Bunch of fixes to scanning and pruning

This commit is contained in:
Dessalines 2019-01-28 15:01:03 -08:00
parent 2b72758365
commit 70223ebbef
3 changed files with 101 additions and 46 deletions

View File

@ -2,42 +2,46 @@
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_adds
mv tmp_adds $torrents_csv
rg "^([^;]*;){7}[^;]+$" $torrents_csv_tmp > tmp_adds
mv tmp_adds $torrents_csv_tmp
# Remove random newlines
sed -i '/^$/d' $torrents_csv
sed -i '/^$/d' $torrents_csv_tmp
# Extract the header
header=$(head -n1 $torrents_csv)
sed -i '1d' $torrents_csv
header=$(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 $torrents_csv
# 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
# 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
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 $torrents_csv
sort --field-separator=';' --key=1 -o $torrents_csv_tmp $torrents_csv_tmp
# Add the header back in
sed -i "1i $header" $torrents_csv
sed -i "1i $header" $torrents_csv_tmp
#truncate -s -1 $torrents_csv # Removing last newline
mv $torrents_csv_tmp $torrents_csv
echo "Pruning done."

52
scripts/prune_currents.sh Executable file
View File

@ -0,0 +1,52 @@
# This refetches the seeder counts for everthing in torrents.csv, and updates the seeder counts
echo "Refetching seeder counts ..."
cd ..
torrents_csv="`pwd`/torrents.csv"
prune_currents_tmps="`pwd`/prune_currents_tmps"
mkdir $prune_currents_tmps
cd $prune_currents_tmps
cp $torrents_csv tmp
# Extract the header
header=$(head -n1 tmp)
sed -i '1d' tmp
cat tmp | cut -d ';' -f1 > tmp2
mv tmp2 tmp
mkdir prune_currents_tmps
# Split these up into 2000 file batches
split -l 2000 tmp tmp_
> no_seeds
for f in tmp_*; do
echo "Fetching seeds..."
echo $f
torrent-tracker-health --torrent "$f" > health
# Select the infohashes with zero seeders
# append to a no seeds file
jq '.results[] | select(.seeders==0) | .hash' health | tr -d \" >> no_seeds
rm $f
done
# Remove those lines from the file
rg -vwF -f no_seeds $torrents_csv > torrents_removed.csv
rm tmp_*
rm health

View File

@ -6,11 +6,7 @@ cd ..
torrents_csv="`pwd`/torrents.csv"
scanned_out="`pwd`/infohashes_scanned.txt"
tmp_torrent_dir="`pwd`/tmp_torrents"
names_out="`pwd`/names.out"
health_out="`pwd`/health.out"
touch $scanned_out
touch $names_out
touch $health_out
help="Run ./scan_torrents.sh [TORRENTS_DIR] \nor goto https://gitlab.com/dessalines/torrents.csv for more help"
if [ "$1" == "-h" ] || [ -z "$1" ]; then
@ -32,47 +28,50 @@ fi
# Loop over all torrents
pushd $torrents_dir
# for torrent_file in *.torrent; do
# Copy the unscanned torrent files to a temp dir
mkdir $tmp_torrent_dir
find `pwd` -name "*.torrent" | sort -n | grep -vFf $scanned_out | while read torrent_file ; do
cp "$torrent_file" "$tmp_torrent_dir"
echo $(basename "$torrent_file" .torrent) >> $names_out
done
# Delete null torrents from the temp dir
find $tmp_torrent_dir -name "*.torrent" -size -2k -delete
# Split these into many directories ( since torrent-tracker-health can't do too many full size torrents)
cd $tmp_torrent_dir
# i=1;while read l;do mkdir $i;mv $l $((i++));done< <(ls|xargs -n100)
ls|parallel -n100 mkdir {#}\;mv {} {#}
if [ -z "$(ls -A $tmp_torrent_dir)" ]; then
echo "No new torrents."
else
# Scrape it
torrent-tracker-health --torrent "$tmp_torrent_dir" > $health_out
for tmp_torrent_dir_sub in *; do
echo "sub dir: $tmp_torrent_dir_sub"
find $tmp_torrent_dir_sub -type f -exec basename {} .torrent \; > names.out
echo -e "$health_out"
# Delete null torrents from the temp dir
find $tmp_torrent_dir_sub -name "*.torrent" -size -2k -delete
# Convert the json results to csv format
results=$(jq -r '.results | map([.hash, .name, .length, (.created | .[0:16] | strptime("%Y-%m-%dT%H:%M") | mktime), .seeders, .leechers, .completed, (now | floor)] | join(";")) | join("\n")' $health_out)
# // "1970-01-01T00:00:00.000Z"
# If there are no results
if [ -z "$results" ]; then
echo "There were no results for some reason."
if [ -z "$(ls -A $tmp_torrent_dir_sub)" ]; then
echo "No new torrents."
else
# Update the torrents.csv and infohashes scanned file
echo -e "$results" >> $torrents_csv
cat "$names_out" >> $scanned_out
# Scrape it
torrent-tracker-health --torrent "$tmp_torrent_dir_sub"/ > health.out
popd
cd scripts
. prune.sh
# Convert the json results to csv format
results=$(jq -r '.results | map([.hash, .name, .length, (.created | .[0:16] | strptime("%Y-%m-%dT%H:%M") | mktime), .seeders, .leechers, .completed, (now | floor)] | join(";")) | join("\n")' health.out)
# If there are no results
if [ -z "$results" ]; then
echo "There were no results for some reason."
else
echo "Torrents.csv updated with new torrents."
# Update the torrents.csv and infohashes scanned file
echo -e "$results" >> $torrents_csv
cat names.out >> $scanned_out
fi
fi
done
fi
popd
cd scripts
. prune.sh
# Remove the temp dir
rm -rf "$tmp_torrent_dir"
rm "$names_out"
rm "$health_out"