ln_cdf & ln_sf

[drtconway]

Hi Statrs,

Thanks for a fantastic library!

I do bioinformatics methods development, and in quite a few situations we find ourselves needing to evaluate in the tails of the distribution (we often need to apply quite aggressive multiple testing corrections). When working in R, we typically use the log.p=TRUE parameter for the cumulative distributions.

Are there any plans to provide log CDF functions in statrs? Obviously, default implementations that simply return the log of the regular cdf or sf functions would be a start, but as you probably know, it's not uncommon to use different numerical algorithms to implement ln_cdf or ln_sf to achieve good precision in the tails of the distribution.

Tom.