refactor: improve symbolic_evaluate performance

[AayushSabharwal]

Before:

julia> @be (expr, subs) symbolic_evaluate(_...)
Benchmark: 156 samples with 1 evaluation
min    453.542 μs (2318 allocs: 66.359 KiB)
median 501.958 μs (2318 allocs: 66.359 KiB)
mean   605.138 μs (2318 allocs: 66.359 KiB)
max    2.379 ms (2318 allocs: 66.359 KiB)

After:

julia> @be (expr, subs) symbolic_evaluate(_...)
Benchmark: 330 samples with 1 evaluation
min    269.209 μs (1183 allocs: 28.891 KiB)
median 274.625 μs (1183 allocs: 28.891 KiB)
mean   282.744 μs (1183 allocs: 28.891 KiB)
max    588.542 μs (1183 allocs: 28.891 KiB)

Checklist

• Appropriate tests were added
• Any code changes were done in a way that does not break public API
• All documentation related to code changes were updated
• The new code follows the
  contributor guidelines, in particular the SciML Style Guide and
  COLPRAC.
• Any new documentation only uses public API

Additional context

Add any other context about the problem here.

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 13 lines in your changes are missing coverage. Please review.

Project coverage is 33.13%. Comparing base (d4b9089) to head (554a212).

Files	Patch %	Lines
src/trait.jl	0.00%	13 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #57      +/-   ##
==========================================
- Coverage   34.33%   33.13%   -1.20%     
==========================================
  Files           7        7              
  Lines         332      341       +9     
==========================================
- Hits          114      113       -1     
- Misses        218      228      +10     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.