Tool to summarise a .bib file

top-ten.R

#!/usr/bin/env Rscript
#'
#' top-ten-0.1.0
#' =============
#'
#' This script prints a list of the top ten journals as ranked by the number of
#' references you have from each of them.
#'
#' Usage
#' -----
#'
#' ./top-ten.R -i references.bib -o journal-count.png
#'
suppressPackageStartupMessages(library(ggplot2))
suppressPackageStartupMessages(library(argparse))
library(bib2df)

parser <- ArgumentParser()

parser$add_argument(
         "-i",
         "--input",
         type = "character",
         help = "Filepath to bibtex file."
       )

parser$add_argument(
         "-o",
         "--output",
         type = "character",
         help = "Filepath of image file to make."
       )

parser$add_argument(
         "-n",
         "--number",
         type = "integer",
         default = 15,
         help = "Number of entries to show (defaults to 15)."
       )

args <- parser$parse_args()


## x1 <- bib2df(file = "references.bib")
x1 <- bib2df(file = args$input)
x1 <- as.character(x1$JOURNAL)

x2 <- as.data.frame(table(x1))


x3 <- x2[order(x2$Freq, decreasing=T),]
names(x3) <- c("journal", "references")

x4 <- rbind(
  head(x3, args$number),
  data.frame(
    journal = "other",
    references = sum(tail(x3$references, -args$number))))

x4$journal <- factor(x = x4$journal, levels = rev(x4$journal))


g <- ggplot() +
  geom_col(
    data = x4,
    mapping = aes(y = journal, x = references)
  ) +
  geom_label(data = x4,
             mapping = aes(y = journal, x = 1.1, label = references),
             hjust = 0,
             vjust = 0.5,
             colour = "white",
             fill = NA,
             label.size = NA,
             size = 6) +
  scale_x_log10() +
  ggtitle(
    args$input,
    subtitle = format(Sys.time(), "%d %b %Y")
  ) +
  theme(axis.text.x = element_text(size = 20),
        axis.title.x = element_blank(),
        axis.title.y = element_blank(),
        axis.text.y = element_text(size = 20),
        title = element_text(size=30))

if (interactive()) {
  print(g)
} else {
  print.data.frame(x4, row.names = FALSE)
  ggsave(
    ## filename = "journal-count.png",
    filename = args$output,
    plot = g,
    height = 7.8,
    width = 14.2
  )
}