Make Zero and One extensional traits

[qryxip]

Makes Zero and One extensional traits for Sum and Product, respectively.

As I said in rust-jp.slack.com 3 years ago IIRC, we might as well regard the "zero value" and "one value" described in the documentation as identity elements.

• std::iter::Iterator::sum
• std::iter::Iterator::product

The implementations should be still inlined.

Affects #102.

[TonalidadeHidrica]

Will you need to implement Sum trait (Sum::sum function) when you want to use the additive segtree for a custom type, if we adopt this PR?

[qryxip]

Yes - instead of impling the Zero from this crate.

struct S(i32);

impl Add for S {
    type Output = Self;

    fn add(self, rhs: Self) -> Self::Output {
        Self(self.0 + rhs.0)
    }
}

impl Sum for S {
    fn sum<I: Iterator<Item = Self>>(iter: I) -> Self {
        iter.fold(Self(0), Add::add)
        //        ^^^^^^^
        //        additive identity here
    }
}

// or:
struct S(i32);

//impl Add for S {
//    type Output = Self;
//
//    fn add(self, rhs: Self) -> Self::Output {
//        Self(self.0 + rhs.0)
//    }
//}

impl Sum for S {
    // only used for `Zero` from ac-library-rs
    fn sum<I: Iterator<Item = Self>>(iter: I) -> Self {
        iter.fold(Self(0), |_, _| unreachable!())
        //        ^^^^^^^
        //        additive identity here
    }
}

Sum and Product are free from ac-library-rs, so personal libraries for non-AtCoder competitive programming services should easily adopt.

[qryxip]

↑ The second code example was not necessary.

[TonalidadeHidrica]

I don't think it is a good idea. "If you want to use + monoid, implement Zero trait." --- that's straightforward. But, "if you want to use + monoid, implement Sum trait." --- that requires an additional effort and not intuitive for me. How do you think?