standard renderers + separate config file

[nanxstats]

• Consider making the liftr metadata fields into an individual yml file instead of inline, like pkgdown: https://hadley.github.io/pkgdown/articles/pkgdown.html

• Consider implement standard renderers, like bookdown: https://bookdown.org/home/getting-started.html

So: standard renderers + .yml config file.

[nanxstats]

• add call to lift() in render_docker() (while still export lift()).

• maybe render_docker() will not call render() furthermore; instead, it will call *_docker functions based on the output_format specified, which will call html_document().

• Mapping:

  render_docker() -> render()

  docker_html() -> html_document() docker_pdf() -> pdf_document() docker_word() -> word_document() docker_md() -> md_document()

• users will write this in the RMD header:

  output:
  liftr::docker_html:
  toc: true
  number_sections: true
  ...

Need to investigate the interface design of rmarkdown and knitr.