sync.sh

#!/bin/bash

readonly rss_file='subscriptions.txt'
readonly memory_file='latest_vids.txt'
readonly working_file="$memory_file.new"
readonly last_sync_file='last_sync.txt'

# Delete temp file if it exists
if test -f "$working_file"; then
    rm "$working_file"
fi

# Check if memory file exists and make a copy for editing
if test -f "$memory_file"; then
    cp "$memory_file" "$working_file"
else
    touch "$working_file"
fi

# Fetch RSS feeds in parallel
echo 'Syncing feed...'
threads=10
total=$(wc -l < "$rss_file")
iter=$(((total + threads - 1) / threads))
index=-1
line=0

# TODO scrape channel page instead of RSS feed for video duration
for i in $(seq 1 $iter); do
    for _ in $(seq 1 $threads); do
        index=$((index + 1))
        line=$((line + 1))
        url=$(sed "${line}q;d" "$rss_file")
        if [ "$url" = "" ]; then
            break
        fi
        # TODO use single line progress counter
        #echo "Syncing $line"
        url=$(sed "${line}q;d" "$rss_file")

        # Save output to a file to enable concurrency
        # TODO test using multi line function with &
        curl -s "$url" > "$line" &
    done
    wait
done

echo 'Sync done. Checking which videos are new.'

# Use 2 arrays, one for video ID and one for video title
declare -a ids
declare -a titles
ctr=0
mem_total=$(wc -l < "$working_file")

# Query for the RSS feed and parse response for the video URLs
for file in $(seq 1 "$total"); do
    #regex=$(awk '/<yt:videoId>/,/<[/]title>/' <<< "${xml}")
    xml=$(cat "$file")
    rm "$file"
    first=true
    count=0
    last_seen=""
    ctr=$((ctr + 1))

    # TODO use channel name, printf
    #echo "Checking $ctr"

    # Determine which video is new and grab the ID of new videos
    while IFS= read id; do
        # Strip grep output for the exact video ID
        id=${id#*"<yt:videoId>"}
        id=${id%"</yt:videoId>"}

        # Check if the newest video matches the one that is recorded
        if [ "$first" = true ]; then
            first=false

            # If channel has never been seen before, add to memory
            if [ "$ctr" -gt "$mem_total" ]; then
                echo "$id" >> "$working_file"
            else
                # If newest video has not been seen, update memory
                last_seen=$(sed "${ctr}q;d" "$working_file")
                if [ "$id" != "$last_seen" ]; then
                    sed -i "${ctr}s/.*/${id}/" "$working_file"
                fi
            fi
        fi

        # Exit loop once all new videos have been seen
        if [ "$id" = "$last_seen" ]; then
            break
        fi

        # Add new video to array
        ids+=("$id")
        count=$((count+1))
    done < <(grep "<yt:videoId>" <<< "$xml")

    # Go to next channel if there are no new videos
    if [ "$count" = 0 ]; then
        continue
    fi

    # For each new video, grab the title as well
    first=true
    channel=""
    while IFS= read -r title; do
        # Strip grep output for the exact title
        title=${title#*"<title>"}
        title=${title%"</title>"}

        # Skip the first one as it is the channel name
        if [ "$first" = true ]; then
            first=false
            channel="$title"
            continue
        fi

        # Add title to array
        titles+=("$channel $title")
        count=$((count-1))

        # Terminate once all titles for new videos have been obtained
        if [ "$count" = 0 ]; then
            break
        fi
    done < <(grep "<title>" <<< "$xml")
done

# Exit if no new videos found
if [ ${#ids[@]} = 0 ]; then
    echo 'No new videos found.'
    rm "$working_file"
    exit 0
fi

# Overwrite old latest_vids.txt with new one
mv "$working_file" "$memory_file"

# Save the current video list
echo "${ids[@]}" > "$last_sync_file"
echo "${titles[@]}" >> "$last_sync_file"

# TODO move below code into its own function/file

# Print each video with index
# TODO prettify
for i in "${!ids[@]}"; do
    duration="$(yt-dlp -O duration_string "https://www.youtube.com/watch?v=${ids[i]}")"
    echo "$i $duration ${titles[$i]}"
done

# Ask for user input to select the videos to download
# TODO loop through 2 menus, first one for options:
# download video, download audio, watch now, bookmark, custom, exit
# second one to select videos
read -a input -p 'Select videos to download: (e.g. "1 2 3") '
#read -a queue -p "Select videos to download: (e.g. \"1 2 3\", \"1-3\" or \"^4\")"

# Exit if no videos selected for download
if [ ${#input[@]} = 0 ]; then
    echo 'No videos selected. Exiting.'
    exit 0
fi

queue_file="queue.txt"
declare -a queue

# Make a queue with the IDs of the selected videos
for i in "${input[@]}"; do
    link="https://www.youtube.com/watch?v=${ids[$i]}"
    queue+=("$link")
done

# Save the download queue into a text file
echo "${queue[@]}" > "$queue_file"

# Use yt-dlp to download each selected video
for i in "${!queue[@]}"; do
    if yt-dlp "${queue[$i]}"; then
        # After a video is downloaded, remove it from the download queue
        unset "queue[$i]"
        echo "${queue[@]}" > "$queue_file"
    fi
done

# Exit
if ! grep -q '[^[:space:]]' "$queue_file"; then
    echo 'All selected videos downloaded successfully!'
    rm "$queue_file"
    exit 0
else
    echo 'Failed to download some videos. Please use "./main -r" to retry.'
    exit 3
fi