Simplify md_compute_lorenz()

[Aeilert]

Hi @tonyfujs

I made an attempt to simplify and improve md_compute_lorenz(). I think this makes it more clear what's going on. But not sure that is actually better. Computation speed is marginally less, but memory allocation is higher.

E.g.

data("md_ABC_2000_income")
df <- wbpip:::md_clean_data(md_ABC_2000_income,
                    welfare = 'welfare',
                    weight = 'weight')$data

# A tibble: 2 x 13
  expression                                            min   median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc
  <bch:expr>                                       <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl> <int> <dbl>
1 wbpip:::md_compute_lorenz(df$welfare, df$weight)    782us   1.11ms      712.    18.2KB     2.86   996     4
2 md_compute_lorenz_3(df$welfare, df$weight)          806us  875.7us     1033.    82.5KB     5.19   995     5

Unit: microseconds
                                             expr   min     lq     mean median      uq     max neval cld
 wbpip:::md_compute_lorenz(df$welfare, df$weight) 842.2 929.25 1289.384 1005.5 1348.15 24203.3  1000   a
       md_compute_lorenz_3(df$welfare, df$weight) 821.5 907.35 1297.613  977.3 1232.10 36638.2  1000   a