Last active
April 3, 2019 17:52
-
-
Save alexshpilkin/73f301ed620993665561f7d404cefc0b to your computer and use it in GitHub Desktop.
Scrape rutracker.org forums
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/bin/sh -eu | |
set -o pipefail | |
TRACE=${TRACE-}; export TRACE | |
tracef() { | |
fmt=$1; shift | |
if [ -t 2 ]; then printf '\033[0K%s'"$fmt"'\r' "$TRACE" "$@" >&2; fi | |
} | |
trap 'tracef ""' EXIT | |
id=${1##*f=} | |
TRACE=$TRACE$(printf '%05d ' "$id") | |
tracef '%s' "$1" | |
rm -f "forum$id.url" | |
page=$1; i=1; trace=$TRACE | |
until | |
TRACE=$trace$(printf '%02d ' "$i") | |
tracef '%s' "$page" | |
echo "$page" >>"forum$id.url" | |
curl -sSL -o - "$page" | \ | |
sed -Ee 's/<(meta charset=[^>]+)>/<!-- \1 -->/' | \ | |
iconv -f cp1251 >"forum$id.$i.html" | |
pup -p a.torTopic attr{href} <"forum$id.$i.html" | \ | |
while read -r url; do | |
url=${page%/*}/$url | |
"$(dirname "$0")/rutopic" "$url" | |
done | |
next=$(pup -p a.pg attr{href} <"forum$id.$i.html" | tail -1) | |
next=${page%/*}/$next | |
grep -qFe "$next" "forum$id.url" | |
do | |
page=$next; i=$((i+1)) | |
done |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/bin/sh -eu | |
set -o pipefail | |
TRACE=${TRACE-}; export TRACE | |
tracef() { | |
fmt=$1; shift | |
if [ -t 2 ]; then printf '\033[0K%s'"$fmt"'\r' "$TRACE" "$@" >&2; fi | |
} | |
trap 'tracef ""' EXIT | |
id=${1##*t=} | |
TRACE=$TRACE$(printf '%07d ' "$id") | |
tracef '%s' "$1" | |
rm -f "topic$id.url" | |
page=$1; i=1; trace=$TRACE | |
until | |
TRACE=$trace$(printf '%02d ' "$i") | |
tracef '%s' "$page" | |
echo "$page" >>"topic$id.url" | |
curl -sSL -o - "$page" | \ | |
sed -Ee 's/<(meta charset=[^>]+)>/<!-- \1 -->/g' | \ | |
iconv -f cp1251 >"topic$id.$i.html" | |
next=$(pup -p a.pg attr{href} <"topic$id.$i.html" | tail -1) | |
next=${page%/*}/$next | |
grep -qFe "$next" "topic$id.url" | |
do | |
page=$next; i=$((i+1)) | |
done | |
curl -sSL -o - -b cookies.txt -d "t=$id" "${1%/*}/viewtorrent.php" | \ | |
iconv -f utf-8 | pup -p '.ftree > li json{}' | jq ' | |
.[] | | |
walk(if type == "object" and .tag == "div" | |
then {"name": .children | map(select(.tag=="b") | .text) | join(""), | |
"size": .children | map(select(.tag=="i") | .text) | join("")} | | |
(if .size == "" then del(.size) else .size |= tonumber end) | |
elif type == "object" and .tag == "li" and .class == "dir" | |
then {"name": .children[0].name, | |
"children": .children[1].children} | |
elif type == "object" and .tag == "li" | |
then .children[0] | |
else . end) | | |
.name |= sub("^\\./"; "") | |
' >"topic$id.json" | |
jq -r ' | |
walk(if type == "object" and .children | |
then .name as $name | | |
[{"name": ""}, .children[][] | | |
(.name |= $name + "/" + .)] | |
elif type == "object" | |
then [.] | |
else . end) | | |
.[] | "\(.size // "")\t\(.name)" | |
' <"topic$id.json" >"topic$id.tsv" |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment