# This prunes torrents.csv, removing those with too many columns, and sorts it
echo "Pruning torrents.csv ..."
cd ..
torrents_csv="`pwd`/torrents.csv"
scanned_out="`pwd`/infohashes_scanned.txt"


# Remove lines that don't have exactly 7 ';'
rg "^([^;]*;){7}[^;]+$" $torrents_csv > tmp_adds
mv tmp_adds $torrents_csv

# Remove random newlines
sed -i '/^$/d' $torrents_csv 

# Extract the header
header=$(head -n1 $torrents_csv) 
sed -i '1d' $torrents_csv

# Sort by seeders desc (so when we remove dups it removes the lower seeder counts)
# TODO this should actually probably do it by scraped date
# sort --field-separator=';' --key=5 -nr -o $torrents_csv $torrents_csv

# Remove dups
sort -u -t';' -k1,1 -o $torrents_csv $torrents_csv
sort -u -t';' -k2,2 -k3,3 -o $torrents_csv $torrents_csv

# Same for the infohashes scanned
sort -u -o $scanned_out $scanned_out

# Remove torrents with zero seeders
awk -F';' '$5>=1' $torrents_csv> tmp
mv tmp $torrents_csv

# Sort by infohash asc
sort --field-separator=';' --key=1 -o $torrents_csv $torrents_csv

# Add the header back in
sed  -i "1i $header" $torrents_csv
#truncate -s -1 $torrents_csv # Removing last newline

echo "Pruning done."