Usage

You can access the service via http://3.29.127.29

NOTE: https was not possible due to the timeline given thus a valid domain was not able to get issued and subsequently the SSL certificate

Playing with the codes

• to run the application locally please run as below and open your browser http://localhost:8000

  $ make up-local

• to run test and linting

  $ uv sync
  $ make checkcodestyle
  $ make test

Tasks

• Ability to resize the frames from 200 to 150 columns.
• Ability to store the resized images in a database.
• Ability to request image frames using an API with depth_min and depth_max query parameters.
• Ability to use colormap to generate the frames.
• Python should be the major langauge used.
• Deploy the application in a cloud service.

Highlights

• A simple UI has been added to simplify the process of viewing the frames.
• Standard software practices were applied such has linters, tests, scripting and docker build.
• A document database (MongoDB) was chosen as the database due to the nature of the data given.
• The Application is deployed in a EC2 machine.
• as a cloud native design, an object store support was added, in the current code based, a minio instance is utilize which can be change easily to s3 by changing some environment variables.

Things I wish I had more time to do

• Authentication layer
• Validation of file upload
• More test specially for failed actions
• Better UI 😅
• Github Action for building the image
• Github Action for running CI job such as linting and tests

Architecture and Stack

• MongoDB: is the main database as the image frames fits well as document.
• Minio: an object store that has local/self-hosted option and is 100% compatible to s3, allowing a easy to switch vendor.
• Traefik: a cloud native loadbalancer.
• Python: My Programming language loved by many 🤓
• FastAPI: Python Framework for quick and easy API
• Polars: a Faster alternative to pandas