ftx-currency-trades

A web service that provides quotes for digital currency trades using data from the​ ​ FTX​ orderbook.

# Getting started

• Ensure to install Node and NPM to your computer.
• Clone the repo and go to the ftx-currency-trades folder.
• Run npm install within your CLI tool.
• After that, run the npm start and ensure to see "Server is running on PORT 4000" message on your console.
• Now you can open Postman and request to the available routes.

# Application routes and Postman collections

Application routes and built-in API request samples exist with Postman Collection folder.

Go to: /postman-collections/README.md

---

# Time consume

1. Planing: Undersanding and planing took almost 1.5 hour
2. MVP-scaffolding: 2 hours
3. Apply API Logic to recive request: 2 hours
4. Planing API Logic to response: 45 minutes
5. Apply API Logic to response: 2.5 hours
6. Code quality improvments: 30 minutes

Total: 9 hours 15 minutes.

# OKr (Objective Key-Result) Planning

1. ✅ - Understand the CASE steps
2. ✅ - Understand the FTX orderbook
3. ✅ - Understand the desired API context: Request and Response

End of TimeConsume.1

4. ✅ - An empty application scaffolding
   • Starting with MVP approach
   • Github; README, gitignore
   • A basic router with dummy-data
   • Accepts JSON body w/ dummy
   • Response JSON body w/ dummy

End of TimeConsume.2

5. ✅ - Match consumer input with request to the FTX's orderbook API.

End of TimeConsume.3

6. ✅ - Match consumer input with response /Response

End of TimeConsume.4 & TimeConsume.5

7. ✅ - Code cleaning and moving into better source-code w/ boilerplate

End of TimeConsume.6

# API logic

1. Determine between 'bids/asks': Consumer's "BUY" request goes to orderbook.asks and "SELL" request goes into the orderbook.bids.
2. Grab the consumer's 'amount': We will need to use 'amount' value provided by the consumer in both case for 'bids/asks'.
3. Compare to orderbook: Find the most closest 'amount' value from orderbook regarding 'bids/asks' differentiation.
4. Left-out amount: So we have compared the values and now store the extra value between two amount of orderbook and consumer.
5. Aggreation process (Modify the amount): In case the closest orderbook's 'amount value is not same to consumer's 'amount' value (which mostly will not be!), then we will need to aggreate orderbook to the consumer's amount;
   • Find the one unit price for orderbook amount on the closest "bids/asks". (amount / price)
   • Multiply one unit amount price to the left-out amount value (left-out amount * one-unit amount price)
6. Response back to the consumer with the final quote which is a weighted average of newly price.

# Possible improvments for the project

• Implementing a Swagger (OpenAPI) API documentation for the consumers.
  • Also a JSdoc for the developers.
• Applying more securty fixes (Helmet, RateLimiter, XSS injection, Snyk scanning etc.)
• Deploying the application to the Heroku cloud services.
• Providing a CI/CD process either with Heroku CI or Travis CI.
  • Enviroment seperations for the Main / Development products.
• Test cases: Providing the NodeJS test pyramid (Unit, Integration, E2E)
• More code quality approches: Lint, Prettier, husky, lint-staged, code-editor standards
• Service health monitors: App Telemetry and App Service status page.
• Validation control: Using Joi and similar project to check consumer inputs.
• A fully-functional boiler-plate adaptation.