/bug-free-octo-funicular

Primary LanguagePython

TMDB App

This app is a multi-Docker container web application that transcodes uploaded video files to .h264 format.

Once the videos are transcoded, it gathers the movie information from TMDB and displays them on the front page.

These uploaded video files will be named like <IMDB-ID>.<VIDEO-EXTENSION. (i.e. tt0241527.mp4). Where the IMDB ID will be used to identify which IMDB movie the video file stores the content of.

Composition

This app is run with 4 Docker containers that serve as different components of the web application:

  • backend -> Serves as the backend application server that interacts with the database and transcodes videos. The application uses FFMPEG to transcode videos. This runs as a Python SQLAlchemy-extended Flask application. A better design pattern may possibly be to have the video transcoder application run as a separate Docker container or application, instead of consolidating the application server with the video transcoder application.
  • frontend -> Serves as the frontend user-interface of the web application. This application runs as a React.js application.
  • db -> The PostgreSQL Database that stores all state information, specifically movie-detail information and the status of each video transcoding process.
  • filesystem_poller -> A simple Python application that serves as a watcher/poller that detects for events on the file system, specifically, the folder where users upload videos. It detects upload events and triggers the downstream tasks of transcoding the video and fetching the movie information from the TMDB database.

Implementation Details

  • backend
    • Flask -> I decided to use Flask as my Python-based web framework of choice because Flask tends to be recommended as a popular choice for any light-weight, extensible, and micro web applications. If the application were to be more heavyweight, then I would have possibly explored other frameworks, like Django.
    • SQLAlchemy -> I chose to use SQLAlchemy with Flask, because they help abstract database interactions which create optimized SQL queries for better performance. It also makes the code more readable, maintainable, and re-usable. It also helps improve security by preventing SQL-injection attacks. And the database abstractions that ORMs introduce can help switching databases easier. Although for this use-case, these benefits are not really applicable, I tend to prefer ORM frameworks when feasible and reasonable.
    • ffmpeg -> This multimedia service was recommended by the interviewer to transcode videos.
  • frontend
    • React.js -> This popular web development framework was recommended by the interviewer to host the front end user-interface.
  • filesystem_poller
    • watcdog -> Watchdog is a Python libarary that monitors file system events.
  • db -> This database service was recommended by the interviewer to store movie details and video transcoding status information.

Usage

Application Args

Here are 3 required, command-line arguments to the application:

  • UPLOADS_FOLDER -> The name of the input/source video folder. Please prepend a / to the path. This will map the folder from your $HOME path to the root of the application filesystem. For example, if you set this to be /uploads, then this will create a mapped volume from the $HOME/uploads path in the local machine to the /uploads path in the root of the Docker/application filesystem.
  • COMPLETE_FOLDER -> The name of the target/destination video folder for transcoded video files. Please prepend a / to the path. This will map the folder from your $HOME path to the root of the application filesystem. For example, if you set this to be /uploads, then this will create a mapped volume from the $HOME/uploads path in the local machine to the /uploads path in the root of the Docker/application filesystem.
  • BACKEND_URL -> The URL to the backend application server.
  • TMDB_API_KEY -> The TMDB V3 API Key to use to authenticate when fetching movie details from the TMDB API

Running Application Locally

Command to run the application:

UPLOADS_FOLDER=/uploads \
COMPLETE_FOLDER=/complete \
BACKEND_URL=localhost \
TMDB_API_KEY=XXXXXX \
docker-compose up --build --detach

Since we're running this application on our local machines, the URL to the backend application server will be localhost.

Copy a video file to the application source folder:

cp /path/to/videos/tt0295297.mp4 ~/uploads/tt0295297.mp4

Running Application in AWS EC2

Command to run the application:

UPLOADS_FOLDER=/uploads \
COMPLETE_FOLDER=/complete \
TMDB_API_KEY=XXXXXX \
BACKEND_URL=$(curl http://169.254.169.254/latest/meta-data/public-hostname) \
docker-compose up --build --detach

The $(curl http://169.254.169.254/latest/meta-data/public-hostname) is a command to fetch the Public IPV4 DNS of the EC2 Instance itself. More information here.

Command to SSH into the AWS EC2 Instance:

ssh -i /path/to/pem/test.pem \ 
ec2-user@<AWS EC2 INSTANCE PUBLIC DNS>

Command to upload files to the AWS EC2 Instance:

scp -i /path/to/pem/test.pem \ 
/path/to/videos/tt0241527.mp4  \
ec2-user@<AWS EC2 INSTANCE PUBLIC DNS>:~/uploads/tt0241527.mp4