revise autoplot() to drop faceting for univariate models

[mdscheuerell]

The default behavior of autoplot_marssPredict() is to add a facet_wrap() for the different observations (or states) with names given by the rownames in the predict object being passed to it (eg, here and here). How about adding functionality to drop the faceting for univariate models so they don't contain the facet label? For example, something like

if(length(unique(df$.rownames)) == 1) {
  p1 <- p1 +
    ggplot2::geom_line(linetype = plotpar$line.type, color = plotpar$line.col, linewidth = plotpar$line.size) +
    ggplot2::xlab("Time") + ggplot2::ylab("Estimate") +
    ggplot2::ggtitle(plottitle)
} else {
  p1 <- p1 +
    ggplot2::geom_line(linetype = plotpar$line.type, color = plotpar$line.col, linewidth = plotpar$line.size) +
    ggplot2::xlab("Time") + ggplot2::ylab("Estimate") +
    ggplot2::facet_wrap(~ df$.rownames, scale = "free_y") +
    ggplot2::ggtitle(plottitle)
}