I started doing a basic experiment with the Reddit API in May 2016. I was initially interested in working with web-based OAuth2 authentication, which I hadn’t done before, so I started hacking away using Flask, PRAW, and SQLAlchemy. There was really no set goal, I just wanted to experiment with extracting data from my account and displaying it on a webpage.
It later dawned on me that if you have had a Reddit account for awhile and have saved a few hundred different items, that there is no readily available tool to extract that and back it up somewhere. I began creating AccountMover.Com, which has evolved into a tool that enables Reddit users to:
- Transfer their subreddit subscriptions to a different account
- Transfer all saved items to a different account
- Export those saved links as an HTML file that can be imported into your browser
For users who are privacy conscious, they also have the option to:
- Delete all their submitted posts
- Delete all their comments
Performing these tasks for a (potentially) large amount of users concurrently, who may or may not have a lot of account data to load, parse, and transmit necessitated some fun choices in software. My tech stack for the website is detailed below.
For the backend, I used:
- Docker to contain, deploy, and network all the components
- Python 2.7 (Why not python 3? I ask myself that all the time and have no answer…)
- Flask, Flask-SQLAlchemy and Gunicorn
- Redis and redis-py for caching, rate limiting, and task queue brokering
- Celery task queue & Flower for monitoring
- PostgreSQL for persistent data storage
The site is hosted on the same DigitalOcean droplet I host my website on.
For the front end I used Materialize for the CSS framework, and React for the task visualization.
There were some hiccups in getting the task queues working effectively, especially with adding the Flask app context to Celery, so I could use the SQLAlchemy models inside the Celery tasks. This is also the first time I’ve put something out into the wild using React since I picked it up this summer. I’m definitely looking forward to hearing feedback regarding UX/UI, and any functional problems user have running the tasks. Feel free to let me know!
