画像で議論

girondl

#!/bin/bash -eu
GIRON_URL='https://rara.jp/giron/photo'
DOWNLOAD_DIR=$HOME/Pictures/tmp
MAX_PAGE=10
CURSOR=$DOWNLOAD_DIR/.last_download_url

VERBOSE=0
CURL_OPT="-Ss"
if [ $# -gt 0 ]; then
    if [ "$1" = "-v" ]; then
        VERBOSE=1
        CURL_OPT=""
    fi
fi


CURL_TMP=`mktemp`

log() {
    if [ $VERBOSE -gt 0 ]; then
        echo $1 1>&2
    fi
}

touch $CURSOR

if [ -s $CURSOR ]; then
    CURSOR_ID=$(cat $CURSOR | grep -oP '(?<=_)[0-9]+(?=\.)')
else
    CURSOR_ID=0
fi
for i in '' $(seq 2 $MAX_PAGE); do
    if [ -n "$i" ]; then
        log "start download page $i."
    else
        log 'start download page 1.'
    fi
    log $GIRON_URL$i
    curl $CURL_OPT $GIRON_URL$i | xmllint --xpath '//img/@data-original' --html - 2>/dev/null | grep -oP '(?<=")http[^"]+(?=")' | sed -e "s/p_/_/g" >> $CURL_TMP
    if test -s $CURSOR && grep -q $(cat $CURSOR) $CURL_TMP; then
        log 'found cursor image.'
        break
    fi
done
for u in $(cat $CURL_TMP); do
    ID=$(echo $u | grep -oP '(?<=_)[0-9]+(?=\.)')
    HASH=$(echo $u | grep -oP '(?<=/)[0-9a-f]+(?=_)')
    EXT=$(echo $u | grep -oP '\.[a-z]+$')
    if [ "$ID" -gt "$CURSOR_ID" ]; then
        log "downloading: $u"
        curl $CURL_OPT -o $DOWNLOAD_DIR/${ID}_${HASH}${EXT} $u
    else
        log "skipped: $u"
    fi
done
head -n1 $CURL_TMP > $CURSOR