Project Name

Collaborative

Overview

This project is an MVP for the Light-weight Collaborative chat space for teams. Follows to the barest minimum, SOLID, DRY, KISS principles. The application is a monolith one as it is not. Uses Socket.io for real-time messaging and includes minimal RBAC, redis implementations for perfomance and rate-limiting

The File structure sees all feature modularized appropriately

Documentation

• Base url: https://collaborative-v1.onrender.com/v1 (HTTP)
• Postman: APIs - https://www.postman.com/kairosclicks/workspace/ko/collection/16936021-d1a99890-7098-4733-a8c0-b36c2c3b0991?action=share&source=copy-link&creator=16936021

Messaging - https://www.postman.com/kairosclicks/workspace/ko/collection/69cbc764b80ea68a9aafebf1?action=share&source=copy-link&creator=16936021

The above are sperated into various collections because the former is HTTP where as the latter is WS and cannot not be in the same collection

• Swagger Open API: https://collaborative-v1.onrender.com/api-documentation

Features

• Authentication:

  Registration/Onboarding:

  • Register (No email service implemented hence for test purposes, it return a verification code)
  • Verify Email (No email service implemented hence for test purposes, it return a verification code)

  Login

  • Basic authentication: Including login, forgot and reset password implementation

• Strict User Input Validations.

• Chat Rooms: User can create chat rooms (or team channel) and then add users to their chat room. where only members are allowed to send or recieve messages.

• Message ReplyTo: a user (who is a member of a chat room) can reply to a single message (For example how it works on Ms Teams, WhatsApp, etc.)

• Message search: A user can search for a keyword amongs all messages effectively using as search term. See /messages endpoint

• Message Status Management

• Open AI Implementation: AI has access to all chat rooms an gets invoked when when mentioned to respond to a message where context is derived by obtaining summary.

  Real time chating events:

  • User upon been added to a chat trigger joinRoom listens to:
    • exception for catching proper error like unauthorized, missing/empty fields
    • newMessage event for new messages
    • aiThinking when "@ai" has been mention and is processing responses

  Rate Limiting:

  • Using throttlers for HTTP request while a rate limit guate for WS

Assumption

• The application is not for person-to-person chat but a shared chat room for teams

What you should know

• Application was deployed on Render free version and as such makes inital request to delay by 50 second or more
• OpenAI free credits may run out via testing,
• Middleware was used for Enrichment

Setup

1. Requires:

• NodeJS >=18.0
• NestJS >= 11
• TypeScript

1. Clone the repository
2. Install dependencies:

   npm install

3. Run:

  npm run migration:run

## Migrations
- To generate migrations run:
```bash
  npm run migration:generate src/migrations/{{nameOfYourMigration}} 

- To migrate run:
```bash
  npm run migration:run

- To revert run:
```bash
  npm run migration:revert

## Testing 
To run test:
  ```bash
    npm run migration:revert

## Improvements with more time
- Logging 
- Authentication Refresh 
- Message threads. (A bit similar to the replyTo feature but more like a tree)
- Message translation
- Caching messages for to tune perfomance
- Chatroom invite link
- Chatroom roles
- As admin Remove user from chat room
- Email service
- More Unit tests