benchmarking http.jl vs download

[drizk1]

was briefly thinking about switchig to use download instead of HTTP.jl to reduce a dependency, but did a little benchmarking.. and it looks like using HTTP.jl is actually faster, so i think we might as well leave it

julia> url_file_path = "https://vincentarelbundock.github.io/Rdatasets/csv/openintro/ucla_f18.csv"
"https://vincentarelbundock.github.io/Rdatasets/csv/openintro/ucla_f18.csv"

julia> @benchmark CSV.read(download(url_file_path), DataFrame)
BenchmarkTools.Trial: 66 samples with 1 evaluation per sample.
 Range (min … max):  58.517 ms … 100.189 ms  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     75.821 ms               ┊ GC (median):    0.00%
 Time  (mean ± σ):   75.524 ms ±   9.110 ms  ┊ GC (mean ± σ):  0.00% ± 0.00%

               ▂   ▅        ▅ ▂    █  ▂                         
  ▅▅▁▅▁▅▅▅█▅▁█▁██▁▅██▅██▅█▁▅█████▁▅██▁██▁▁▁▅▁▁▁▁▁▅█▅▁▁▁▁▅▁▁▁▁▅ ▁
  58.5 ms         Histogram: frequency by time         98.1 ms <

 Memory estimate: 2.01 MiB, allocs estimate: 24063.

julia> @benchmark  read_csv(url_file_path)
BenchmarkTools.Trial: 101 samples with 1 evaluation per sample.
 Range (min … max):  37.511 ms … 78.963 ms  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     48.330 ms              ┊ GC (median):    0.00%
 Time  (mean ± σ):   49.855 ms ±  9.051 ms  ┊ GC (mean ± σ):  0.00% ± 0.00%

    ▃ █ ▃▃▁▆█  ▃▃ ▃▆▃▃ █ ▃  ▁          ▃                       
  ▄▇█▇█▁█████▇▁██▁████▇█▁█▄▁█▇▁▁▄▄▁▁▇▄▄█▄▁▁▁▄▁▁▁▁▇▁▁▄▁▁▄▁▁▄▁▄ ▄
  37.5 ms         Histogram: frequency by time        75.8 ms <

 Memory estimate: 4.19 MiB, allocs estimate: 23570.

julia> @benchmark CSV.read(local_file_path, DataFrame)
BenchmarkTools.Trial: 527 samples with 1 evaluation per sample.
 Range (min … max):  7.186 ms … 329.770 ms  ┊ GC (min … max): 0.00% … 93.91%
 Time  (median):     8.608 ms               ┊ GC (median):    0.00%
 Time  (mean ± σ):   9.507 ms ±  14.238 ms  ┊ GC (mean ± σ):  7.04% ±  5.91%

   █▆    ▂▁                                                    
  ▅██▆▅▅▆██▇▆▅▄▅▄▃▃▃▃▂▃▂▂▂▂▂▂▂▁▁▁▁▁▁▁▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▂▁▁▁▁▂ ▃
  7.19 ms         Histogram: frequency by time        19.2 ms <

 Memory estimate: 1.46 MiB, allocs estimate: 22579

julia> @benchmark read_csv(local_file_path)
BenchmarkTools.Trial: 527 samples with 1 evaluation per sample.
 Range (min … max):  7.199 ms … 284.456 ms  ┊ GC (min … max): 0.00% … 94.96%
 Time  (median):     8.905 ms               ┊ GC (median):    0.00%
 Time  (mean ± σ):   9.476 ms ±  12.137 ms  ┊ GC (mean ± σ):  6.49% ±  6.37%

    ▅█▃       ▁▂▂▅                                             
  ▅▇███▆▇▄▃▅▄▅████▇▇▇▄▅▅▄▄▄▁▃▃▃▃▃▂▃▃▂▂▃▁▃▁▁▁▃▃▁▁▁▁▂▁▂▁▁▂▂▁▁▁▂ ▃
  7.2 ms          Histogram: frequency by time          15 ms <

 Memory estimate: 1.69 MiB, allocs estimate: 22677.