Automating video downloads

Being a dad comes with opportunities to automate mundane chores. To curate a sensible kids video collection, I have to download select videos from YouTube. To organize this content, I have recently setup a container running the jellyfin media (offline) server. This exercise of curation and downloading videos manually requires a lot of unnecessary effort and should be have been automated.

Yes, it is time to automate!

Solution architecture

Below is the solution architecture. The input or the video urls are saved on a csv with headers VideoURL & Type.

---
title: Video download automation
---
flowchart TD
subgraph 0[Video download]
1[RequiredVideos.csv] -.-> |VideoURL, <br> Type| 2[ytDownload.sh<br>yt-dlp]

2 -.->|mkdir| 3[Folder_Category_1]

2 -.->|mkdir| 4[Folder_Category_2]
2  -.-> |mkdir| 5[Folder_Category_N]

3 -.->|Download video, description, <br> thumbnail, subtitles| 6[Individual_VideoFiles]
end

6 -.-> |Copy|7

subgraph Jellyfin server
        10[Jellyfin_Server]
        7[Folder_Category_1]
        8[Folder_Category_2]
        9[Folder_Category_N]
    end
  
subgraph Tools used
11[Bash<br>yt-dlp<br>github <br>Jellyfin]
end

This .csv file then is an input file to a bash script. The bash script iterates over the rows in the csv file and uses the yt-dlp project/cli tool to download videos. In short, it will help me build a video collection (for personal use) with minimal resistance.

The next time when I or my wife want to add a video, we add a row in the .csv file with the video url and a corresponding type (category).

ytDownload.sh script

#!/bin/bash

# Reading the csv file and using Videos column
VIDEO_URLS= readarray -t eCollection < <(cut -d, -f1,2 RequiredVideos.csv)

# Set the directory where you want to save the downloaded files
OUTPUT_DIR="."

# Create the output directory if it doesn't exist
mkdir -p "$OUTPUT_DIR"

for row in ${eCollection[@]:1} 
do
url=$(echo "$row" | cut -d, -f1)
typeVideo=$(echo "$row" | cut -d, -f2)
echo "Starting to download : "$url" "
# Download the video and its subtitles no special characters do not overwrite video if already found(if available)
yt-dlp -f 'bestvideo[ext=mp4]+bestaudio[ext=m4a]' --write-auto-sub --write-description --sub-lang en --restrict-filename --no-force-overwrites --output "$typeVideo/%(title)s/%(title)s" "$url"
echo "Completed download : "$url" "
done

Command explanation

• Get the best video resolution,
• Best audio quality,
• Download subtitles if available in english,
• Download description if available,
• Remove symbols and special characters from video filename,
• Skip download if resource already found in target folder,
• Save to respective type/videofilename folder

yt-dlp -f 'bestvideo[ext=mp4]+bestaudio[ext=m4a]' --write-auto-sub --write-description \
 --sub-lang en --restrict-filename --no-force-overwrites --output "$typeVideo/%(title)s/%(title)s" "$url"

Further improvements

The input .csv should be made accessible via a network-attached-storage device which supports a native mobile app. This way we always have access to the input file.

A further step would be to run a cron job on a server to look for new rows added in the .csv file and run the script. This will ensure that my jellyfin collection is always up-to-date.

Credits

Thank you to all the maintainers of yt-dlp!