feat(r): Implement infer schema methods

[paleolimbot]

This PR implements infer_nanoarrow_schema(), which gives a hypothetical type of what schema an object would have if it were converted to Arrow format. Arrow's arrow::infer_type() does the same thing, and before this PR, nanoarrow just fell back on Arrow to infer the type conversion. Now that types can be created natively, we don't need Arrow for this.

This PR implements schema inference using S3 dispatch. It would be faster to implement this in C, and this probably will be done in the future; however for now it seems like the overhead is acceptable for most uses (basically, you need more than 10,000 columns in a data frame to start noticing).

library(arrow, warn.conflicts = FALSE)
#> Some features are not enabled in this build of Arrow. Run `arrow_info()` for more information.
library(nanoarrow)

make_big_df <- function(n) {
  cols <- lapply(seq_len(n), function(...) integer())
  names(cols) <- paste0("col", seq_len(n))
  tibble::new_tibble(cols, nrow = 0)
}

df <- make_big_df(1e5)

bench::mark(
  infer_nanoarrow_schema(df),
  arrow::infer_type(df),
  check = F
)
#> Warning: Some expressions had a GC in every iteration; so filtering is
#> disabled.
#> # A tibble: 2 × 6
#>   expression                      min   median `itr/sec` mem_alloc `gc/sec`
#>   <bch:expr>                 <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl>
#> 1 infer_nanoarrow_schema(df)  582.4ms  582.4ms      1.72    2.43MB     13.7
#> 2 arrow::infer_type(df)        85.1ms   86.6ms     11.5     1.02MB      0

^{Created on 2023-02-08 with reprex v2.0.2}