Load the Raw Data using PDFTOOLS

Code

raw <- pdftools::pdf_text("homo_deus_chapter_1.pdf")

Extract the Manuscript

Code

manuscript <- paste(raw, collapse = "")

Split the manuscript into section heading and text.

Code

section_text <- dplyr::tibble(manuscript) |> 
  dplyr::mutate(section = stringr::str_split(manuscript, "\n\n\n")) |> 
  tidyr::unnest(cols = "section") |> 
  dplyr::select(-manuscript) |> 
  dplyr::filter(!stringr::str_detect(section, "\n\n")) |> 
  dplyr::slice(3:dplyr::n()) |> 
  dplyr::mutate(section = stringr::str_trim(section)) |> 
  tidyr::separate(section, c("section", "text"), "\n", extra = "merge") |> 
  dplyr::mutate(section_id = 1:dplyr::n())

Lexrankr

Code

lex_top_3 <- seq_along(section_text$text) |> 
  purrr::map_df(~{
    lexRankr::lexRank(
      section_text$text[[.x]],
      docId = .x, 
      n = 5, 
      continuous = TRUE, 
      Verbose = FALSE
    ) }) |> 
  tibble::as_tibble() |> 
  dplyr::mutate(docId = as.numeric(docId))

Results

Code

dplyr::left_join(section_text, lex_top_3, by = c("section_id" = "docId")) |> 
  dplyr::select(section_id, section, sentenceId, sentence, value) |> 
  DT::datatable(options = list(pageLength = 5))