Allow user-defined printing

[MilesCranmer]

This creates keyword arguments for string_tree (and print_tree) which can take a custom function for printing variables and constants in a tree. e.g.,

    operators = OperatorEnum(;
        binary_operators=(+, *, /, -), unary_operators=(cos, sin)
    )
    x1, x2, x3 = [Node(Float64; feature=i) for i in 1:3]
    tree = x1 * x1 + π
    f_constant(val::Float64, args...) = string(round(val; digits=4))
    @test string_tree(tree, operators; f_constant=f_constant) == "((x1 * x1) + 3.1416)"

[coveralls]

Pull Request Test Coverage Report for Build 5306782871

• 28 of 33 (84.85%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.01%) to 92.421%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/Equation.jl	25	27	92.59%
src/Utils.jl	3	6	50.0%

Totals
Change from base Build 5095018727:	-0.01%
Covered Lines:	878
Relevant Lines:	950

---

💛 - Coveralls

[github-actions]

Benchmark Results

	master	e5494ea...	t[master]/t[e5494ea...]
eval/ComplexF32/evaluation	13.5 ± 0.88 ms	13.4 ± 0.98 ms	1.01
eval/ComplexF64/evaluation	16.8 ± 1.2 ms	16.8 ± 1.6 ms	1
eval/Float32/derivative	25 ± 3.8 ms	25.3 ± 3.2 ms	0.989
eval/Float32/derivative_turbo	25.4 ± 3.8 ms	25.2 ± 3.1 ms	1.01
eval/Float32/evaluation	4.68 ± 0.4 ms	4.51 ± 0.41 ms	1.04
eval/Float32/evaluation_turbo	1.48 ± 0.15 ms	1.42 ± 0.14 ms	1.04
eval/Float64/derivative	0.0346 ± 0.0014 s	0.0336 ± 0.0012 s	1.03
eval/Float64/derivative_turbo	0.0383 ± 0.0016 s	0.0368 ± 0.0014 s	1.04
eval/Float64/evaluation	5.04 ± 0.45 ms	4.98 ± 0.44 ms	1.01
eval/Float64/evaluation_turbo	2.71 ± 0.22 ms	2.67 ± 0.2 ms	1.02
time_to_load	2.3 ± 0.014 s	2.26 ± 0.033 s	1.01
utils/combine_operators	0.0691 ± 0.015 ms	0.067 ± 0.013 ms	1.03
utils/convert/break_sharing	0.0644 ± 0.017 ms	0.0656 ± 0.02 ms	0.982
utils/convert/preserve_sharing	0.27 ± 0.053 ms	0.269 ± 0.061 ms	1
utils/copy/break_sharing	0.0631 ± 0.015 ms	0.0638 ± 0.014 ms	0.989
utils/copy/preserve_sharing	0.279 ± 0.044 ms	0.277 ± 0.041 ms	1.01
utils/count_constants	30.2 ± 4.1 μs	0.0326 ± 0.0042 ms	0.926
utils/count_depth	0.0331 ± 0.0053 ms	0.0319 ± 0.0051 ms	1.04
utils/count_nodes	28.8 ± 5.3 μs	30.7 ± 4.7 μs	0.938
utils/get_set_constants!	0.109 ± 0.02 ms	0.114 ± 0.019 ms	0.957
utils/has_constants	9.82 ± 2 μs	9.96 ± 1.7 μs	0.986
utils/has_operators	3.07 ± 0.37 μs	2.8 ± 0.4 μs	1.1
utils/index_constants	0.0905 ± 0.019 ms	0.096 ± 0.017 ms	0.943
utils/is_constant	9.52 ± 2.1 μs	14.8 ± 2.5 μs	0.642
utils/simplify_tree	0.206 ± 0.041 ms	0.205 ± 0.039 ms	1
utils/string_tree	1.39 ± 0.17 ms	0.432 ± 0.072 ms	3.21

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.
Go to "Actions"->"Benchmark a pull request"->[the most recent run]->"Artifacts" (at the bottom).

[MilesCranmer]

Fixes #38