vaccine 1.2.1

Fixed a bug that displayed an incorrect package version number at startup.
Tweaked the SuperLearner library involved in standard error estimation for est_ce(..., type="NP"), resulting in improved performance.
Fixed a bug related to plot_ce(..., density_type="kde edge").
Fixed a minor bug related to the SuperLearner package introduced in version 1.2.0.

vaccine 1.2.0

Added the trim function, which can be used in conjunction with the plot_ce function to truncate the display of estimate objects produced by est_ce. Specifically, estimates with X-coordinates that lie outside specified quantiles of the observed distribution of the marker are removed. See package vignette and examples.
Added the as_table function, which formats estimate objects produced using est_ce as a table.

Implemented various changes to the plot_ce function: added background kernel density plotting, changed default plot styling, fixed a bug that caused CR plots to be displayed instead of CVE plots when plot_ce(..., which="CVE") was called, and added to the plotting section of the main package vignette.

In est_ce, added the option return_p_value; if set to TRUE, a P-value will be returned corresponding to the null hypothesis that the CVE curve (or equivalently, the CR curve in the vaccine group) is constant. If type='Cox' is specified, this will be a Wald-type test using the estimated Cox model parameter vector. If type='NP', this will be a nonparametric test assuming monotonicity of the curve.

Added functions for nonparametric estimation and inference for mediation analysis parameters, including the natural direct effect, the natural indirect effect, and the proportion mediated.
Adapted nonparametric and Cox-based CVE estimation and inference to handle settings in which some or all covariates are measured only in phase two (e.g., applicable to settings in which researchers want to control for a baseline biomarker measurement).
Added confidence band monotonization for nonparametric CVE inference.