A simple text-b 80C5 ased terminal/command prompt application that connects to the WebSocket market data feed of Gemini, a digital asset exchange. On each update it prints out the best bid and ask price and quantity of the BTCUSD asset pair.
# optionally use venv to manage python enviroments
python3 -m venv venv
source ./venv/bin/activate
# install dependencies
pip3 install -r requirements.txt
# run program
python3 orderbook.pyadditionally you can use the provided Dockerfile to build a runnable container image
# build image from Dockerfile and name it orderbook
docker build -t orderbook .
# create and run a container named: orderbookcontainer using the image we created earlier
docker run -d --name orderbookcontainer orderbook
# get snapshot of stdout from running container
docker logs orderbookcontainerpython3 orderbook.py| best bid | quantity | best ask | quantity |
|---|---|---|---|
| 6748.70 | 0.03700000 | 6748.71 | 4.27506690 |
| 6739.70 | 0.20000000 | 6739.71 | 4.63391087 |
Whenever any of these values changes, the application prints out a line with the latest values. ## Additonal Information
The protocol specification for the WebSocket market data feed is available at:
https://docs.gemini.com/websocket-api/#market-data
For the purposes of this assignment, you should either omit the top_of_book parameter or set it to false. You are free to implement the application in a programming language of your choice. You are also free to use any appropriate libraries e.g. for handling JSON and WebSocket. However, you should not use any ready-made Gemini client libraries. Hint: If you use an efficient data structure for the orderbook reconstruction then this is a plus.
To represent both sides of the book I use two identical defaultdict(list) structures for ASK and BID
Since Python 3.7 dicts maintain insertion order
If I were to redo this; I would change my implementation.
- Class Order: Double linked list data structure referencing Orders in same price level. Arranged by price time priority
class Order:
def __init__(self, price, quantity, timestamp):
self.price = price
self.quantity = quantity
self.timestamp = timestamp
self.prev = None
self.next = None- Class PriceLevel: would represent all orders at a given price and this class would handle responsibilities like fufilling a trade.
class PriceLevel:
def __init__(self, price, quantity):
self.price = price
self.quantity = quantity
self.head = None- Class Orderbook: representing both sides of the book (bid/ask), possible use a sortedcontainers. SortedDict here or use perhaps use pythons std lib heapq which is a minheap (bid) and inverse heapq for a maxheap (asks) Would have helper methods suchas: add_order() remove_order() add_level() remove_level() print_summary()
class Orderbook:
def __init__(self):
"""
self.bids =>
{ (price): PriceLevel(1, 1, Side.BUY) }
{ (price): PriceLevel(2, 1, Side.BUY) }
self.asks =>
{ (price): PriceLevel(2, 1, Side.SELL) }
{ (price): PriceLevel(3, 1, Side.SELL) }
"""
self.bids = sortedcontainers.SortedDict(PriceLevel)
self.asks = sortedcontainers.SortedDict(PriceLevel)
def add_order(self):
pass
def remove_order(self):
pass
def add_level(self, price, qty):
pass
def remove_level(self, price, qty):
pass
def print_summary():
pass