gitbook 和 split_by 的 bookdown html 格式问题

Question

I am generating a gitbook report with Rstudio the bookdown package.

It is fairly simple in terms of underlying R code, just a few recent tweaks for:

1. reducing text to 80% of page width
2. using double columns and
3. adding line number in the R code displayed.

Everything works well, except when I added " split_by: rmd " in the _output.yml. When doing so the resulting output doesn't respect the margin around the text anymore.

I don't know much about html yet, but looking at the html inspector revealed that the sections are located outside the inner-page formatting when using " split_by: rmd "

Default (no split_by argument):

With split_by: rmd

This is a shot in the dark as I cannot share the code and I am not able to reproduce the error with the minimal bookdown example from Yihui: https://github.com/rstudio/bookdown-demo .

Any leads to identify the origin of the error or even better propose a solution would be very welcomed!

Building the book from a R script:

bookdown::render_book(
  input = "index.Rmd",
  output_format = "bookdown::gitbook",
  output_dir = paste0("gitbook-", format(Sys.time(), format = "%Y-%m-%d-%H%M%S"))
)

index.rmd YAML header:

---
title: "blahblah"
subtitle: "blahblahblah"
author: "DRAFT"
date: "August 2020"
documentclass: article
fontsize: 12pt
geometry: margin=2cm
link-citations: yes
#mainfont: Arial
bibliography: packages.bib
site: bookdown::bookdown_site
biblio-style: apalike
urlcolor: blue
---

```{r setup, include=FALSE}

knitr::opts_chunk$set(echo = TRUE, eval = FALSE, attr.source='.numberLines')

table_format <- knitr::opts_knit$get('rmarkdown.pandoc.to')

if (table_format %in% c("html", "latex")) {
  library(kableExtra)
  knitr::opts_chunk$set(fig.pos='H', fig.align='center', out.width='80%')
}

## Automatically create a bib database for R packages
knitr::write_bib(c(.packages(), 'bookdown', 'knitr', 'rmarkdown', 'Hmisc'), 'packages.bib')
```

_output.yml:

bookdown::gitbook:
      css: style.css
      config:
        toc:
          before: |
            <li><a href="./">My book title</a></li>
          #after: |
          #  <li><a href="https://github.com/rstudio/bookdown" target="blank">Published with bookdown</a></li>
        edit: null
        download: null
        sharing: null
        info: null
      split_bib: FALSE
      split_by: rmd

style.css:

p.caption {
  color: #777;
  margin-top: 10px;
}
p code {
  white-space: inherit;
}
pre {
  word-break: normal;
  word-wrap: normal;
}
pre code {
  white-space: inherit;
}

/* watermark for draft report
.watermark {
  opacity: 0.2;
  position: fixed;
  top: 45%;
  left: 45%;
  font-size: 500%;
  color: #606099;
  z-index: 1000000;
}
*/

.book .book-body .page-wrapper .page-inner {
  max-width: 80% !important;
}

/* Increase space to display line number in R chunks correctly */
pre.numberSource code > span > a:first-child::before {
  left: -0.3em;
}

/* for multi cols */
/*.cols {display: flex; } /* uncomment for flex column size */
.cols {display: grid; grid-template-columns: 30% 50% 20%;} /* for fixed column size */

Answer 1

I found that the origin of the problem was the \newpage I added at the beginning of each rmd file for building the report in pdf format.

Solution 1: was to change split_by: rmd to split_by: section as my rmd files correspond mostly to level 2 sections.

Solution 2: was to put a wrapper around \newpage so that they are not evaluated when the output is html:

`r if (knitr::opts_knit$get('rmarkdown.pandoc.to') != "html") '
\\newpage
'`

Other approaches to solution 2 have been described here: How to add \newpage in Rmarkdown in a smart way?