import_bib.R

# import_bib.R
# To convert from Bibtex to a data frame for working with the data in R.

library(dplyr)
library(stringr)
library(tidyr)

# 0. Set filename for the bibfile
the_bibfile <- "~/path/to/my.bib"

# 1. read the bib file as a vector of lines
bibdata <- scan(the_bibfile, 
                what="character", 
# skip omits junk from header; uncomment and adjust if needed
                # skip = 4, 
# nline limits reading to clear junk from footer; uncomment and adjust if needed
                # nlines = 2285,
                sep = "\n")

# 2. Index the start and end of each entry
startlines <- grep("^@",bibdata)
endlines <-
  (startlines[2:length(startlines)]-1) |>
  c(length(bibdata))

# 3. Get a list of cite keys in order.
citekeys <- bibdata[startlines] |>
  str_remove("^@[a-z]*\\{") |>
  str_remove(",")

# 4. Convert the whole vector into a table, with one row per entry
get_bibitem <- function(data = bibdata, 
                        n){
  from <- startlines[n]
  to <- endlines[n]
  this_item <- data[from:to] |> 
    paste(collapse = "\n")
  
  return(this_item)
}

bib_vec <- c()
for (i in 1:length(citekeys)) {
  bib_vec <- c(bib_vec, 
               get_bibitem(n=i))
}

bib_df <- data.frame(key = unname(citekeys),
                     item = bib_vec)

# 5. Extract entrytype and tidy entries
bib_df_old <- bib_df
bib_df <- bib_df_old |> 
  mutate(entrytype = item |> 
           {\(x) 
             gsub("@([a-z]+).*", 
                  "\\1", x)}(),
         item = item |> 
           {\(x) 
             gsub("@[a-z]+\\{[a-zA-Z0-9:-]+,[\n](.*)", 
                  "\\1", x)}() |> 
           {\(x) 
             gsub(".{1}$", "", x)}()) |> 
  select(key, entrytype, item)

# 6. Parse and unnest items, then pivot wider to columns
bib_df_old2 <- bib_df
bib_df <- bib_df_old2 |> 
  mutate(field = item |> 
           {\(x) gsub("\t", "", x)}() |> 
           strsplit("\n")) |> 
  select(key, entrytype, field) |> 
  unnest(field) |> 
  mutate(name = field |> 
# next line presumes a spaced equal sign is always used between fieldname and data
           strsplit(" = ") |> 
           sapply(`[`, 1),
         value = field |> 
           strsplit(" = ") |> 
           sapply(`[`, 2) |> 
           {\(x) 
             gsub(",$", "", x)}()) |>
  arrange(name) |> 
  select(-field) |> 
  pivot_wider(id_cols = c(key, entrytype))

The code assumes some peculiarities of the bib file: for example, that field names and data are separated by spaced equal-signs " = ", etc., but the code and comments should make it flexible enough to adapt.