feat: add center function

[richtia]

PR to add center function.

Originally worked on in #248.

Instead of allowing a string with multiple characters to be used as padding, like lpad and rpad, center will only allow a single character for padding.

[jvanstraten]

The description is not very explicit about it needing to be a single character; character vs characters could easily be misinterpreted as a typo. This is especially problematic because the prototypes allow anything from 0 to 2^31-1 characters to be specified for character. The varchar<L2> could at least be made a varchar<1> to make it 0 or 1 character in that case, though that doesn't really solve the problem completely and doesn't solve anything for the string variant... The closest thing to a character type that Substrait has is fixedchar<1>, but using that bends the rules of #251.

I'm on the fence about what the right approach is, but if we stick to the implementations as currently defined, the description should state what happens if != 1 characters are passed to the character argument (throw error that fails the whole query? return null? use first character for >1 and use space for 0 characters? something else? explicitly undefined/up to the consumer to decide?)

[richtia]

The description is not very explicit about it needing to be a single character; character vs characters could easily be misinterpreted as a typo. This is especially problematic because the prototypes allow anything from 0 to 2^31-1 characters to be specified for character. The varchar<L2> could at least be made a varchar<1> to make it 0 or 1 character in that case, though that doesn't really solve the problem completely and doesn't solve anything for the string variant... The closest thing to a character type that Substrait has is fixedchar<1>, but using that bends the rules of #251.

I'm on the fence about what the right approach is, but if we stick to the implementations as currently defined, the description should state what happens if != 1 characters are passed to the character argument (throw error that fails the whole query? return null? use first character for >1 and use space for 0 characters? something else? explicitly undefined/up to the consumer to decide?)

Updated the description and arg implementation to specify 1 character. For now, i've also just explicitly stated the undefined behavior if the number of characters passed to the character argument is not 1.

[jvanstraten]

LGTM. CI failure is due to a network failure.