rob
/
torrents.csv
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Fixing scanning.

This commit is contained in:
Dessalines 2018-11-26 10:08:18 -07:00
parent ad17496f28
commit 8a2d934048
1 changed files with 52 additions and 41 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

85
scripts/scan_torrents.sh
View File

@ -4,6 +4,7 @@
# Help line # Help line
cd .. cd ..
torrents_csv="`pwd`/torrents.csv" torrents_csv="`pwd`/torrents.csv"
scanned_out="`pwd`/infohashes_scanned.txt"
help="Run ./scan_torrents.sh [TORRENTS_DIR] \nor goto https://gitlab.com/dessalines/torrents.csv for more help" help="Run ./scan_torrents.sh [TORRENTS_DIR] \nor goto https://gitlab.com/dessalines/torrents.csv for more help"
if [ "$1" == "-h" ] || [ -z "$1" ]; then if [ "$1" == "-h" ] || [ -z "$1" ]; then
@ -27,60 +28,70 @@ fi
pushd $torrents_dir pushd $torrents_dir
for torrent_file in *.torrent; do for torrent_file in *.torrent; do
# Scrape it file_infohash=$(basename $torrent_file | cut -d'.' -f 1)
health_text=$(torrent-tracker-health --torrent $torrent_file --timeout 1000 --addTrackers={udp://tracker.coppersurfer.tk:6969/announce,udp://tracker.internetwarriors.net:1337/announce,udp://tracker.opentrackr.org:1337/announce,udp://exodus.desync.com:6969/announce,udp://explodie.org:6969/announce})
# echo $health_text if rg -Nq $file_infohash $scanned_out; then
echo "$file_infohash already scanned"
else
infohash=$(jq -r '.hash' <<< $health_text) # Scrape it
name=$(jq -r '.name' <<< $health_text) health_text=$(torrent-tracker-health --torrent $torrent_file --timeout 1000 --addTrackers={udp://tracker.coppersurfer.tk:6969/announce,udp://tracker.internetwarriors.net:1337/announce,udp://tracker.opentrackr.org:1337/announce,udp://exodus.desync.com:6969/announce,udp://explodie.org:6969/announce})
size_bytes=$(jq -r '.length' <<< $health_text)
seeders=$(jq -r '.seeds' <<< $health_text)
leechers=$(jq -r '.peers' <<< $health_text)
completed=$(jq -r '.completed' <<< $health_text)
date_string=$(jq -r '.created' <<< $health_text)
created_date=$(date -d "${date_string}" +"%s")
scraped_date=$(date +%s)
# Construct add line echo -e "$health_text"
add_line="$infohash;$name;$size_bytes;$created_date;$seeders;$leechers;$completed;$scraped_date"
# echo -e $add_line
if (( $seeders > 0 )); then infohash=$(jq -r '.hash' <<< $health_text)
name=$(jq -r '.name' <<< $health_text)
size_bytes=$(jq -r '.length' <<< $health_text)
seeders=$(jq -r '.seeds' <<< $health_text)
leechers=$(jq -r '.peers' <<< $health_text)
completed=$(jq -r '.completed' <<< $health_text)
date_string=$(jq -r '.created' <<< $health_text)
created_date=$(date -d "${date_string}" +"%s")
scraped_date=$(date +%s)
found_line=$(rg -n $infohash $torrents_csv) # Construct add line
found_seeders=$(echo -e $found_line | cut -d';' -f 5) add_line="$infohash;$name;$size_bytes;$created_date;$seeders;$leechers;$completed;$scraped_date"
# echo -e $add_line
# Only re-add if the infohash doesn't exist, or if (( $seeders > 0 )); then
if [ ! -z "$found_line" ]; then
# Seeder counts are different found_line=$(rg -n $infohash $torrents_csv)
if [ "$found_seeders" != "$seeders" ]; then found_seeders=$(echo -e $found_line | cut -d';' -f 5)
# Delete the original infohash line # Only re-add if the infohash doesn't exist, or
grep -v "$infohash" $torrents_csv > torfile.tmp.2; mv torfile.tmp.2 $torrents_csv if [ ! -z "$found_line" ]; then
# Seeder counts are different
if [ "$found_seeders" != "$seeders" ]; then
# Delete the original infohash line
rg -N -v "$infohash" $torrents_csv > torfile.tmp.2; mv torfile.tmp.2 $torrents_csv
# Append the add lines to the torrents.csv file
echo -e "\n$add_line" >> $torrents_csv
# truncate -s -1 $torrents_csv # Removing last newline
echo -e "Updating Seeders: $torrent_file | $name | $infohash | $seeders"
else
echo -e "Not adding $name, had identical seeders"
fi
else
# Append the add lines to the torrents.csv file # Append the add lines to the torrents.csv file
echo -e "\n$add_line" >> $torrents_csv echo -e "\n$add_line" >> $torrents_csv
# truncate -s -1 $torrents_csv # Removing last newline # truncate -s -1 $torrents_csv # Removing last newline
echo -e "Updating Seeders: $torrent_file | $name | $infohash | $seeders" echo -e "New Torrent: $torrent_file | $name | $infohash | $seeders"
else
echo -e "Not adding $name, had identical seeders"
fi fi
else else
# Append the add lines to the torrents.csv file # Deleting the line if it existed
echo -e "\n$add_line" >> $torrents_csv if [ ! -z "$infohash" ]; then
# truncate -s -1 $torrents_csv # Removing last newline # This removes lines that have no seeders
echo -e "New Torrent: $torrent_file | $name | $infohash | $seeders" echo -e "$name has no seeders, removing if existed."
grep -v "$infohash" $torrents_csv > torfile.tmp.2; mv torfile.tmp.2 $torrents_csv
fi
fi fi
echo $infohash >> $scanned_out
else
# Deleting the line if it existed
if [ ! -z "$infohash" ]; then
echo -e "$name has no seeders, removing if existed."
grep -v "$infohash" $torrents_csv > torfile.tmp.2; mv torfile.tmp.2 $torrents_csv
fi
fi fi
done done