library("readxl")
library("writexl")
library("dplyr")
library("xlsx")
library("stringi")
library("SPARQL")
library("stringr")
library("tictoc")
library("rlang")

#Funktionen einbinden
source("sparql.R")
source("config/config.R")


### Einlesen ###
inDB <-  get_current_geoera_lit_db()

to_replace_original <- read_excel(paste(PFAD_EXCEL, FILENAME_EXCEL, sep=""))
to_replace_done <- to_replace_original

###


#Schleife über den Spaltenvektor aus der Excel
for (spalte_excel in SPALTEN[1]:SPALTEN[length(SPALTEN)]) { 
  
  #Start der Zeitmessung für aktuelle Spalte
  tic(paste("starte mit Spalte",spalte_excel))
  
  for (zeile_excel in 1:nrow(to_replace_original[, spalte_excel])) {
    
    current_excel <- tolower(to_replace_original[zeile_excel, spalte_excel])
    
    if (is.na(current_excel)) {
      next
    }
    
    if ((nchar(current_excel) < 20)) {
      to_replace_done[zeile_excel, spalte_excel] <- ""
      next
    }
    
    excel_search_all <- unlist(strsplit(gsub("[^[:alnum:] ]", "", current_excel), " +"))
    excel_search_numbers <- unlist(str_extract_all(current_excel, "(?:19|20)\\d{2}"))
  
    
    for (zeile_db in 1:nrow(inDB[, INHALTE_DB])) {
      
      current_db <- tolower(inDB[zeile_db,INHALTE_DB])
      
      db_search_all <- unlist(strsplit(gsub("[^[:alnum:] ]", "", current_db), " +"))
      db_search_numbers <- unlist(str_extract_all(current_db, "(?:19|20)\\d{2}"))
      
      count_all <- 0
      count_numbers <- 0
      percent_match_numbers <- 0.1
      
      for (k in 1:length(excel_search_all)) {
        if (excel_search_all[k] %in% db_search_all) {
          
          count_all <- count_all + 1
        }
      }
       
      if ((length(db_search_numbers) > 0) && (length(excel_search_numbers) > 0)) {
        for (l in 1:length(excel_search_numbers)) {
          if (excel_search_numbers[l] %in% db_search_numbers) {
            
            count_numbers <- count_numbers + 1
          }
        }
      }
      
      percent_match_all <- count_all / length(excel_search_all)
      
      if (length(excel_search_numbers) > 0) {
        percent_match_numbers <- count_numbers / length(excel_search_numbers)
      }
      
      if ((percent_match_all > 0.80) && (percent_match_numbers > 0.99)) {
        
        to_replace_done[zeile_excel, spalte_excel] <- str_remove_all(inDB[zeile_db,ID], "[<>]")
      }
      
    }
  }
  toc()
}

toc()

pfad_output <- paste(PFAD_OUT, "replaced_", format(Sys.time(), "%Y_%m_%d_%H%M%S"), FILENAME_EXCEL, sep="")
write_xlsx(to_replace_done, pfad_output)