Rank distribution for models

[mqwilber]

We need to implement a consistent rank function between models (Issue #73) but that is not too slow for practical use. For example, plnorm_ztrunc and logser_uptrunc have quite slow rank functions when the number of species gets large. How can we speed this up?

[mqwilber]

Here is some more information on this:

Try the following calls (Warning, you will probably have to Crtl-C)

logser_uptrunc.rank(800, .99, 100)

plnorm_ztrunc.rank(50, 2, 2, crit=.5)

How can we speed this up? Maybe we can use the old make_rank function or switch between the ppf call and the old make_rank function.