#!/bin/bash echo -e "\n-========================================================================-" echo -e "-=- -=-" echo -e "-=- SCRAPE_LPS.SH: Downloads LPS committee agendas and minutes -=-" echo -e "-=- -=-" echo -e "-=- https://gist.github.com/rvtr/******************************** -=-" echo -e "-=- Lillian Skinner -=-" echo -e "-=- -=-" echo -e "-========================================================================-" source ./functions/.functions MEETINGS_PAGE="./tmp.html" # London seems to have recently blocked unusual user agents. Can't use wget or even ping. Thankfully pretend to be a real person! WGET_UA="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/125.0.0.0 Safari/537.36 Viewer/99.9.8782.87" mkdir "./LondonArchive" mkdir "./LondonArchive/LPS" mkdir "./tmp" wget --user-agent="$WGET_UA" "https://londonpoliceserviceboard.com/board-meetings/" -O "./tmp/index.html" -q #--show-progress current_year=$(date +%Y) current_month=$(date +%m) current_day=$(date +%d) # If I don't set these values then "10#: invalid integer constant" ITEM_YEAR="0000" ITEM_MONTH="00" ITEM_DAY="00" while IFS= read -r LINE_PRE; do LINE=$(echo $LINE_PRE | sed 's/\xC2\xA0/ /') FOUND_ATTACH_TYPE="$(echo $LINE | grep '

') if [[ "$ATTACH_TYPE" != "" ]] && [[ "$FOUND_LINK" != "" ]]; then _time_parse_helper "$(echo $FOUND_LINK | sed 's/.*]*>\([^<]*\)<[\/:-]a>.*/\1/' | sed -e 's/\([0-9]\{4\}\).*/\1/' | sed -e 's/^[[:space:]]*//g; s/[[:space:]]*$//g')" echo "$ITEM_YEAR/$ITEM_MONTH/$ITEM_DAY" echo "$(echo $FOUND_LINK | sed 's/.*href="\([^"]*\)".*/\1/')" mkdir -p "./LondonArchive/LPS/Board/$ITEM_YEAR/$ITEM_MONTH-$ITEM_DAY/" _utils_download_helper "$(echo $FOUND_LINK | sed 's/.*href="\([^"]*\)".*/\1/')" "./LondonArchive/LPS/Board/$ITEM_YEAR/$ITEM_MONTH-$ITEM_DAY/$ATTACH_TYPE.pdf" fi done < "./tmp/index.html"