mtt_haum/code/01_preprocessing_haum.R

# 01_preprocessing_haum.R
#
# content: (1) Parse raw log files
#          (2) Create event logs
#          (3) Add meta data
#
# input:  raw log files from ../data/haum/*.log
#         ../data/metadata/feiertage.csv
#         ../data/metadata/schulferien_2016-2018_NI.csv
#         ../data/metadata/schulferien_2019-2025_NI.csv
# output: raw_logfiles_<timestamp>.csv
#         event_logfiles_<timestamp>.csv
#         event_logfiles_<timestamp>.csv
#
# last mod: 2023-10-23, NW

# setwd("C:/Users/nwickelmaier/Nextcloud/Documents/MDS/2023ss/60100_master_thesis/code")

#library(mtt)
devtools::load_all("../../../../software/mtt")

now <- format(Sys.time(), "%Y-%m-%d_%H-%M-%S")
#now <- "2023-09-23_01-31-30"

#--------------- (1) Parse raw log files ---------------

#path <- "../data/haum/LogFiles/"
#folders <- dir(path)
#folders <- "2016"

#datraw <- parse_logfiles(folders, path)

datraw <- read.table("results/haum/raw_logfiles_2023-10-25_16-20-45.csv",
                     sep = ";", header = TRUE)

## Export data

#write.table(datraw, paste0("results/haum/raw_logfiles_small_", now, ".csv"),
#            sep = ";", row.names = FALSE)

#--------------- (2) Create event logs ---------------

datlogs <- create_eventlogs(datraw,
           xmlpath = "../data/haum/ContentEyevisit/eyevisit_cards_light/",
           glossar = TRUE, save = TRUE)

artworks <- unique(datlogs$artwork)
topics <- extract_topics(artworks, xmlfiles = paste0(artworks, ".xml"),
                         xmlpath = "../data/haum/ContentEyevisit/eyevisit_cards_light/")

datlogs_topics <- add_topic(datlogs, topics = topics)

#--------------- (3) Add meta data ---------------

## Read data for holiday

hd0 <- read.table("../data/metadata/feiertage.csv", sep = ";", header = TRUE)
hd0$X.br. <- NULL

hd <- hd0[hd0$Abkuerzung == "NI", ]
names(hd) <- c("state", "stateCode", "date", "holiday")
hd$date <- as.POSIXct(hd$date)

## Read data for school vacations

# https://ferien-api.de/#holidaysPerStateAndYear
# Data extracted (on Linux) via:
# curl https://ferien-api.de/api/v1/holidays/NI > schulferien_NI.json

# library(jsonlite)
#
# dat <- read_json("data/metadata/schulferien_NI.json", simplify = TRUE)
# dat$slug <- NULL
#
# dat$name <- paste0(gsub("^(.*).niedersachsen.*", "\\1", dat$name),
#                    gsub("^.*niedersachsen [0-9]{4}(.*)", "\\1",
#                               dat$name))
#
# write.table(dat, "data/metadata/schulferien_2019-2025_NI.csv", sep = ";",
#             row.names = FALSE, quote = FALSE)

sf1 <- read.table("../data/metadata/schulferien_2016-2018_NI.csv", sep = ";",
                  header = TRUE)
sf2 <- read.table("../data/metadata/schulferien_2019-2025_NI.csv", sep = ";",
                  header = TRUE)
sf <- rbind(sf1, sf2)
sf$start <- as.Date(sf$start)
sf$end <- as.Date(sf$end)

sfdat <- NULL

for (i in seq_len(nrow(sf))) {
  date <- seq(sf$start[i], sf$end[i], by = 1)
  sfdat <- rbind(sfdat, data.frame(date, vacations = sf$name[i],
                                   stateCodeVacations = sf$stateCode[i]))
}

# TODO: How to handle stateCode? There will be several for certain types of
# data sets... Not important here, since I only do NI.

## Merge data

datlogs_topics$date <- as.Date(datlogs_topics$date.start)

dat1 <- merge(datlogs_topics, hd, by.x = "date", by.y = "date", all.x = TRUE)
dat2 <- merge(dat1, sfdat, by.x = "date", by.y = "date", all.x = TRUE)

## Export data

write.table(dat2, paste0("results/haum/event_logfiles_glossar_", now, ".csv"),
            sep = ";", row.names = FALSE)

# TODO: Maybe add infos about artworks?