More string trimming capabilities

The existing `string.trim`, `string.trim_left` and `string.trim_right` functions currently only trim whitespace.

Could we add equivalent trimming functions that take the codepoints to be trimmed as an input, similar to Erlang's [`string.trim/3`](https://www.erlang.org/doc/man/string#trim-3)?

E.g.

```gleam
import gleam/list

/// Removes all occurrences of the specified codepoints from the start and
/// end of a `String`.
///
pub fn trim_codepoints(s: String, codepoints: List(UtfCodepoint)) -> String {
string_trim(s, Both, codepoints_to_ints(codepoints))
}

/// Removes all occurrences of the specified codepoints from the start of a
/// `String`.
///
pub fn trim_codepoints_left(s: String, codepoints: List(UtfCodepoint)) -> String {
string_trim(s, Leading, codepoints_to_ints(codepoints))
}

/// Removes all occurrences of the specified codepoints from the end of a
/// `String`.
///
pub fn trim_codepoints_right(
s: String,
codepoints: List(UtfCodepoint),
) -> String {
string_trim(s, Trailing, codepoints_to_ints(codepoints))
}

fn codepoints_to_ints(codepoints: List(UtfCodepoint)) {
codepoints
|> list.map(string.utf_codepoint_to_int)
}

type Direction {
Leading
Trailing
Both
}

@target(erlang)
@external(erlang, "string", "trim")
fn string_trim(a: String, b: Direction, c: List(Int)) -> String
```

There are other ways to slice this, such as taking a `String` as the final parameter and using its codepoints as the ones to trim, rather than a `List(UtfCodepoint)`.

I haven't looked into what the JavaScript implementation would be, but can do so.

On a related note, my understanding is that the existing methods with the words "left" and "right" in them assume left-to-right reading order, which doesn't hold for all languages (e.g. Arabic, Hebrew). Could consider using "start" and "end" instead, or adding these as aliases.