lincoln23/discuss-share

This is an a web app for users with similar interests to follow each other and share ideas through a live-chat or written posts and discovery posts via autocomplete!

Discuss & Share

This app will allow people with similar interests to follow each other, live chat, and stay up to date with each other through their posts!

Setup

Requires:

• Ruby 2.5.3
• Rails 5.2.2
• MySQL
• ElasticSearch version >= 6.4

Clone this repo

• Update config/database.yml with your database settings
• run bundle install
• run rails db:migrate
• rails db:seed if you want a sample dataset
• start server by rails s
• start ElasticSearch server

Features

Signing up

You will need to create a username and password along with your email address to sign up

• Your email address is used to activate your account and reset your password if forgotten
• When logging in, if “remember me” is checked, you will not have to re-login even after closing your browser
• The profile picture associated with your email address is fetched on Gravatar

Share posts

Share thoughts and updates with your followers by writing posts

• Now supports file uploads so you can upload images and attachments

Follow other users and stay up to date with their posts in your feed

• Note: If you follow a certain user, they will not automatically follow you back

Searching

Want to find a post but can’t remember who wrote it? Want to find a post but you can only remember parts of it?

• Go to the search page and type in what you remember!
  • If you remember who wrote it, type in their user_name or email address
  • If you remember only parts of a post, type that in and it will return auto-complete suggestions or posts that have partial matches

Live chat

Discussion in real time! Use the live chat feature to talk to other about topics you care about!

• You can ping other users in chat by typing @[their name]

Technical Details

Autocomplete suggestion and fuzzy matching

• Using ElasticSearch, I have created search functionality that will predict what you are typing and match documents even though it is not a 100% match
  • To do this, I first apply a standard tokenizer on the input to separate the text into words. I then apply lowercase, asciifolding and Ngram filters to create lowercase N-grams of words to be indexed and searched
  • I index the contents of the posts, username and email addresses to allow users to find posts created by a certain user or search for posts

Live chat

• Real-time multi-way communication for users to chat with each other was built using ActionCable (Websockets)
• All users are subscribed to a general channel. Messages submitted by any user in the chat are broadcasted
• @mention feature: Users are subscribed to their own private channel using their unique user_id.
  • When any message is submitted, it is parsed for @Name using regex and if it matches a user, an alert to broadcasted to their channel

Login system

• Implemented using bycrpyt to keep user password safe in the database by storing the hashed version of the password
• User session is persisted across pages using Rails session
• Remember me feature is implemented by generating a unique hash then encrypting and storing it on the browser’s cookies and database.
  • When a user visits this application after closing their browser, the app authenticates by decrypting the stored browser cookie with the database and checks if the current user is valid
  • When a user logs out of their account, the stored hash is deleted both in the browser and database. Doing so will prevent someone from having permanent access to a user's account if the database is compromised

Mailer

Account activation and password resets

• Users are required to activate their account by going to their email address and confirming
• Mailers are sent using Action Mailer
• A unique hash is generated when the user first creates an account and a unique link with the hash as a query parameter is sent to the account's email address.
  • When the user clicks the "activate my account" button, the hash is authenticated with the one stored in the database
• This works the same for password resets

Database

• Using MYSQL database
• Modelled following users using a many to many relationship and a join table