add shape argument to mcmc_scatter

[behramulukir]

Until now, users were able to control the look of their divergent points via the np_style argument and scatter_style_np function. Even though the mcmc_scatter function provided some control over the look of main points, such as alpha and size of them, users were not able to change the shape of the main points.

This PR adds shape argument to mcmc_scatter to give users control over the shape of their main points, as requested by @avehtari in #45.

Tasks:

• implement the argument
• implement tests
• update the documentation

Examples

x <- example_mcmc_draws(params = 6)
color_scheme_set("brightblue")

mcmc_scatter(x, pars = c("alpha", "sigma"), transform = list(sigma = "log"))
mcmc_scatter(x, pars = c("alpha", "sigma"), transform = list(sigma = "log"), shape = 23)

Similarly, users can control the shapes of their main points when they have divergent points.

library(rstanarm)
fit <- stan_glm(
     mpg ~ ., data = mtcars,
     iter = 1000, refresh = 0,
     # this combo of prior and adapt_delta should lead to some divergences
     prior = hs(),
     adapt_delta = 0.9
)

posterior <- as.array(fit)
np <- nuts_params(fit)

mcmc_scatter(posterior, pars = c("wt", "sigma"), np = np)
mcmc_scatter(posterior, pars = c("wt", "sigma"), np = np, shape=25)

div_style <- scatter_style_np(div_color = "green", div_shape = 4, div_size = 4)

mcmc_scatter(posterior, pars = c("sigma", "(Intercept)"), np = np, np_style = div_style)
mcmc_scatter(posterior, pars = c("sigma", "(Intercept)"), np = np, np_style = div_style, shape=22)

[avehtari]

Looks good, and I tested that it worked in my use case

[behramulukir]

Great then! I'll see what's going wrong with the tests and finish this PR

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.66%. Comparing base (89fea18) to head (b4866e3).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #375      +/-   ##
==========================================
+ Coverage   98.62%   98.66%   +0.03%     
==========================================
  Files          35       35              
  Lines        5748     5749       +1     
==========================================
+ Hits         5669     5672       +3     
+ Misses         79       77       -2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[jgabry]

Great, thanks @behramulukir. Everything looks good, so I'll go ahead and merge it now.