Custom table of contents

DESCRIPTION

@@ -63,7 +63,8 @@ Imports:
     knitr (>= 1.22),
     rmarkdown (>= 2.3.5),
     xfun (>= 0.13),
-    tinytex (>= 0.12)
+    tinytex (>= 0.12),
+    stringr (>= 1.2.0)
 Suggests:
     htmlwidgets,
     rstudioapi,

NAMESPACE

@@ -36,6 +36,7 @@ export(tufte_html_book)
 export(word_document2)
 import(stats)
 import(utils)
+import(stringr)
 importFrom(xfun,in_dir)
 importFrom(xfun,read_utf8)
 importFrom(xfun,same_path)

R/gitbook.R

@@ -27,7 +27,7 @@ gitbook = function(
 ) {
   html_document2 = function(..., extra_dependencies = list()) {
     rmarkdown::html_document(
-      ..., extra_dependencies = c(extra_dependencies, gitbook_dependency(table_css))
+      ..., extra_dependencies = c(gitbook_dependency(table_css), extra_dependencies)
     )
   }
   gb_config = config
@@ -174,7 +174,46 @@ gitbook_toc = function(x, cur, config) {
   i2 = find_token(x, '<!--bookdown:toc2:end-->')
   x[i1] = ''; x[i2] = ''
   if (i2 - i1 < 2) return(x)
+
+  #Changed to allow for toc length to change over the course of the function
+  prior = x[1:i1]
   toc = x[(i1 + 1):(i2 - 1)]
+  post = x[i2:length(x)]
+
+  if(is.character(rmarkdown:::yaml_load_file('_bookdown.yml')[['custom_toc']])){  
+    toc_frame <- readRDS("toc.Rds")
+    #First add in divisions and headers
+    for(i in (1:nrow(toc_frame))){
+      if(toc_frame[i,6] == "HEADING"){
+        toc <- append(toc, paste('<li class="header">', toc_frame[i,4], '</li>', sep=""), i)
+      } else if(toc_frame[i,6] == "BREAK"){
+        toc <- append(toc, '<li class="divider"></li>', i)
+      } 
+    }    
+
+    #Check that the in length difference is due to the leading and trailing <ul> tags in toc
+    if(length(toc) - nrow(toc_frame) != 2){
+      warning("The toc file listed in _bookdown.yml has more elements than expected. \n
+          Potential causes include links to non-existent files or lines that are not one of: dividers, section headings, or pages. \n
+          Consider checking your toc file, or setting clean_book to FALSE and inspecting toc.Rds.")
+    }     
+
+    #Now sort out custom page titles and indents
+    for(i in 1:(nrow(toc_frame))){
+      #Finds pandoc-generated title (from first H1) and replaces it with title specified in the toc_frame
+      if(!is.na(toc_frame[i,2])){
+        toc[i+1] <- str_replace(toc[i+1], gsub(".*>(.+)</a>.*", "\\1", toc[i+1]), paste(" ", toc_frame[i,2]))
+      }
+      #Adds in <ul> tags so subchapters work as per the tabs in the toc_frame file
+      diff <- toc_frame[i,5] - toc_frame[i-1,5]
+      if(i == 1) diff <- 0
+      if(diff == 1){
+        toc[i] <- gsub('</li>$', '<ul>', toc[i])
+      } else if(diff < 0){
+        toc[i] <- paste(toc[i], strrep("</li></ul>", times = diff/-1), sep = "")
+      }
+    }
+  }
 
   # numbered sections
   r = '^<li><a href="([^#]*)(#[^"]+)"><span class="toc-section-number">([.A-Z0-9]+)</span>(.+)(</a>.*)$'
@@ -222,7 +261,8 @@ gitbook_toc = function(x, cur, config) {
       toc[n] = paste(c('<li class="divider"></li>', extra, toc[n]), collapse = '\n')
     }
   }
-  x[(i1 + 1):(i2 - 1)] = toc
+
+  x = c(prior, toc, post)
   x
 }
 

R/utils.R

@@ -68,37 +68,88 @@ book_filename = function(config = load_config(), fallback = TRUE) {
 }
 
 source_files = function(format = NULL, config = load_config(), all = FALSE) {
-  subdir = config[['rmd_subdir']]; subdir_yes = isTRUE(subdir) || is.character(subdir)
-  # a list of Rmd chapters
-  files = list.files('.', '[.]Rmd$', ignore.case = TRUE)
-  # content in subdir if asked
-  subdir_files = unlist(mapply(
-    list.files,
-    if (is.character(subdir)) subdir else '.', '[.]Rmd$', ignore.case = TRUE,
-    recursive = subdir_yes, full.names = is.character(subdir), USE.NAMES = FALSE
-  ))
-  subdir_files = setdiff(subdir_files, files)
-  files = c(files, subdir_files)
-  # if rmd_files is provided, use those files in addition to those under rmd_subdir
-  if (length(files2 <- config[['rmd_files']]) > 0) {
-    # users should specify 'docx' as the output format name for Word, but let's
-    # make 'word' an alias of 'docx' to avoid further confusion:
-    # https://stackoverflow.com/q/63678601/559676
-    if ('word' %in% names(files2) && identical(format, 'docx')) format = 'word'
-    if (is.list(files2)) files2 = if (all) unlist(files2) else files2[[format]]
-    # add those files to subdir content if any
-    files = if (subdir_yes) c(files2, subdir_files) else files2
+  toc_loc = config[['custom_toc']]
+  if(is.character(toc_loc)) {
+    files = toc_build(toc_loc)
+  } else {
+    subdir = config[['rmd_subdir']]; subdir_yes = isTRUE(subdir) || is.character(subdir)
+    # a list of Rmd chapters
+    files = list.files('.', '[.]Rmd$', ignore.case = TRUE)
+    # content in subdir if asked
+    subdir_files = unlist(mapply(
+      list.files,
+      if (is.character(subdir)) subdir else '.', '[.]Rmd$', ignore.case = TRUE,
+      recursive = subdir_yes, full.names = is.character(subdir), USE.NAMES = FALSE
+    ))
+    subdir_files = setdiff(subdir_files, files)
+    files = c(files, subdir_files)
+    # if rmd_files is provided, use those files in addition to those under rmd_subdir
+    if (length(files2 <- config[['rmd_files']]) > 0) {
+      # users should specify 'docx' as the output format name for Word, but let's
+      # make 'word' an alias of 'docx' to avoid further confusion:
+      # https://stackoverflow.com/q/63678601/559676
+      if ('word' %in% names(files2) && identical(format, 'docx')) format = 'word'
+      if (is.list(files2)) files2 = if (all) unlist(files2) else files2[[format]]
+      # add those files to subdir content if any
+      files = if (subdir_yes) c(files2, subdir_files) else files2
+    }
   }
   # exclude files that start with _, and the merged file
   files = files[grep('^[^_]', basename(files))]
   files = setdiff(files, with_ext(book_filename(config), c('.md', '.Rmd')))
   files = unique(gsub('^[.]/', '', files))
   index = 'index' == with_ext(files, '')
-  # if there is a index.Rmd, put it in the beginning
+  #if there is an index.Rmd, put it in the beginning
   if (any(index)) files = c(files[index], files[!index])
   check_special_chars(files)
 }
 
+
+toc_build <- function(toc_loc){
+  toc <- readLines(toc_loc)
+  toc <- as.data.frame(toc[nchar(toc) > 0])
+  names(toc)[1] <- "base"
+  toc$title <- str_sub(str_extract(toc$base, "(\\[.*\\])"), 2, -2) #Couldn't work out how to regex match between-but-not-including square brackets, so cheated
+  toc$relref <- str_sub(str_extract(toc$base, "\\(.*\\)"), 2, -2)
+  toc$relref <- str_replace(toc$relref, "(^/)", "")
+  toc$heading <- str_extract(toc$base, "(?:^#+(\ |)).+") ##Matches headings including #s; can't do variable length look-behinds?
+  toc$heading <- sapply(str_split(toc$heading, "^#+(\ |)"), `[`, 2) #Drops the #s
+  toc$indent <- str_count(str_extract(toc$base, ".+?(?=\\*)"), pattern = "\\t")
+  toc$indent[is.na(toc$indent)] <- 0
+  toc$type <- apply(toc, 1, function(r){
+    if(str_detect(r[1], "(---)")){
+      "BREAK"
+    } else if(!is.na(r[4])){
+      "HEADING"
+    } else if(is.character(r[3])){
+      "PAGE"
+    }
+  })
+  toc$exists <- sapply(toc$relref, function(r){ #Check that files listed in toc actually exist
+    file.exists(r)
+  })
+
+  #Check for files in toc.md without associated files in directory, and for files in toc_loc without an entry in toc.md
+  #Flags but allows site to be built anyway
+  missing <- toc[with(toc, is.na(relref) == FALSE & exists == FALSE),]$relref
+  if(length(missing) > 0){
+    warning('These files are listed in the table of contents but cannot be found: \n',
+            lapply(missing, function(r) paste(r, "\n")))
+  }
+  missed <- list.files('.' , full.names = TRUE, pattern = "(?i)([.]Rmd$)|([.]md$)", recursive = TRUE)
+  missed <- str_replace(missed, "(\\.\\/)", "")
+  missed <- setdiff(missed, toc$relref)
+  missed <- missed[!str_detect(missed, pattern = "(stubs/)|(toc.md)")]
+  if(length(missed) > 0){
+    warning('These files found in the "', getwd(), '" directory or subdirectories, but are not listed in toc.md: \n',
+            lapply(missed, function(r) paste(r, "\n")))
+  }
+  saveRDS(toc, file = "toc.Rds")
+  #Returns to source_files only the files that are listed in the toc and exist
+  return(toc[toc$exists == TRUE,]$relref)
+
+}
+
 output_dirname = function(dir, config = load_config(), create = TRUE) {
   if (is.null(dir)) {
     dir2 = config[['output_dir']]