Plot graph using shell for total file count and date created - linux

Question

I want to create a histogram with total file count intervals of 50 on Y-axis and time created in weeks on X-axis (i.e if new files were created between week 1 and 2 and so on)
Something like
200, 150, 100 , 50 files created during a certain week
7, 14, 21, 28 days on Y-axis. Kind of lost on how to implement this. Any help is appreciated
Update: I am trying along these lines
find <dirname> -type f -ctime -1 -ctime -7 | wc -l
find <dirname> -type f -ctime +7 -ctime -14 | wc -l
Find the max number and use this as my X-axis upper limit. Then divide this number into equal intervals to plot my X-axis

Answer 1

This is a start using GNU awk for time functions (untested since you didn't provide concise, testable sample input that we could test against):
find "$1" -type f -printf '%T# %p\0' |
awk -v RS='\0' '
BEGIN {
nowSecs = systime()
}
{
fileName = gensub(/\S+\s+/,"",1)
fileModSecs = int($1)
fileAgeSecs = nowSecs - fileModSecs
fileAgeDays = int(fileAgeSecs / (24 * 60 * 60))
fileAgeWeeks = int(fileAgeDays / 7)
weekNr = fileAgeWeeks + 1
fileCnts[weekNr]++
numWeeks = (weekNr > numWeeks ? weekNr : numWeeks)
maxFileCnt = (fileCnts[weekNr] > maxFileCnt ? fileCnts[weekNr] : maxFileCnt)
print nowSecs, fileModSecs, fileAgeSecs, fileAgeDays, fileAgeWeeks, weekNr, fileName | "cat>&2"
}
END {
for (fileCnt=maxFileCnt; fileCnt>0; fileCnt--) {
for (weekNr=1; weekNr<=numWeeks; weekNr++) {
if (weekNr in fileCnts) {
char[weekNr] = "*"
}
printf "%s%s", char[weekNr], (weekNr<numWeeks ? OFS : ORS)
}
}
for (weekNr=1; weekNr<=numWeeks; weekNr++) {
printf "%s%s", weekNr, (weekNr<numWeeks ? OFS : ORS)
}
}
'
You need to figure out the details of the loops in the END section for printing the histogram but the above at least shows you how to get the count of files by week without calling find multiple times and hard-coding the number of days week by week.

Answer 2

Apologies being ksh instead of bash (bash level is near echo "Hello World") :)...
Would that do what you need ?
#!/bin/ksh
######################################
#
# statDirReport.sh
#
version="1.0"
# Andre Gelinas, 2018
#
######################################
#############
# Variables
#############
typeset -F2 SCALE
# Max value of X
X_SCALE=30
#############
# Main
#############
if [[ -n $1 ]]; then
DIRNAME=$1
else
print -n "Enter full path to stat : "; read DIRNAME
fi
if [[ ! -d $DIRNAME || ! -r $DIRNAME || ! -x $DIRNAME ]]; then
print "ERROR - Directory unusable - Exiting"
exit
fi
## Getting the data
CTIME1=1
CTIME2=0
for ((i=1;i<=4;i++)); do
CTIME2=$(($i*7))
FILE_COUNT[$i]=$(find $DIRNAME -type f -ctime +$CTIME1 -ctime -$CTIME2 | wc -l)
#To find late on the max amount
F_COUNT[${FILE_COUNT[$i]}]=${FILE_COUNT[$i]}
#
CTIME1=$CTIME2
done
#Doing some math
## Highest number of file
MAX_COUNT=${F_COUNT[-1]}
## Find the value of each tick
SCALE=$(($MAX_COUNT/$X_SCALE))
## Find the real length of the histogram for each week
## having the highest amount using full x scale (integer mathematics)
for ((i=1;i<=4;i++)); do
DATA_2_SCALE[$i]=$(((${FILE_COUNT[$i]}*$X_SCALE)/$MAX_COUNT))
done
# Getting the report
typeset -L2 Col1
typeset -L1 Col2
typeset -L$(($X_SCALE+5)) Col3
typeset -L5 Col4
Col1="Wk"
Col2=" "
Col3="Data"
Col4="Real"
clear
print "statDirReport v$version\tScale is #=$SCALE\n"
print "$Col1$Col2$Col3$Col4\n"
for ((i=1;i<=4;i++)); do
Col1=$i
Col2="|"
graph=""
Col4=${FILE_COUNT[$i]}
for ((j=1;j<=${DATA_2_SCALE[$i]};j++)); do
graph+="#"
done
Col3=$graph
print "$Col1$Col2$Col3$Col4"
done
Edit to modify to add dates as title for the histograms. Modify the last part, right after the "DATA_2_SCALE" loop, with :
#Setting the title of each histogram
## Finding how many sec since the beginning of time
TODAY_SEC=$(date +"%s")
## Finding real date for find range
SEC_PER_DAY=86400
lastDate=$(date -u -d #"$TODAY_SEC" +"%m/%d")
for ((i=1;i<=4;i++)); do
firstDate=$(date -u -d #"$(($TODAY_SEC-(7*$i*$SEC_PER_DAY)))" +"%m/%d")
WEEK[$i]=$firstDate" to "$lastDate" "
lastDate=$firstDate
done
# Getting the report
typeset -L15 Col1
typeset -L1 Col2
typeset -L$(($X_SCALE+5)) Col3
typeset -L5 Col4
Col1="Wk"
Col2=" "
Col3="Data"
Col4="Real"
clear
print "statDirReport v$version\tScale is #=$SCALE\n"
print "$Col1$Col2$Col3$Col4\n"
for ((i=1;i<=4;i++)); do
Col1=${WEEK[$i]}
Col2="|"
graph=""
Col4=${FILE_COUNT[$i]}
for ((j=1;j<=${DATA_2_SCALE[$i]};j++)); do
graph+="#"
done
Col3=$graph
print "$Col1$Col2$Col3$Col4"
done

Answer 3

Using feedgnuplot on a home directory:
dirname=~
e=0
for f in `seq 7 7 28` ; do
find "${dirname}" -type f -ctime +$e -ctime -$f | wc -l
e=$f
done 2> /dev/null |
feedgnuplot --terminal 'dumb 50,15' --with boxes --unset grid --exit
Output:
5500 +-+-----+-------+------+-------+-----+-+
5000 +-+ ********* + + + +-+
4500 +-+ * * ******** +-+
4000 +-+ * * * * +-+
3500 +-+ * * * * +-+
3000 +-+ * * * * +-+
2500 +-+ * ********* * +-+
2000 +-+ * * * * +-+
1500 +-+ * * * * +-+
1000 +-+ * + * + * + ********* +-+
500 +-+-********************************-+-+
0 1 2 3 4 5