Is there a way to parse coefficients from generic function like f(x) = 3x^2 - 5x + 4
a = 3, b = -5, c = 5
But i couldn't find a way other than string.
#Dynamicly parse coefficients
steel stream
odd maple
Using https://docs.julialang.org/en/v1/manual/metaprogramming/#Program-representation and Meta.parse, you can probably get something that's a little better than pure strings to get the answer
expr = "f(x) = 3x^2 - 5x + 4"
ex1 = Meta.parse(expr)
dump(ex1)
And you get a syntax tree that you can visit to get the answer
Maybe it's not simpler than string though
julia> ex1.args[2].args[2].args[2].args[2]
:(3 * x ^ 2)
going down in the tree you can get to individual elements
steel stream
Oh that's literally what i need
odd maple
julia> dump( ex1.args[2].args[2].args[2].args[2])
Expr
head: Symbol call
args: Array{Any}((3,))
1: Symbol *
2: Int64 3
3: Expr
head: Symbol call
args: Array{Any}((3,))
1: Symbol ^
2: Symbol x
3: Int64 2
for a single element you can extract the value here
But you need to be careful on how you actually process the tree
steel stream
Oh now i realized it's kind of hard
if a is negative etc
using Plots
f(x) = 3x^2 + 4x - 5
a = 3
b = 4
c = -5
vertex_x = -b / (2a)
vertex_y = f(vertex_x)
x_vals = -20:1:20
y_vals = f.(x_vals)
points = [(x_vals[1], y_vals[1], "f(-20) = $(y_vals[1])", :green), (x_vals[end], y_vals[end], "f(20) = $(y_vals[end])", :orange), (vertex_x, vertex_y, "f($vertex_x) = $vertex_y", :blue)]
p = x_vals |> x -> plot(x, y_vals, xlabel="x", ylabel="f(x)", title="f(x)", label="f(x)", size=(800, 600), legend=:topleft)
for (x, y, label, color) in points
p |> p -> scatter!(p, [x], [y], markersize=5, markercolor=color, label=label)
end
display(p)
languid path
It'd be a lot easier to go the other way around, coefficients to the function
But either could work
steel stream
Yeah i figured that out too but i need to change 3 parameters everytime instead of single generic function
I came from C# so it affects on me too probably
Like this it works fine
I just don't like it
odd maple
I think with a careful tree algo you could make this work. Like 2x - 3x^2 gives you a expressions that's -, 2x, 3x^2, parse the right children first (last element), then when it's done apply the - to the value you've just obtained and then go on with the left children ...
But yeah, there's a lot of helper functions to represent polynomials in vector form and evaluate them otherwise
steel stream
using Plots
(a, b, c) = (-5, 10, 9)
f(x) = a*x^2 + b*x + c
vertex_x = -b / (2a)
vertex_y = f(vertex_x)
x_vals = -20:1:20
y_vals = f.(x_vals)
points = [(x_vals[1], y_vals[1], "f(-20) = $(y_vals[1])", :green), (x_vals[end], y_vals[end], "f(20) = $(y_vals[end])", :orange), (vertex_x, vertex_y, "f($vertex_x) = $vertex_y", :blue)]
p = x_vals |> x -> plot(x, y_vals, xlabel="x", ylabel="f(x)", title="f(x)", label="f(x)", size=(800, 600), legend=:topleft)
for (x, y, label, color) in points
p |> p -> scatter!(p, [x], [y], markersize=5, markercolor=color, label=label)
end
display(p)
I'll keep it like this
Thanks btw for quick response
I'm still learning so i don't bother with it (im kinda lazy)
odd maple
(I would have in mind something like that usually https://juliamath.github.io/Polynomials.jl/stable/#Construction-and-Evaluation)
Having a polynomial as a vector [c,b,a] and a package that can evaluate it (it's a very common representation for polynomials, though maybe more complex to do that in that case)
