R/make_history.R

In choisy/dictionary: Dictionary of South East Asia Admin1, Admin2 and Admin3 Names

# Splits the vector by `EVENT`
split_event <- function(vect){
  vect <- gsub("splits", "split", vect)
  vect <- gsub("renames", "rename", vect)
  vect <- gsub("merges", "merge", vect)
  strsplit(vect, split = "split|merge|rename|complex split|complex merge")
}

# ------------------------------------------------------------------------------
# Function to translate the Vietnase admin1 names in UNICODE and English.
translate_v <- function(vect, hash) {
  vect <- gsub(" in ", "", vect)
  vect <- gsub("\\.", "", vect)
  vect <- trimws(vect)
  translate(vect, hash)
}

# ------------------------------------------------------------------------------
# Function to identify the event and return a vector of character
id_event <- function(vect) {
  ifelse(agrepl("complex split", vect), "complex split",
         ifelse(agrepl("complex merge", vect), "complex merge",
                ifelse(agrepl("split", vect), "split",
                       ifelse(agrepl("merge", vect), "merge", "rename"))))
}

# ------------------------------------------------------------------------------
# Function to identify the date and return a vector of character
id_date <- function(vect) {
  vect <- gsub("[^[:digit:]]", "", vect)
  vect <- paste0(vect, "-01-01")
  vect <- as.character(as.Date(vect))
}

# ------------------------------------------------------------------------------
# Function to identify the admin1 name and return a vector of character,
# the extractor parametes permit to select the admin1 names before (1) or
# after (2) the event. Hash for the translation of the admin1 name in a
# standardized format
id_admin1 <- function(vect, extractor, hash) {
  lst <- lapply(split_event(vect), "[", extractor)
  lst <- strsplit(unlist(lst), ";")
  lst <- lapply(lst, function(x) {
    sel <- regexpr("[.*(?=(\\(.+\\)))|.*]", x)
    if (any(grepl("-1", sel)))
      sel <- replace(sel, grep("-1", sel), max(nchar(x)) + 1)
    substr(x, 1, sel - 1)
  })
  lst <- lapply(lst, function(x) unlist(strsplit(x, ", ")))
  admin1 <- lst
  admin1 <- lapply(admin1, function(x) x[which(!grepl("In ", x))])
  admin1 <- lapply(admin1, translate_v, hash)
  admin1 <- lapply(admin1, as.list)
}

# ------------------------------------------------------------------------------
# Function to identify the admin2s name and return a vector of character,
# the extractor parametes permit to select the admin2 names before (1) or
# after (2) the event. Hash for the translation of the admin2 name in a
# standardized format
id_admin2 <- function(vect, extractor, hash_p, hash_d) {
  lst <- lapply(split_event(vect), "[", extractor)
  lst <- strsplit(unlist(lst), ";|[[:digit:]],")
  lst <- unlist(lst)
  lst <- lst[which(lst != "" & !grepl("In ", lst))]
  lst <- data.frame(admin1 = lst, stringsAsFactors = FALSE)

  sel <- regexec("\\(.+\\)", lst$admin1)
  sel <- replace(sel, grep(-1, sel), max(nchar(lst$admin1)))

  admin2 <- substr(lst$admin1, sel, max(nchar(lst$admin1)))
  admin2 <- gsub("\\(|\\)", "", admin2)
  admin2 <- strsplit(admin2, ", ")
  admin2 <- lapply(admin2, function(x) translate_v(x, hash_d))
  test <- grepl(0, sapply(admin2, length))
  if (any(test)) admin2[test] <- NA

  admin1 <- substr(lst$admin1, 1, unlist(sel) - 1)
  admin1 <- gsub(" in ", "", admin1)
  admin1 <- lapply(admin1, function(x) translate_v(x, hash_p))
  admin1 <- rep(admin1, lapply(admin2, length))

  df <- data.frame(admin1 = unlist(admin1), admin2 = unlist(admin2))
  list(df)
}

# ------------------------------------------------------------------------------
#' Creates a list of event for a country
#'
#' From a text file (see prerequisite), make a list of a list of 4 or 6 elements
#' for each event (see Details).
#'
#' From a text file (see prerequisite), make a list of event (list of 4 or 6
#' elements):
#' \enumerate{
#'    \item 'year': date of event in character,
#'    \item 'event': character either split, merge or rename,
#'    \item 'before': name of the admin1(s) before the event in a list and
#'    \item 'after': name of the admin1(s) after the event in a list
#'    \item 'd.before' : name of the admin2s concerned by the event
#' (only for complex event)
#'    \item 'd.after' : name of the admin2s concerned by the event
#' (only for complex event)
#' }
#' \cr\cr
#' In the file .txt, each event should be written in a specific format:
#' \code{"In DATE, ADMIN1(s) (ADMIN2(s)) EVENT in ADMIN1(s)."}. \cr
#' ADMIN1(s): if multiple, separated by ";" \cr
#' If the details of the split/merge event is available at admin2 level, the
#' admin2 should be written in (), separated by "," and the event should be
#' written complex EVENT. \cr
#' DATE: written by year as "YYYY" or can be a full date written as "YYYY-mm_dd"
#' \cr
#' EVENT: choose betwrepreen: split(s)/merge/rename(s) \cr
#' For example: In 1992, Hau Giang splits in Can Tho; Soc Trang. \cr
#' For example: In 2013, Vientiane (Longsan, Xaysomboun, Phun, Hom);
#'                       Xiengkhuang (Thathon) complex splits in Xaisomboun.
#'
#'
#' @param file txt file containing the information (try should be written in a
#'   specific format, see \code{details})
#' @param hash named vector to translate the name of the admin1 geographic unit
#' @param d.hash named vector to translate the name of the admin2 geographic
#'   unit
#'
#' @importFrom utils read.delim unzip
#' @importFrom stats na.omit setNames
#' @export
#' @examples
#' \dontrun{
#'  library(dictionary)
#'  make_history("path/to/file.txt", vn_admin1)
#' }
make_history <-  function(file, hash, d.hash) {

  df <- read.delim(file, header = FALSE, stringsAsFactors = FALSE)
  df[["year"]] <- id_date(df$V1)
  df[["event"]] <- id_event(df$V1)
  df[["before"]] <- id_admin1(df$V1, 1, hash)
  df[["after"]] <- id_admin1(df$V1, 2, hash)

  if (any(grepl("complex", df$event))) {
    df[["d.before"]] <- id_admin2(df$V1, 1, hash, d.hash)
    df[["d.after"]] <- id_admin2(df$V1, 2, hash, d.hash)
  }

  df <- df[, - which(names(df) == "V1")]
  lapply(seq_len(nrow(df)), function(x) as.list(unlist(df[x, ], FALSE)))
}