R/count_words.R

In jhollway/iheiddown: For Writing Geneva Graduate Institute Documents

#' A function for counting words in an Rmarkdown document.
#'
#' This function takes a path to a .Rmd file and returns a word count.
#' For best results each sentence should start on a new line.
#' Inline code chunks should also start on a new line.
#' 
#' @name countwords
#' @param file A path to a .Rmd file
#' @import dplyr
#' @importFrom rlang .data
#' @importFrom readr read_lines
#' @importFrom tibble as_tibble
#' @importFrom rstudioapi getSourceEditorContext
#' @importFrom tidytext unnest_tokens
#' @importFrom pdftools pdf_text
#' @return A scalar representing the number of words in the document.
#' @examples
#' rmarkdown::draft(file = "test", template = "html_vignette",
#' package = "rmarkdown", create_dir = TRUE, edit = FALSE)
#' iheiddown::count_words("test/test.Rmd")
#' unlink("test", recursive = TRUE)
#'
NULL

#' @rdname countwords
#' @export
count_words <- function(file) {
  # Get currently viewed panel in RStudio if file is not specified
  if (missing(file)) file <- rstudioapi::getSourceEditorContext()$path
  wc <- readr::read_lines(file) %>%
    tibble::as_tibble() %>%
    remove_front_matter() %>%
    remove_code_chunks() %>%
    remove_inline_code() %>%
    remove_html_comment() %>%
    tidytext::unnest_tokens(output = .data$words, input = .data$value) %>%
    nrow()
  return(wc)
}

#' @rdname countwords
#' @export
count_words2 <- function(file) {
  # Get currently viewed panel in RStudio if file is not specified
  if (missing(file)) {
    file <- rstudioapi::getSourceEditorContext()$path
    file <- stringr::str_replace(file, ".Rmd", ".pdf")
  } 
  
  pdf <- pdftools::pdf_text(file)
  wc <- sum(unlist(lapply(wc, function(x) length(x))))

  return(wc)
}


# Helper function for removing unwanted lines from count
# Checks is value is odd.
is_odd <- function(x, val) {
  x %% 2 == 1
}

# Helper function to remove front matter (lines starting with "---" and
# anything between)
remove_front_matter <- function(x) {
  mutate(x, is_code = cumsum(grepl("^---", .data$value))) %>%
    group_by(.data$is_code) %>%
    mutate(start_end = lag(.data$is_code, 1)) %>%
    ungroup() %>%
    filter(!is_odd(.data$is_code), !is.na(.data$start_end)) %>%
    select(-.data$is_code, -.data$start_end)
}

# Helper function for removing knitr code chunks (lines starting with "```"
# and anything in between)
remove_code_chunks <- function(x) {
  mutate(x, is_code = cumsum(grepl("^```", .data$value))) %>%
    group_by(.data$is_code) %>%
    mutate(start_end = lag(.data$is_code, 1)) %>%
    ungroup() %>%
    mutate(is_odd = is_odd(.data$is_code)) %>%
    filter(.data$is_odd != TRUE, !is.na(.data$start_end)) %>%
    select(-.data$is_code, -.data$start_end, -.data$is_odd)
}

# Helper function to remove inline code (lines starting with "`r")
remove_inline_code <- function(x) {
  filter(x, !grepl("`r", .data$value))
}

# Helper function to remove HTML comments (lines starting with "<!--" and "-->")
remove_html_comment <- function(x) {
  mutate(x, start_comment = cumsum(grepl("^<!-- ", .data$value)),
         end_comment = cumsum(grepl(" -->", .data$value))) %>%
    group_by(.data$start_comment, .data$end_comment) %>%
    mutate(start_end = lag(.data$start_comment, 1)) %>%
    ungroup() %>%
    mutate(remove = if_else(.data$start_comment - .data$end_comment == 1 |
                              is.na(.data$start_end), 1, 0)) %>%
    filter(remove != TRUE) %>%
    select(-.data$start_comment, -.data$end_comment, -.data$remove)
}