Option to specify measure names in `evaluate` and `benchmark`

[pat-alt]

It would be nice to be able to pass measure_names::Vector{Strin} or something along these lines to evaluate and benchmark. This would be particularly useful for passing on anonymous functions to measure. Currently, only meta_data` can help, but it's not ideal:

julia> benchmark([ce]; measure=[((x;kwrgs...) -> feature_sensitivity(x, 1))], meta_data=[Dict(:measure_name=>"sens_1")])
Benchmark(1×7 DataFrame
 Row │ sample                                num_counterfactual  variable  value    measure_name  target  factual 
     │ Base.UUID                             Int64               String    Float32  String        Int64   Int64   
─────┼────────────────────────────────────────────────────────────────────────────────────────────────────────────
   1 │ 50193892-009d-11f0-2009-f1b3b93e7d48                   1  #35       1.34479  sens_1             2        1, nothing)

For non-anonymous functions, the function name is used as expected, but it might still be nice to be able to specify something more readable:

julia> sens1(x;kwrgs...) = feature_sensitivity(x, 1)
sens1 (generic function with 2 methods)

julia> benchmark([ce]; measure=[sens1])
Benchmark(1×8 DataFrame
 Row │ sample                                num_counterfactual  variable  value    generator                          model                              t ⋯
     │ Base.UUID                             Int64               String    Float32  Symbol                             Symbol                             I ⋯
─────┼───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
   1 │ f23365a6-009f-11f0-1c9a-1bc824dccac1                   1  sens1     1.34479  GradientBasedGenerator(nothing, …  CounterfactualExplanations.Model…    ⋯
                                                                                                                                            2 columns omitted, nothing)