#!/bin/bash
# -- prototype script

echo "Run this script in an empty folder."
echo "This probably only works on YouTube videos which have comments which have replies."
sleep 0.5
echo "YouTube video ID: $1"
echo "YouTube API key (SHA1 hashed): $(echo $2 | sha1sum - | sed 's/\s.*//g')"
echo "Cleanup option: $3"
echo "Download start: $(date --utc)"
curl "https://www.googleapis.com/youtube/v3/commentThreads?key=$2&part=snippet,replies&videoId=$1&maxResults=100" > "youtube_comments-$1.txt"

# -- Download all pages of the comments
nextpage=$(cat "youtube_comments-$1.txt" | grep '"nextPageToken"' | awk {'print $2'} | sed "s/\"\|,//g")
if [ -n "$nextpage" ]; then
    echo "nextPageToken found."
    while [ -n "$nextpage" ];
    do
        nextpagedup=$nextpage
        curl "https://www.googleapis.com/youtube/v3/commentThreads?key=$2&part=snippet,replies&videoId=$1&maxResults=100&maxResults=100&pageToken=$nextpage" >> "youtube_comments-$1.txt"
        nextpage=$(cat "youtube_comments-$1.txt" | grep '"nextPageToken"' | tail -n 1 | awk {'print $2'} | sed "s/\"\|,//g")
        echo "Old nextpage: $nextpagedup"
        echo "New nextpage: $nextpage"
        sleep 0.7
        if [[ "$nextpagedup" == "$nextpage" ]]; then
            echo "No new nextPageToken found."
            break
        fi
    done
else
    echo "nextPageToken not found (there is none)."
fi

# -- Download missing replies
echo 'execute "%s/\\(.\\+\\n\\)\\{28} \\{8}\"totalReplyCount\":\\s\\([1-9]\\|\\d\\d\\+\\),\\(\\n.\\+\\)\\{3}\\n\\( \\{4}{\\)/\\4/ge"' > vim.txt
echo 'execute "%s/,\\n\\(.\\+\\n\\)\\{28} \\{8}\"totalReplyCount\":\\s\\([1-9]\\|\\d\\d\\+\\),\\(\\n.\\+\\)\\{3}\\n\\( \\{2}\]\\)/\\r\\4/ge"' >> vim.txt
echo 'execute "%s/,\\n\\(.\\+\\n\\)\\{28} \\{8}\"totalReplyCount\":\\s0,\\(\\n.\\+\\)\\{3}\\n\\( \\{2}\]\\)/\\r\\3/ge"' >> vim.txt
echo 'execute "%s/\\(.\\+\\n\\)\\{28} \\{8}\"totalReplyCount\": 0,\\(\\n.\\+\\)\\{3}//ge"' >> vim.txt
echo 'execute "%s/\\(.\\+\\n\\)\\{6} \\{14}\"textDisplay\": \"\\(.\\+\\n\\)\\{15}.\\+/./ge"' >> vim.txt
echo 'execute "%s/\\(^\\.$\\n\\)\\+/\\=strlen(substitute(submatch(0), \"\\n\", \"\", \"g\")) . \"\\r\""' >> vim.txt
echo 'execute "%s/^ \\{8}\"totalReplyCount\": \\(\\d\\+\\),\\(\\n.\\+\\)\\{4}\\n\\1/---Has all replies/ge"' >> vim.txt
echo 'execute "%s/^ \\{8}\"totalReplyCount\": \\(\\d\\+\\),\\(\\n.\\+\\)\\{4}\\n\\d\\+/---Does NOT have all replies/ge"' >> vim.txt
echo 'execute "g!/^---\\|^ \\{6}\"id\"/d"' >> vim.txt
echo 'execute "%s/^\\s\\+\"id\":.\\+\\n---Has all replies\\n//g"' >> vim.txt
echo 'execute "%s/^\\s\\+\"id\": \"\\([^\"]\\+\\)\",\\n---Does NOT have.\\+/curl \"https:\\/\\/www.googleapis.com\\/youtube\\/v3\\/comments?key=[KEY_HERE]\\&part=snippet,id\\&parentId=\\1\\&maxResults=100\" --header \"Accept: application\\/json\" > \\1.json/g"' >> vim.txt
echo 'execute "w dl_missing.sh | q!"' >> vim.txt
vim "youtube_comments-$1.txt" -c 'execute "so vim.txt"'
cat "dl_missing.sh" | sed "s/\[KEY_HERE\]/$2/g" > "download_missing.sh"
rm "dl_missing.sh"
bash "download_missing.sh"

somemissing=$(wc --bytes download_missing.sh | sed "s/\s.*//g")

if [ "$somemissing" -gt 1 ]; then
    echo "Patching in missing replies..."
# -- Patch in missing replies
    cp "youtube_comments-$1.txt" "youtube_comments-$1_patched.txt"
    dlmissing=$(cat download_missing.sh | sed "s/.\+ //g")
# -- Do not run '; vim "$one" -c "execute \"so ""$args""_do.txt\""' in xargs, it breaks GNOME Terminal
    echo -en "$dlmissing" | xargs -d "\n" sh -c 'for args do one=$(ls | grep _patched.txt | tr -d \\n); argsnojson=$(echo -n $args | sed s/.json//g); echo "execute \"edit $args\"" > ""$args""_do.txt; echo "execute \"%s/^/\\\\\=\\\"\\\\\x20\\\\\x20\\\\\x20\\\\\x20\\\\\x20\\\\\x20\\\"/g | %s/ \\\"items\\\": / \\\"comments\\\": /g\"" >> ""$args""_do.txt; echo "execute \"normal! gg6ddGddgg\\\"ayG\"" >> ""$args""_do.txt; echo "execute \"bdelete!""\"" >> ""$args""_do.txt; echo "execute \"edit $one\"" >> ""$args""_do.txt; echo "execute \"%s/^\\\\\( \\\\\{10}\\\"id\\\": \\\"$argsnojson\\\".*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\\\\n.*\\\\\)\\\\\\\\n \\\\\{8}\\\"comments\\\": \\\\\(.\\\\\+\\\\\\\\n\\\\\)\\\\\{-} \\\\\{8}\\\\\]\\\\\\\\n/\\\\\\\\1\\\\\\\\r\\\"\\\"\\\"\\\"\\\"/g | %s/\\\"\\\"\\\"\\\"\\\"/\\\\\=@a/g\"" >> ""$args""_do.txt; echo "execute \"wq\"" >> ""$args""_do.txt; echo "vim \"$one\" -c \"execute \\\"so ""$args""_do.txt\\\"\"" >> do.txt; done' _
# -- done: make a loop which executes each line
    donum=$(expr $(wc -l do.txt | sed "s/\s.*//g") + 1)
    doloop=1; while [ $doloop -lt $donum ]; do eval $(tail +$doloop do.txt | head -n 1); doloop=$(expr $doloop + 1); done
    if [ -n "$3" ]; then
        if [[ "$3" == "gzip" ]]; then
            historyfiles=$(ls -1 . | grep -v "^youtube_comments-..........._patched\.txt" | sed 's/$/ /g' | tr -d \\n)
            echo "History files = $historyfiles"
            cat download_missing.sh | sed "s/$2/[KEY_HERE]/g" > dl_missing.sh
            cp dl_missing.sh download_missing.sh
            rm dl_missing.sh
            tar -I 'gzip -9' -cf "youtube_comments-"$1"_history.tar.gz" $historyfiles
            rm $historyfiles
        fi
    fi
else
    echo "No missing replies."
    echo "Deleting download_missing.sh..."
    rm download_missing.sh
    echo "Deleting vim.txt..."
    echo "SHA1 hash of vim.txt: $(sha1sum vim.txt | sed 's/\s.*//g')"
    rm vim.txt
fi

echo "Download end: $(date --utc)"