[Feature]: Support for Binary data?

### Feature Request I would love to use SR in a binary context: ```julia using SymbolicRegression A = rand(Bool, (1000, 5)) B = @. (A[:, 1] | A[:, 2]) & (!A[:, 3]) ⊻ A[:, 4] # I could have added some noise but let's keep it simple # Define options for symbolic regression options = SymbolicRegression.Options(; binary_operators=[&, |, ⊻], unary_operators=[!], progress=true, verbosity=1, ) # Run symbolic regression hall_of_fame = equation_search( A, B; niterations=300, options=options, parallelism=:serial ) # => ERROR: InexactError: Bool(-100.0) ``` As I saw in the [docs](https://ai.damtp.cam.ac.uk/symbolicregression/dev/examples/custom_types) you can apply SR to strings, I am not entirely sure who to recreate it with binary data or if it's something that can't be done?