/thumbnail-server

Primary LanguageJavaScript

Thumbnail Server

A server that returns a thumbnail for an uploaded image.

This is an assignment from Cogent Labs.

Requirements

Docker and docker-compose should be installed.

Run the app server:

$ docker-compose up --build

Run the test:

$ docker-compose -f docker-compose-test.yml up --build

Usage

Once the server is running, the APIs are aviable.

API List

  • POST /upload: post an image with a POST request, and get an url for later check.
    • attach: image file (png or jpg)
    • code 202
      • return: {status: "in_process", url: string}
    • code 500
      • return: {status: "server_error"}
  • GET /check/:id: ask if the task is finished. Once the task is finished, the output image is encoded as base64.
    • code: 200
      • return: {status: "done", image: base64_image}
    • code: 202
      • return: {status: "in_process", url: string}
    • code: 404
      • return: {status: "not_found"}
    • code: 500
      • return: {status: "server_error"}

example:

Run the server at localhost, and use the other shell to run:

# POST /upload
$ curl -X POST  -H "Content-Type: multipart/form-data" -F length=200 -F image=@image.jpg http://localhost:8080/upload
{"status":"in_process","url":"/check/617d334a-d277-4351-912a-ff3a9aac8e6f"}

# GET /check/:id
$ curl http://localhost:8080/check/617d334a-d277-4351-912a-ff3a9aac8e6f
{"status":"done","image":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAGQAAABkCAIAAAD/gAIDAAAACXBIWXMAAC4jAAAuIwF4pT92AAAgAElEQVR4nFy6BXhc55n2r4Wv9O1u ... "}

Architecture

Architecture

  • Client: user
  • Express: express on Node.js
  • Rabbitmq: message queue server
  • Redis: cache DB for task information, which include the task ID, original image, and output image
  • Worker: Node.js worker thread for handling tasks

Express is used since it's the most famous HTTP framework. Since there may be a lot of requests, a message queue is nessasary and Rabbitmq is adopted due to its robustness. We also need a DB to store information for later processing, and Redis, an in-cache memory DBMS, is used because (1) we want a rapid response to save the uploaded image and get the output image, and (2) we don't need to keep the data for a long time, which means we can erase the data soon.

Future Work

  • add a machenism to erase the data in the DB
  • add a cancel API
  • add a machenism to identify the user
  • provide more parameters for API
  • add a monitor for each components
  • add stress tests (e.g. 10k requests/second)
  • add more unit tests for corner cases

License

MIT