Skip to content

Latest commit

 

History

History
 
 

OrderBook

README.md

How to run order_book.py script

Standard approach

Firstly you need to be in the OrderBook/app directory and then execute in the terminal the following command:

./process_records.zsh

This is a zsh script that reads from standard input and saves data to a data.json file. In the backgroud order_book.py program will be running.

Additionally, you can use defined functions in the process_records.zsh script (this is additional funcionallity):

  • clear : that will delete everything from data.json

  • edit_record : this will update selected record

  • remove_record : this will remove a selected record from order book (data.json)

Each time a new order is added/removed or updated the order_book.py script will be executed and the final status of the order book will be displayed to standard output (in so called standard approach).

As an additional feature of process_records.zsh - input orders are also saved to the history.json file. For tracking purposes.

Docker container

As an alternative approach, docker image was built. The assumption here is that you have Docker / Docker Desktop installed on your machine. Then easily, the docker image could be built from Dockerfile which is part of this folder. With following instructions.

Docker container is built based on official images. Moreover, as a simple way to interact with running containers Flask app was developed. You can get access to running image in your web browser on 5001 port - local host.

Go to root directory (Dockerfile level) and next:

docker build -t order_book_image .

docker run -d --name order_book_container -p 5001:5001 order_book_image

You can acess that under following http:

http://127.0.0.1:5001

When you want to stop docker container

docker stop order_book_container

When you want delete docker container

docker -rm order_book_container

This way of building app has almost the same functionality as standard approach. Nevertheless there are some differences:

  • each time when you want provide new record, you need to manually delete previous one

  • you do not have functionality to update or delete particullar record

  • you can clear data.json file just providing clear key word to box

  • as output you will receive final status of order book and matched orders (ransactions) but that is not formatted

Software requirements

Example of input data

{"type": "iceberg", "order": {"direction": "sell", "id": 1, "price": 100, "quantity": 200,"peak": 100}}
{"type": "iceberg", "order": {"direction": "sell", "id": 2, "price": 100, "quantity": 300,"peak": 100}}
{"type": "iceberg", "order": {"direction": "sell", "id": 3, "price": 100, "quantity": 200,"peak": 100}}
{"type": "iceberg", "order": {"direction": "buy", "id": 4, "price": 100, "quantity": 500,"peak": 100}}

  • After each new order is pasted to stdin press Enter / Return on your keyboard (docker approach is different at this point, please read above)

  • When you want to stop program press CTRL+D (I'm using UNIX base OS)

More business logic and explanation

Explanations and examples could be found in the general_description.md file. Where description was provided. The reader wants to read this file twice before attempting to understand the implementation.

Additional thoughts - optimization

From an efficient data structure perspective it seems to be a good idea to use heapq (https://docs.python.org/3/library/heapq.html) library.

Nevertheless in the cons.md file has been written some optimization steps that could be a perfect fit for a task similar to this one.

As PoC of using heapq one function has been developed: alternative_match_orders.py. But that was not tested in an end-to-end solution.

TODO

  • Formatting: to be in line with description of the task. As it is not implemented yet: final OrderBook status and transactions

Other instruction for developers

pre-commit

  • Install pre-commit tool by running pip install pre-commit.

  • Create a .pre-commit-config.yaml file in the root directory of the project.

  • Define the pre-commit hooks you want to use in the file, using the appropriate syntax.

  • Run pre-commit install to install the hooks specified in the configuration file to the project's Git repository.

  • Run pre-commit run to execute the hooks on the files that are about to be committed.

Tests

Tests are in OrderBook/tests. Pytest was used for those purposes.

In your terminal, make sure you are in the OrderBook directory:

cd OrderBook

Run the tests from the OrderBook directory:

pytest tests/test_generic.py

or

pytest tests/test_orderbook.py

Additional resources

Reinforcement Learning and OrderBook example and mathematical proof.