I’ve baked some stuff together which might help some. Is it complete, probably not. Can it be improved, most likely.
One bash script which reads every new entry in the Icecast log file.
One service ctl script which now let’s the batch file be a service.
One Awk Script that makes every entry uniq, adding up the number of listeners and the number of times it played.
#!/bin/bash
Service_name=“Your station Name”
Transmission_category=“B” # determined by SoundExchange in our case it is B
week=date +%U
#set -x
logFile=./IcecastLogs/playlist.log
# Function to process each line
process_line() {
line=“$1”
# Split the line by the '|' delimiter
IFS='|' read -ra parts <<< "$line"
# Extract date and time from the first part
datetime_part=${parts[0]}
stream_name=${parts[1]}
listeners=${parts[2]}
track_info=${parts[3]}
# Further split the datetime part
IFS=':' read -ra date_time <<< "$datetime_part"
date_part=${date_time[0]}
time_part=(${date_time[@]:1})
# Extract the day, month, year from the date part
IFS='/' read -ra date_parts <<< "$date_part"
day=${date_parts[0]}
month=${date_parts[1]}
year=${date_parts[2]}
# Extract hour, minute
hour=${time_part[0]}
minute=${time_part[1]}
# Split the track_info into show name, performer, and song title
track_info=`echo "${track_info// - /-}"`
IFS='-' read -ra track_parts <<< "$track_info"
show_name=${track_parts[0]}
performer=${track_parts[1]}
song_title=${track_parts[2]}
# Escape single quotes in the variables
performer=$(echo "$performer" | sed "s/'/''/g")
song_title=$(echo "$song_title" | sed "s/'/''/g")
album_name=$(docker exec libretime-postgres psql -U XXXXXX -q -t -c "select album_title from cc_files WHERE artist_name='$performer' and track_title='$song_title';" --pset footer=off | tail -2 | head -1 | cut -c2-) 2> /dev/null
check_label=$(docker exec libretime-postgres psql -U XXXXXX -q -t -c "select label from cc_files WHERE artist_name='$performer' and track_title='$song_title';" --pset footer=off | tail -2 | head -1 | cut -c2-) 2> /dev/null
#==============================================================================
if [ -z "$check_label" ]
then
jsondata=$(curl -s -G \
--url "https://api.discogs.com/database/search" \
--data-urlencode "artist=$performer" \
--data-urlencode "track=$song_title" \
--data-urlencode "release_title=$album_name" \
--data-urlencode "type=release" \
--data-urlencode "format=CD" \
-H "Authorization: Discogs token=XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX" \
-H "User-Agent: EWCMI_Radio/1.0") 2> /dev/null
label=`echo $jsondata | jq -r '.results[0].label[0]'`
fi
if [ -z "$check_label" ] && [ "$label" != "null" ]
then
lab=$label
perf=$performer
song=$song_title
album=$album_name
label=$(echo "$label" | sed "s/'/''/g")
performer=$(echo "$performer" | sed "s/'/''/g")
song_title=$(echo "$song_title" | sed "s/'/''/g")
album_name=$(echo "$album_name" | sed "s/'/''/g")
docker exec libretime-postgres psql -U XXXXXX -t -c \"update cc_files set label='$label' WHERE artist_name='$performer' and track_title='$song_title' and album_title='$album_name';\"
else
label=$check_label
fi
#==============================================================================
# Get missing info: Publisher, icrs
# search on artist, recording(track title),
# Print the extracted variables (or use them as needed in your script)
LogLine="$Service_name|$Transmission_category|$performer|$song_title|ISRC|$album_name|$label|$listeners|(aggregate Tuning Hours)|$show_name|Total Plays"
LogLine=`echo $LogLine | sed "s/''/'/g"`
LogLineTest=$(echo "$LogLine" | tr '[:upper:]' '[:lower:]' | grep -iE "jingle|promo|---|legal|need him")
if [ -z "$LogLineTest" ]
then
echo "$LogLine" >> ./SoundExchange/$year-$month-$week.raw
fi
}
# Start monitoring the log file
tail -Fn0 “$logFile” | while read -r line; do
test=echo $line | grep offline
if [ -z “$test” ]
then
process_line “$line”
fi
done
The AWK Script to process the raw data log file
cat SoundExchangeFormat.sh
#!/bin/bash
#ls | grep -v SoundExchangeFormat.sh |cut -c10-11 | sort
Path=“/xxxxxxxxx/xxxx/SoundExchange”
declare -i Week
Year=date +%Y
#Month=date +%b
Week=date +%U
Find the first day of the specified week
first_day_of_week=$(date -d “$Year-01-01 +$((Week - 1)) weeks” +%Y-%m-%d)
Extract the month from that date
month=$(date -d “$first_day_of_week” +%B)
Week=$Week-1
Month=echo $month | cut -c1-3
#set -x
if [ -f “$Path/$Year”-“$Month”-“$Week”“SE.csv” ]
then
#ls “$Path/$Year”-“$Month”-“$Week”“SE.csv”
exit
fi
Define the log file
logfile=$Path/$Year"-“$Month”-“$Week”.raw"
Define the log file
date=echo $1 | awk -F"." '{print $1}'
#set -x
Process the log file
awk -F ‘|’ ’
{
# Create a key using fields 3 (Performer), 4 (Album), 6 (Title), and 7 (Label)
key = $3 FS $4 FS $6 FS $7
# Sum the listeners for this key
listeners[key] += $8
# Count the number of plays for this key
plays[key]++
# Store the first occurrence of each unique entry
if (!(key in first_occurrence)) {
first_occurrence[key] = $0
}
}
END {
# Reprint unique log entries with updated Total Plays and listener count
for (k in first_occurrence) {
# Update the Total Plays and Listener count in the original entry
split(first_occurrence[k], fields, FS)
fields[8] = listeners[k] # Update number of listeners
fields[11] = plays[k] # Update total plays
# Print the updated entry
print fields[1] FS fields[2] FS fields[3] FS fields[4] FS fields[5] FS fields[6] FS fields[7] FS fields[8] FS fields[9] FS fields[10] FS fields[11]
}
}
’ “$logfile” > “$Path/$Year”-“$Month”-“$Week”_“SE.csv”
#echo “Unique log file has been created with updated Total Plays and listener counts.”
The SoundExchangeformat.sh is executed each Monday morning from Cron
0 2 * * 1 /xxxxxxxxx/xxxx/SoundExchange/SoundExchangeFormat.sh > /dev/null 2>&1
Starting tracklog.sh as a service
/etc/systemd/system/tracklog.service
cat /etc/systemd/system/tracklog.service
[Unit]
Description=Tracklog Bash Script
After=network.target
[Service]
ExecStart=/home/container/libretime/tracklog.bash
Restart=always
RestartSec=5
User=root
WorkingDirectory=/home/container/libretime
StandardOutput=journal
StandardError=journal
#Environment=YOUR_ENV_VAR=value
[Install]
WantedBy=multi-user.target
Hopefully this might help some people, or instigate the creation of the log file from within Libretime.
Chris.
