Add unified wrapper functions for plot families

[ishaan-arora-1]

Fixes #433

This adds high-level wrapper functions that let users access each plot family through a single function + type argument, instead of needing to know the exact function name upfront.

New wrappers

PPC: ppc_error(), ppc_dist(), ppc_discrete(), ppc_loo()
PPD: ppd_dist()
MCMC: mcmc_dist(), mcmc_trace_w(), mcmc_diag(), mcmc_nuts(), mcmc_recover()

For example, instead of remembering ppc_dens_overlay() vs ppc_ecdf_overlay() vs ppc_hist(), users can call:

ppc_dist(y, yrep, type = "dens_overlay")
ppc_dist(y, yrep, type = "hist")
ppc_dist(y, yrep, type = "ecdf_overlay", grouped = TRUE, group = g)

Partial matching on type is supported. Families with _grouped variants accept grouped = TRUE. All other arguments are forwarded via ....

The underlying functions remain fully available — these wrappers are thin dispatchers that just make the API a bit more discoverable.

What's included

• R/wrappers.R — all wrapper implementations with a shared .dispatch() helper
• man/bayesplot-wrappers.Rd — consolidated docs listing valid types per family
• tests/testthat/test-wrappers.R — tests covering dispatch, grouped variants, error cases, and partial matching
• NEWS.md and NAMESPACE updated

[jgabry]

Let's wait on this until we've had a chance to discuss in #433

[jgabry]

In general for things like this that require discussion we should wait to open PRs until we've had the discussion and agreed on the new design. For things like fixing deprecations or fixing documentation (or other things like that that don't require discussion) it's ok to just open the PR.

[ishaan-arora-1]

In general for things like this that require discussion we should wait to open PRs until we've had the discussion and agreed on the new design. For things like fixing deprecations or fixing documentation (or other things like that that don't require discussion) it's ok to just open the PR.

Alright, thankyou for the feedback!! Will keep in mind.