readme.md

Trade Management System

System built using Python, Flask, SQLite and SQLAlchemy. Additional static frontend shown on home ('/') route.

To run

• To build the system, clone the repository and (optional) setup a virtual environment using python -m venv venv To activate the virtual environment run venv\Scripts\activate
• Install requirements (Flask, SQLAlchemy) with pip install -r requirements.txt
• To run tests, run python run_test.py This performs all 4 required database executions using the api: Add and Get Trader, and Add and Get Trade
• To build the project, run python run.py You can interact with the api by referencing the Endpoints section below

Database

Built using SQLite and SQLAlchemy, the database is created in the initializer in the utilities section, and imported from database.py to avoid circular importing issues. The tables in the database are configured in the app/models/models.py file according to the requirements. By using the model class provided by SQLAlchemy, we can build the tables (along with their internal data types, relationships, and primary / foreign keys) as python data structures which can easily interact with both systems. Interaction with the database is all governed by the abstraction layer provided by SQLA. The classes are:

Trader (includes backreference from trade object so program can call trade.trader)
Table name: 'traders'

trader_id = Integer
name = String(10)

Trade
Table name: 'trades'
trade_id = Integer
currency_pair = String(10)
amount = Float
price = Float
trade_date = DateTime
identifier = String
trader_id = Integer (ForeignKey-> traders.trader_id)

(no columns are nullable for both tables)

Procedures

Interaction with the database all happens in the routes/api_handlers.py file. Each endpoint has it's own function, which ingests the query information and follows standard response protocol for either adding or reading information from the database. The main functions for database are object based, so we query depending on if we use the Trader or Trade object, these functions are Object.query, and DB.add and .commit. These are built in to SQLA, so instead of using cursors or raw SQL code to INSERT INTO a database, we can use a model and database class. This also helps with checking for common errors like rejecting a trader insert if they already exist, or even breaking down our get requests so we can query differently depending on parameters provided, as shown in the endpoints section.

Endpoints

• Add Trader POST request /api/add_trader
  Add a new trader to the traders table
  POST

{
	"name" : string
}

RESPONSE Successful (201)

{
  "message": string,
  "trader_id": int
}

• Add Trade POST request /api/add_trade
  Add a new trade to the trades table
  POST

{
    "currency_pair": string, 
    "amount": float,
    "price": float,
    "trader_id": int
}

RESPONSE Successful (201)

{
  "message": string,
  "trade_id": int
}

• Get Trader GET request /api/get_trader
  Retrieve a trader from the traders table
  Additional functionality to query by either name, trader_id, or both
  GET Query Parameters
  name : string
  or
  trader_id: int
  or
  name : string
  trader_id: int
  RESPONSE Successful (200)

{
  "name": string,
  "trader_id": int
}

• Get Trade(s) GET request /api/get_trade
  Retrieve a trade or multiple trades from the trades table
  Additional functionality to query by:
  trader id and date: returns list of trades from given trader from specified date
  trader id: returns list of all trades by given trader
  GET Query Parameters
  trade_id: int
  or
  trader_id: string
  date: DateTime
  or
  trader_id
  RESPONSE Successful (200)

{
    "currency_pair": string, 
    "amount": float,
    "identifier": UUID (128 bit string - 36 characters),
    "price": float,
    "trader_id": int,
    "trade_id" int,
    "trade_date" DateTime
}

or
Array of trades sorted by trade_date