From 45bf0d9af96d6f972869341513b2bf7368e2996c Mon Sep 17 00:00:00 2001
From: nwickel <n.wickelmaier@iwm-tuebingen.de>
Date: Fri, 7 Jul 2023 17:18:09 +0200
Subject: [PATCH] Adjusted some things after getting complete set of log files

---
 code/01_parse-logfiles.R | 19 ++++++++++---------
 code/03_specs.R          | 28 ++++++++++++++++------------
 2 files changed, 26 insertions(+), 21 deletions(-)

diff --git a/code/01_parse-logfiles.R b/code/01_parse-logfiles.R
index 84dab52..9084333 100644
--- a/code/01_parse-logfiles.R
+++ b/code/01_parse-logfiles.R
@@ -14,6 +14,8 @@
 
 # setwd("C:/Users/nwickelmaier/Nextcloud/Documents/MDS/2023ss/60100_master_thesis/code")
 
+#+ setup, include = FALSE
+knitr::opts_chunk$set(warning = FALSE, message = FALSE) 
 
 #' # Preprocessing raw log files into data frame
 
@@ -34,17 +36,11 @@
 
 #' Choose which folders with raw log files should be included:
 
-folders <- c(
-             "_2016"
-             , "_2017a"
-             , "_2017b"
-             , "_2018"
-            )
+folders <- "all"
 
-dirpaths <- paste0("../data/HAUM Logs/", folders)
+dirpaths <- paste0("../data/haum_logs_2016-2023/", folders)
 
 fnames <- dir(dirpaths, pattern = "*.log", full.names = TRUE)
-# TODO: Enter all data
 length(fnames)
 head(fnames)
 
@@ -59,6 +55,7 @@ head(dat$logs)
 #   incomplete final line found on '_2016/2016_11_18-11_31_0.log'
 #   incomplete final line found on '_2016/2016_11_18-11_38_30.log'
 #   incomplete final line found on '_2016/2016_11_18-11_40_36.log'
+#   ...
 
 ## --> files have a last line that looks like a binary entry??
 
@@ -74,6 +71,7 @@ d2 <- dim(dat)[1]
 
 #' The files contain `r d1-d2` corrupt lines that were remooved from the data.
 
+
 #' ### Extract relevant infos
 
 date <- sapply(dat$logs, gsub,
@@ -124,9 +122,12 @@ dat$y         <- moves[,2]
 dat$scale     <- moves[,3]
 dat$rotation  <- moves[,4]
 
+dat$logs <- NULL
+# remove original log files from data so file becomes smaller
+
 str(dat)
 
-head(dat[, 3:ncol(dat)], 20)
+head(dat[, 2:ncol(dat)], 20)
 
 ## TODO: Replace artwork and popup numbers with informative strings
 
diff --git a/code/03_specs.R b/code/03_specs.R
index 768e40c..79de837 100644
--- a/code/03_specs.R
+++ b/code/03_specs.R
@@ -14,8 +14,10 @@ dat <- read.table("../data/rawdata_logfiles.csv", header = TRUE, sep = ";")
 #                                           "ShowPopup", "HidePopup"))
 #
 
-#dat$logs <- NULL                  # do not need original log files
 dat$date <- as.POSIXct(dat$date)  # create date object
+dat <- dat[order(dat$date), ]
+
+rownames(dat) <- NULL
 
 str(dat)
 head(dat)
@@ -31,16 +33,16 @@ table(year(dat$date))
 #' Number of events total and per year
 
 table(dat$event)
-# Start Application  Show Application 
-#              1679              1656 
-#   Transform start    Transform stop 
-#           2119815           1944618 
-#         Show Info        Show Front 
-#             71955             71043 
-#  Artwork/OpenCard Artwork/CloseCard 
-#             64990             56750 
-#         ShowPopup         HidePopup 
-#             44070             43813 
+# Artwork/CloseCard  Artwork/OpenCard
+#             80026             92588
+#         HidePopup  Show Application
+#             59932              3179
+#        Show Front         Show Info
+#            106945            108113
+#         ShowPopup Start Application
+#             60350              3210
+#   Transform start    Transform stop
+#           3219109           2966724
 # --> more "openups" than "closes" (not surprisingly)
 
 table(dat$event, year(dat$date))
@@ -58,7 +60,9 @@ length(unique(as.Date(dat$date)))
 table(dat$artwork)
 lattice::barchart(table(dat$artwork))
 
-
+#' ## Corona Gaps
+table(diff(as.Date(dat$date)))
+dat[diff(as.Date(dat$date)) > 50, ]
 
 #' ## Plots from Visualization Tool