Can't convert Num to Float64

[nathanrboyer]

The result of substitute sometimes is type Float and sometimes is type Num. If the output is Num, then I cannot seem to use it generally. (These are the same variables from the problem in #533.)

dict = Dict(P_i => 60, P_o => 0, r_i => 6/2, r_o => 20/2)
typeof(substitute(a_solution, dict)) # BigFloat
typeof(substitute(r_i, dict)) # Num
r_range = substitute(r_i, dict) : 0.1 : substitute(r_o, dict) # Error

Errror Output:

r_range =
Failed to show value:

MethodError: no method matching div(::Symbolics.Num, ::Symbolics.Num, ::RoundingMode{:ToZero})

Closest candidates are:

div(::Real, ::Real, ::RoundingMode) at C:\Users\nboyer.AIP\.julia\juliaup\julia-1.7.2+0~x64\share\julia\base\div.jl:288

div(::Number, !Matched::Missing, ::RoundingMode) at C:\Users\nboyer.AIP\.julia\juliaup\julia-1.7.2+0~x64\share\julia\base\missing.jl:130

div(!Matched::MultivariatePolynomials.AbstractPolynomialLike, ::Number, ::Any...) at C:\Users\nboyer.AIP\.julia\packages\MultivariatePolynomials\JG6mC\src\division.jl:10

...

div(::Symbolics.Num, ::Symbolics.Num, ::RoundingMode{:ToZero})@div.jl:289
div(::Symbolics.Num, ::Symbolics.Num)@div.jl:37
length(::StepRange{Symbolics.Num, Symbolics.Num})@range.jl:693
axes(::StepRange{Symbolics.Num, Symbolics.Num})@range.jl:643
axes@abstractarray.jl:74[inlined]
var"#_latexarray#30"(::Symbol, ::Bool, ::Bool, ::Bool, ::Base.Pairs{Symbol, Symbol, Tuple{Symbol}, NamedTuple{(:env,), Tuple{Symbol}}}, ::typeof(Latexify._latexarray), ::StepRange{Symbolics.Num, Symbolics.Num})@latexarray.jl:20
var"#_latexraw#9"(::Base.Pairs{Symbol, Symbol, Tuple{Symbol}, NamedTuple{(:env,), Tuple{Symbol}}}, ::typeof(Latexify._latexraw), ::StepRange{Symbolics.Num, Symbolics.Num})@latexraw.jl:112
var"#process_latexify#115"(::Base.Pairs{Symbol, Symbol, Tuple{Symbol}, NamedTuple{(:env,), Tuple{Symbol}}}, ::typeof(Latexify.process_latexify), ::StepRange{Symbolics.Num, Symbolics.Num})@latexify_function.jl:25
#latexraw#[email protected]:58[inlined]
latexraw@latexraw.jl:58[inlined]
var"#_latexequation#60"(::Bool, ::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}, ::typeof(Latexify._latexequation), ::StepRange{Symbolics.Num, Symbolics.Num})@latexequation.jl:5
_latexequation(::StepRange{Symbolics.Num, Symbolics.Num})@latexequation.jl:5
var"#process_latexify#115"(::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}, ::typeof(Latexify.process_latexify), ::StepRange{Symbolics.Num, Symbolics.Num})@latexify_function.jl:25
process_latexify@latexify_function.jl:18[inlined]
var"#latexify#114"(::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}, ::typeof(Latexify.latexify), ::StepRange{Symbolics.Num, Symbolics.Num})@latexify_function.jl:3
latexify@latexify_function.jl:2[inlined]
show(::IOBuffer, ::MIME{Symbol("text/latex")}, ::StepRange{Symbolics.Num, Symbolics.Num})@latexify_recipes.jl:85
__binrepr@iobuffer.jl:0[inlined]
var"#repr#1"(::Nothing, ::typeof(repr), ::MIME{Symbol("text/latex")}, ::StepRange{Symbolics.Num, Symbolics.Num})@multimedia.jl:147
repr(::MIME{Symbol("text/latex")}, ::StepRange{Symbolics.Num, Symbolics.Num})@multimedia.jl:147
show_richest(::IOContext{IOBuffer}, ::Any)@PlutoRunner.jl:1014
var"#sprint_withreturned#54"(::IOContext{Base.DevNull}, ::Int64, ::typeof(Main.PlutoRunner.sprint_withreturned), ::Function, ::StepRange{Symbolics.Num, Symbolics.Num})@PlutoRunner.jl:954
format_output_default(::Any, ::Any)@PlutoRunner.jl:862
var"#format_output#42"(::IOContext{Base.DevNull}, ::typeof(Main.PlutoRunner.format_output), ::Any)@PlutoRunner.jl:879
formatted_result_of(::Base.UUID, ::Bool, ::Vector{String}, ::Nothing, ::Module)@PlutoRunner.jl:785
top-level scope@none:1

Attempt to Fix:

BigFloat(substitute(r_i, dict)) # Error

Error Output:

MethodError: no method matching BigFloat(::Symbolics.Num)

Closest candidates are:

BigFloat(::Real, !Matched::RoundingMode; precision) at C:\Users\nboyer.AIP\.julia\juliaup\julia-1.7.2+0~x64\share\julia\base\mpfr.jl:262

(::Type{T})(::Real, !Matched::RoundingMode) where T<:AbstractFloat at C:\Users\nboyer.AIP\.julia\juliaup\julia-1.7.2+0~x64\share\julia\base\rounding.jl:200

BigFloat(::Real, !Matched::Int64) at C:\Users\nboyer.AIP\.julia\juliaup\julia-1.7.2+0~x64\share\julia\base\deprecated.jl:70

...

top-level scope@[Local: 1](http://localhost:1234/edit?id=c3a57e20-9583-11ec-2279-b500f3a93791#)[inlined]

I am also wondering if my approach is just wrong. In general the steps I am taking are:

1. Define @variables x ...
2. x_solution = solve_for(exp, x)
3. x_solution_value = substitute(x_solution, dict_of_input_parameters)
4. Use x_solution_value with other numerical data to make plots.

Keeping track of whether I should be using x, x_solution, or x_solution_value has been tough since they all represent the same value. I'd rather just transform x at some point in the code I guess, but then I can't use Pluto.jl.

[nathanrboyer]

I found that you have to use the function Symbolics.value to convert Nums to usable Floats.

julia> typeof(Symbolics.value(substitute(r_i, dict)))
Float64

julia> r_range = Symbolics.value(substitute(r_i, dict)) : 0.1 : Symbolics.value(substitute(r_o, dict))
3.0:0.1:10.0

Couldn't the conversion be done automatically if there are no @variables left? The syntax is very verbose and not well documented.

help?> Symbolics.value
  No documentation found.

It is mentioned here on the website, but I can't make sense of it.

[andreasvarga]

I wonder what is the simplest way to recover a Float64 matrix from a constant Num matrix

Julia> S = Num.(zeros(2,2))
2×2 Matrix{Num}:
 0.0  0.0
 0.0  0.0

Substitution apparently is not working (there is no variable to substitute).

[Leebre]

@andreasvarga what I've done is to just create my own wrapper function:

function Float64(a::Num)
    b = Symbolics.value(a)
end

which then seems to work with:

A = Float64.(S)

but it would be nice if this was included in the Symbolics package by default.