R/parse_dje.R

In courtsbr/dje: A scraper that downloads DJEs

parse_dje_tjms <- function(text_file){
  txt <- readr::read_file(text_file)

  clean_text <- txt %>%
    remove_header() %>%
    remove_footer()

  sumario <- clean_text %>%
    find_index() %>%
    stringr::str_split("(SUM\u00c1RIO|\b)\n") %>%
    dplyr::first() %>%
    stringr::str_subset("Juizado|Vara") %>%
    stringr::str_trim()

  courts <- sumario %>%
    paste0("\n", ., "\n") %>%
    break_text(clean_text, .) %>%
    append(-1)

  d <- purrr::map(1:(length(courts)-1),
                  ~stringr::str_sub(clean_text, courts[.x], courts[.x+1])) %>%
    purrr::map(~break_chunk(.x) %>%
                 dplyr::first() %>%
                 purrr::map_df(parse_docket_entry)) %>%
    purrr::map2(names(courts)[-length(courts)],
                ~dplyr::mutate(.x, location = .y)) %>%
    dplyr::bind_rows() %>%
    dplyr::filter(stringr::str_detect(lawsuit_id, "Processo"))
}


#' Parse DJE file
#'
#' parse TJSP file obtained from [dje_to_text()].
#'
#' @param text_file path of the file to parse.
#'
#' @return tibble containing lawsuit and court information
#' @export
parse_dje_tjsp <- function(text_file) {

  # clean_text <- text_file %>%
  #   remove_header() %>%
  #   remove_footer()
  txt <- readr::read_file(text_file)
  cnj_format_sp <- stringr::regex("[0-9]{7}\\-[0-9]{2}\\.[0-9]{4}\\.8\\.26\\.[0-9]{4}")
  clean_text <- paste0(txt, "@fim_do_texto@")

  classify_content <- function(raw_content) {
    dplyr::case_when(
      stringr::str_detect(raw_content, "Distribuidor") ~ "D",
      stringr::str_detect(raw_content, "[XVI ]+-|F\u00f3rum|^[\u00c7A-Z\u00c3\u00c2\u00c1\u00cd\u00d3\u00da\u00c9\u00ca ]+$") ~ "C",
      TRUE ~ "Outros")
  }

  index <- clean_text %>%
    find_index(index_regex = "\nSUM\u00c1RIO\nTRIBUNAL") %>%
    stringr::str_split("SUM\u00c1RIO|\n") %>%
    dplyr::first() %>%
    stringr::str_trim() %>%
    stringr::str_subset("^[XVI ]+-|F\u00f3rum|^[\u00c7A-Z\u00c3\u00c2\u00c1\u00cd\u00d3\u00da\u00c9\u00ca -]+$|Vara|Anexo|Distribuidor|Juizado|Fiscais|Criminal|C\u00edvel") %>%
    tibble::as.tibble() %>%
    purrr::set_names("valor") %>%
    dplyr::mutate(tipo = classify_content(valor))

  clean_breaks <- function(x, y) {
    if (stringr::str_detect(x[[1]][length(x[[1]])], "Distribuidor")) {
      tab2 <- tibble::tibble(valor = ifelse(is.na(y), "@fim_do_texto@", y))
      x <- dplyr::bind_rows(x, tab2)
    } else {
      x
    }
  }

  breaks <- index %>%
    dplyr::filter((tipo == "D") |
                    (dplyr::lag(tipo) == "D") |
                    (tipo == "C" & dplyr::lead(tipo) != "C")) %>%
    dplyr::filter(is.na(dplyr::lead(tipo)) |
                    !(tipo == "C" & dplyr::lead(tipo) == "C")) %>%
    dplyr::mutate(classe = ifelse(tipo == "C" & !is.na(dplyr::lead(tipo)),
                                  valor, NA),
                  valor = paste0("\n", valor, "\n")) %>%
    tidyr::fill(classe) %>%
    dplyr::mutate(classe = paste0("\n", classe, "\n")) %>%
    dplyr::select(classe, valor) %>%
    dplyr::group_by(classe) %>%
    tidyr::nest(.key = "valor") %>%
    dplyr::mutate(valor = purrr::map2(valor, dplyr::lead(classe), clean_breaks))

  breaks_counties <- clean_text %>%
    break_text(breaks$classe) %>%
    append(-1)

  cut_text <- function(x, y) {

    if (breaks_counties[x+1] == -1){
      text <- clean_text %>%
        stringr::str_sub(breaks_counties[x], breaks_counties[x+1])
    } else {
      text <- clean_text %>%
        stringr::str_sub(breaks_counties[x],
                         breaks_counties[x + 1] + nchar(breaks_counties[x + 1]) - 1)
    }

    # print(y$valor[1])
    if(length(y$valor) == 3) {
      lim_inf <- text %>%
        break_text(y$valor[2])
      points <- text %>%
        stringr::str_sub(lim_inf, -1) %>%
        break_text(y$valor[2:3])
      r <- stringr::str_sub(stringr::str_sub(text, lim_inf, -1), points[1], points[2])
    } else {
      distribuidor <- which(stringr::str_detect(y$valor, "Distribuidor"))
      r <- character(length(distribuidor))
      for(i in 1:length(distribuidor)) {
        if(i == 1){
          points <- text %>%
            break_text(y$valor[distribuidor[1]:(distribuidor[1] + 1)])
          r[i] <- stringr::str_sub(text, points[1], points[2])
        } else {
          lim_inf <- text %>%
            break_text(y$valor[distribuidor[i]])
          points <- text %>%
            stringr::str_sub(lim_inf, -1) %>%
            break_text(y$valor[distribuidor[i]:(distribuidor[i]+1)])
          r[i] <- stringr::str_sub(stringr::str_sub(text, lim_inf, -1),
                                   points[1], points[2])
        }
      }
      # points1 <- text %>%
      #   break_text(y$valor[distribuidor[1]:(distribuidor[1]+1)])
      #
      # lim_inf <- text %>%
      #   break_text(y$valor[distribuidor[2]])
      #
      # points2 <- text %>%
      #   stringr::str_sub(lim_inf, -1) %>%
      #   break_text(y$valor[distribuidor[2]:(distribuidor[2]+1)])
      #
      # c(stringr::str_sub(text, points1[1], points1[2]),
      #   stringr::str_sub(stringr::str_sub(text, lim_inf, -1), points2[1], points2[2]))
      r
    }
  }

  inner_breaks <- purrr::map2(seq_along(breaks$classe), breaks$valor, cut_text) %>%
    unlist() %>%
    stringr::str_split(pattern = stringr::regex(paste0("(?=PROCESSO ?:",
                                                       cnj_format_sp, ")"),
                                                ignore_case = TRUE))

  d <- breaks %>%
    tidyr::unnest(valor) %>%
    dplyr::filter(stringr::str_detect(valor, "Distribuidor"))  %>%
    dplyr::mutate(processos = inner_breaks) %>%
    tidyr::unnest(processos)
  d
}