Add several methods to use string cache

[sorhawell]

Close #350, close #234

[sorhawell]

@etiennebacher I think we can just merge this one in as is.

[etiennebacher]

@sorhawell thanks! I tweaked the docs a bit, nothing major. I just have one question regarding #234: it looks like I can make it work if I use pl$enable_string_cache() before creating the DataFrame

pl$enable_string_cache(TRUE)
pl_letters_cat <- pl$DataFrame(list(a = factor(letters[1:3])))
pl_letters_cat$filter(
  pl$col("a")$is_in(pl$lit("a"))
)

shape: (1, 1)
┌─────┐
│ a   │
│ --- │
│ cat │
╞═════╡
│ a   │
└─────┘

but after resetting the cache to FALSE and creating the DataFrame in pl$with_string_cache() doesn't seem to work:

pl$with_string_cache({
  pl_letters_cat <- pl$DataFrame(list(a = factor(letters[1:3])))
})
pl_letters_cat$filter(
  pl$col("a")$is_in(pl$lit("a"))
)
Error: Execution halted with the following contexts
0: In R: in $collect():
0: During function call [pl_letters_cat$filter(pl$col("a")$is_in(pl$lit("a")))]
1: Encountered the following error in Rust-Polars:
joins/or comparisons on categoricals can only happen if they were created under the same global string cache

How should I use pl$with_string_cache()?

[etiennebacher]

My bad, I just saw that the whole thing should be in pl$with_string_cache():

pl$with_string_cache({
  pl_letters_cat <- pl$DataFrame(list(a = factor(letters[1:3])))
  pl_letters_cat$filter(
    pl$col("a")$is_in(pl$lit("a"))
  )
})

shape: (1, 1)
┌─────┐
│ a   │
│ --- │
│ cat │
╞═════╡
│ a   │
└─────┘