Add fit!(o, y, n) to fit multiple observations of the same value

[adknudson]

This closes #284 that I opened. It's a minimum implementation where simple stats such as Sum, Mean, and CountMap (and a few others) can efficiently update their sufficient stats given a value and the number of times it was observed.

Methods without a specialized update will fall back to

for _ in 1:n
    fit!(o, y)
end

If you prefer to use fitn!(o, y, n), then I can make the change and update the pull request.

[adknudson]

Other things to check:

1. I added a default implementation for _fit!(o, y) that throws an error if it is not implemented for an OnlineStat.
2. The specialized method for Sum(Integer) needs review. _fit(Sum(Integer), 3.5, 10) will first multiply 3.5 by 10, then round to the nearest integer, which will give a different result than fitting 3.5 ten times in a row.

[joshday]

This won't be correct for Mean{T,W} when W is something other than EqualWeight

[adknudson]

Does this imply there is also a problem with _merge!(o::Mean, o2::Mean)? My thinking is along the lines that fitting y k times is equivalent to merging two means where the second one has k values and a mean of y. What is the appropriate way to deal with weights? Like this?

function _fit!(o::Mean{T}, y, n) where {T}
    o.n += n
    o.μ = smooth(o.μ, y, o.weight(o.n / n))
end

[adknudson]

Never mind, in testing that doesn't work. I can define a special version for Mean with equal weight, and then use the generic fallback otherwise

[joshday]

I'll merge when CI goes green!