# setwd("C:/Users/nwickelmaier/Nextcloud/Documents/MDS/2023ss/60100_master_thesis/data/ContentEyevisit/eyevisit_cards_light")
rm(list=ls())

dat0 <- read.table("../../event_logfiles.csv", sep = ";", header = TRUE)
dat0$artwork <- sprintf("%03d", dat0$artwork)

# artwork names
artworks <- sort(unique(dat0$artwork))

# create data frame with file names and topics for each artwork

dat <- NULL
file_order <- NULL

for (artwork in artworks) {
  fnames <- dir(pattern = paste0(artwork, "_"), path = artwork, full.names = TRUE)
  topic <- NULL
  for (fname in fnames) {
    topic <- c(topic, gsub("^<card type=.(.*).>$", "\\1",
                           grep("^<card type=", trimws(readLines(fname)), value = T)))
    
  }
  index <- paste(artwork, "index.xml", sep = "/")
  file_order <- c(file_order, gsub("^<card src=.(.*)./>$", "\\1",
                       grep("^<card src=", trimws(readLines(index)), value = TRUE)))
  in_index <- fnames %in% file_order
  dat <- rbind(dat, data.frame(artwork, file_name = fnames, in_index, topic))
}

table(dat$artwork)
table(dat$topic)

# take only the ones that are actually displayed and sort in the same order
# as indicated in index.html

dat2 <- dat[dat$in_index, -3]
dat2 <- dat2[order(file_order, dat2$file_name), ]

dat2$index <- unlist(sapply(table(dat2$artwork), seq_len))

write.table(dat2, file = "../../topics.csv", sep = ";", row.names = FALSE)