include tvpack.jl with bivariate normal cdf

[magerton]

This reincludes the tvpack.jl file with the bivariate normal CDF function. I've added a bvncdf(x,y,r) function that corresponds to the standard normal cdf Φ(x,y;ρ) where r=ρ. The function also handles Inf and -Inf arguments correctly now.

[magerton]

Addresses #43 and comment by @dmbates #36 (comment)

[codecov-commenter]

Codecov Report

Merging #67 (40222d8) into master (28d732a) will increase coverage by 6.82%.
The diff coverage is 18.14%.

@@            Coverage Diff             @@
##           master      #67      +/-   ##
==========================================
+ Coverage   48.64%   55.47%   +6.82%     
==========================================
  Files          13       13              
  Lines         516      530      +14     
==========================================
+ Hits          251      294      +43     
+ Misses        265      236      -29     

Impacted Files	Coverage Δ
src/tvpack.jl	16.34% <18.14%> (+16.34%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 28d732a...40222d8. Read the comment docs.

[spaette]

Possibly throw could be misspelled

$ tail -n 15 StatsFuns.jl/src/tvpack.jl
    # here we either have 0 or the marginal
    else
        if dh == typemax(T) || dk == typemax(T)
            return zero(T)
        elseif dh == typemin(T)
            return normcdf(-dk)
        elseif dk == typemin(T)
            return normcdf(-dh)
        else
            thow(error())
        end
    end
end

bvncdf(dh, dk, r) = bvnuppercdf(-dh, -dk, r)
$ 

[magerton]

Possibly throw could be misspelled

$ tail -n 15 StatsFuns.jl/src/tvpack.jl
    # here we either have 0 or the marginal
    else
        if dh == typemax(T) || dk == typemax(T)
            return zero(T)
        elseif dh == typemin(T)
            return normcdf(-dk)
        elseif dk == typemin(T)
            return normcdf(-dh)
        else
            thow(error())
        end
    end
end

bvncdf(dh, dk, r) = bvnuppercdf(-dh, -dk, r)
$ 

Good catch, @spaette. Just submitted PR #160