Just to make the VC tip a clean release after init syntax thrash

That's it.

Qualify imports with std/ meta-package

Really just that minor std/ qualification thing & comments.

Add "DiGamma" ψ (psi) function for integer arguments

Well, probably this should bump to 0.8 with a whole new special function, but it's all good since ψ (psi) here is only implemented to the raw bare minimum needed for a mutual information estimate and would clearly suck for large arguments.

Adapt to recent nim-devel changes

Adapt to nim-lang/Nim@a93c5d7 which broke template lentz*(). Newer layering is more clear anyway since den0 immediately follows den(). { Commit explains why caller itself does not add in den0, although a future convergence common-infrastructure will also probably be a breaking change. }

Add `maxTry` to `spfun/studentT.ccPv`

This prevents some pathological and not useful very long-lived loops.

Add `χ²` & `χ²c` entry points

That's it. More or less trivial, but some people love utf8 idents.

Some ease of use additions

Just these two:

• Add binomp.initBinomP - a little constructor proc for simpler usage at call sites
• Add studentT.ccPvEvalCount - a global evaluation counter (to easily id many-evals situations)

There was preliminary work on a CDF plotter, but it moved to https://github.com/c-blake/bu

Add good binomial p CI's & use for correlation significance

Add Wilson & Agresti scoring (as well as the roll-out-of-bed Wald stuff) for good CIs of binomial p parameter with a little cligen CLI utility for testing.

Use this new binomial proportion estimator for very strong simulation-based p-Values of correlations (Pearson linear or Spearman rank). { This largely nullifies the advantage of rank correlation, but sometimes people do still over-interpret "correlations close to 1". }

Fix oversight of using seq not openArray

For a quick patch release since the API is brand new anyway.

Add permutation test pValue for correlations

For small samples, this can have 2-4X better false positive control as per Yu & Hutson 2020/2021 which can make all the difference in even being distracted by insignificant correlations.