geekgawd / branch-assignment

Branch International Assignment - Chat Application

Hello Everyone at Branch, This is my chat application. Let me walk you through the assignment what all I implemented and my design choices.

Flow and how to run the project

• /support/customer/ For the customer flow
• /support/agent/ For the agent flow

Login Credentials for Admin Panel available on /admin/:

• Username: admin
• Password: admin

Setup

Docker Setup

If you have docker installed you can simply just do the following.

docker-compose up --build -d

To Create Local Setup

1. Clone the repository:

git clone https://github.com/GeekGawd/branch-assignment

To run the server, you need to have Python installed on your machine. If you don't have it installed, you can follow the instructions here to install it.

2. Install & Create a virtual environment:

pip install virtualenv
virtualenv venv

3. Activate the virtual environment:

./venv/scripts/activate
# For Ubuntu
source venv/bin/activate

4. Install the dependencies:

pip install -r requirements.txt

5. Setup postgreSQL

• Install postgreSQL on your local machine by searching on google
• Create a table branch

6. Setup ENV You need to create a .env file with the following key/value, change it according to your needs

SECRET_KEY=django-insecure-t1xmw&dqyhc+_&b@x86!kt7o1@p+0_v%g2t5@b3+pz9_#t)7fq
CHANNEL_HOST=localhost
REDIS_URL=redis://localhost:6379/0
POSTGRES_DB_NAME=branch
POSTGRES_DB_USER=postgres
POSTGRES_DB_PASSWORD=admin
POSTGRES_DB_HOST=localhost
POSTGRES_DB_PORT=5432

7. Make the database table and add migration

python manage.py makemigrations
python manage.py migrate

8. Run the server

python manage.py runserver

Features and Design Choices

• Websockets: I implemented websockets because we wanted persistent connections for chats, and this is one of the ways we can achieve it in real time.

• PostgreSQL: Its a solid SQL database, provides great Read performance, I also implemented Full Text Search and Faceted Search with the help of search vector available in PostgreSQL.

• Django: Extremely fast to develop with and very similar to Ruby on Rails (which I believe branch uses in the backend).

Also, implemented the following bonus features:

• ✅ No two agents can work on the same support request and they cannot also forcefully enter a conversation, the connection will be closed if a agent tries to enter forcefully through url.

• ✅ Implemented Faceted Search to better priotise and focus on support requests that matter.

• ✅ Implemented Search Functionality on user, username and messages. Also, implemented GIN index for fast retrieval and search.

• ✅ Implemented preconfigured canned messages that changes as conversation goes on.

• ✅ Made UI using Django MVT architecture.

• ✅ Improved the search and lookup performance by storing the search vector in the database and also implement gin indexes.

Database Design and Scalability

• Implemented AutoIncrement ID as primary key and UUID as external_id to reference other models objects. This is done for the following reasons:

  • To easily index the AutoIncrement field and database is also easier to maintain.
  • There is security threats and issues if the simple auto increment id is exposed to users.
  • The UUID is hard to index but can't be predicted so it is passed in urls and json to reference other table and objects.

• There is conversation table that stores which customer the support request (aka Conversation) belongs to, it has fields like agent_active (that helps to know if it is being worked upon).

To make scale this application in the future.

• We can vertically scale the application at first. Like increasing redis memory and increasing number of redis nodes for websockets.

• Database: As the database size increases:

  • Sharding: We can shard the database, to accomodate more messages in future.
  • Partition: We can partition the database on the basis on timestamps per month.
  • Move to more performant databases like ScyllaDB which is used by big companies like Discord to really scale.

• Kubernetes to maintain clusters of Django Application, to scale horizontally.

About Me:

I am Suyash Singh, a final year student. I am well versed with Python, Django, Redis, Apache Kafka, Celery and Websockets.

A Little bit about me: I have strong problem solving skills and regularly practice on major coding platforms. I have Three Internship Experiences, Open Source Contributions and some notable hackathon wins are KAVACH-2023 Winner, 2x times NIT Jalandhar Hackmol Winner, Smart India Hackathon Winner, Bajaj HackRx 4.0 Runner Up, which showcase my skills as a great backend developer.

My Work: Silive Website: The Software Development Center of the College. Link: www.silive.in Tech Stack: Python, Django, HTML, CSS

Cryptocurrency Investigation Tool: Kavach Cybersecurity 2023 Hackathon Winning project. Tech Stack: React, Django, Python https://github.com/I-polusta/kavach2023

More-Fit: https://github.com/GeekGawd/morefit-ml My machine learning work done for the winning idea in Smart India Hackathon. Tech Stack: Android(Kotlin), Django, Tensorflow Lite

Log-Tracer Application: This uses Golang, Grafana and InfluxDB. This is a personal project to store the api status and timestamps and error which can be easily visualised in the grafana dashboard. https://github.com/GeekGawd/logtracer

Cryptocurrency Tracking Application: https://github.com/GeekGawd/cryptocurrency-tracker-application This project sets alerts to the cryptocurrency and sends it using email. I used cron jobs, websockets and kafka to make this system.

I have also contributed to Wagtail and other open source repositiories, which can be found on my github: https://github.com/GeekGawd

Resume: https://drive.google.com/file/d/1Ogrce3iTqOOcygo69csCF1SG6udPAYj4/view?usp=sharing

Resources Used

• Django Documentation
• Django Channels for Websockets
• Redis