in eachindex(a)
		if typeof(a[index]) <: AbstractFloat || typeof(a[index]) <: ComplexF64
			a[index] = Base.rationalize(a[index])
		end
	end
	a
end

function decompose_compose(expr)
	if typeof(expr) <: SymbolicUtils.Sym || typeof(expr) <: Number
		return expr
	end
	if ! isnothing(sr(expr))
		subexpr = sr(expr)
		compose=sum
	elseif ! isnothing(mr(expr))
		subexpr = mr(expr)
		compose=prod
	elseif ! isnothing(pr(expr))
		subexpr = pr(expr)
		compose=^
	else
		display(expr)
		error("no rule defined")
	end
	subexpr = rationalize!(subexpr)
	subexpr = decompose_compose.(subexpr)
	(compose == ^) ? compose(subexpr...) : compose(subexpr)
end

@syms x

expr=(0.014378665475956662 - 0.0645731363698366im)*((72 + 132x + 2262(x^4) + 6743(x^5) - 754(x^2) - 234(x^6) - 1641(x^3) - 9180(x^7) - 5400(x^8))^-1) + (-0.0995313498923154 + 0.07922575111068406im)*x*((72 + 132x + 2262(x^4) + 6743(x^5) - 754(x^2) - 234(x^6) - 1641(x^3) - 9180(x^7) - 5400(x^8))^-1)

decompose_compose(expr)
1 Like