Analytica - Gaming data analysis tool

Tech stack used

• Python
• FastAPI
• Docker
• AWS EC2
• Clickhouse

Setting up app locally

• Clone the repository

git clone  https://github.com/GaurKS/analytica.git

• Change directory to root and install all the packages in requirements.txt

cd analytica
pip download -r requirements.txt

• Create an environment file using the .env.sample reference
• Run the app and navigate to localhost:8000/health

uvicorn main:app --reload

Running the app using docker image

• Make sure docker is downloaded
• Pull the docker image for the app

docker pull krugarr/fastapi-app-analytica

• Run the container

docker run -d -p 8000:8000 --name my-app krugarr/fastapi-app-analytica

• Navigate to localhost:8000/health to test the app

Server endpoints

GET: http://3.108.184.193/health

• This is a simple health check endpoint to test the status of server. It is a public route

GET: http://3.108.184.193/query

• This is a query endpoint to query the data stored using multiple filters. It supports various filters, aggregation and search parameters. It is a protected route and can be accessed using a valid api-key.

POST: http://3.108.184.193/upload

• This is a post endpoint. It takes a csv file url as a body and stores the data in the database. File url can be passed as an url parameter in request body. A sample csv file url is attached below. Currently the endpoint only supports google drive download links which are accessible to public.

https://drive.google.com/uc?id=1ZIMlNQ226uaFvTId8ineajrBmWdQB2Ew&export=download

Complete details related to application endpoints, request body, query parameters and sample response can be found in the API documentation

Docker Image is available on DockerHub

All the deployed endpoints can be accessed at http://3.108.184.193/

For deployment, T2.micro instance is used under AWS free cloud credits. For database, Clickhouse free cloud instance is used.

Note: Initial database call can take upto several seconds as the DB instance is in idle state, Once the first call is made the following calls will have much lesser response time. The intial response time is mainly because of the cold start of the Clickhouse DB instance.