torrents.csv/scripts/prune.sh

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

cp $torrents_csv $torrents_csv_tmp

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

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

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

# 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_tmp $torrents_csv_tmp

# Remove dups, keeping the last ones
sort -r -t';' -k1,1 -o $torrents_csv_tmp $torrents_csv_tmp
sort -r -u -t';' -k1,1 -o $torrents_csv_tmp $torrents_csv_tmp
# sort -u -t';' -k2,2 -k8,8 -o $torrents_csv_tmp $torrents_csv_tmp


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

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

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

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

mv $torrents_csv_tmp $torrents_csv

echo "Pruning done."