search

yiwensong / coinbasepro-python

The unofficial Python client for the Coinbase Pro API
MIT License
0 stars 2 forks source link
coinbase cryptocurrency python python3 python36 python37

readme

coinbasepro-python

Build Status Coverage Status Documentation Status PyPI version

The Python client for the Coinbase Pro API (formerly known as the GDAX)

Note

This is a fork of the original cbpro project. I created this fork because between September and October of 2018, the maintainers of that project seemed to be inactive in reviewing and merging pull requests. The goal of this fork is as follows:

Benefits

Under Development

Getting Started

This README is documentation on the syntax of the python client presented in this repository. See function docstrings for full syntax details.
This API attempts to present a clean interface to CB Pro, but in order to use it to its full potential, you must familiarize yourself with the official CB Pro documentation.

Public Client

Only some endpoints in the API are available to everyone. The public endpoints can be reached using PublicClient

import cbpro
public_client = cbpro.PublicClient()

PublicClient Methods

Authenticated Client

Not all API endpoints are available to everyone. Those requiring user authentication can be reached using AuthenticatedClient. You must setup API access within your account settings. The AuthenticatedClient inherits all methods from the PublicClient class, so you will only need to initialize one if you are planning to integrate both into your script.

import cbpro
auth_client = cbpro.AuthenticatedClient(key, b64secret, passphrase)
# Use the sandbox API (requires a different set of API access credentials)
auth_client = cbpro.AuthenticatedClient(key, b64secret, passphrase,
                                  api_url="https://api-public.sandbox.pro.coinbase.com")

Pagination

Some calls are paginated, meaning multiple calls must be made to receive the full set of data. The CB Pro Python API provides an abstraction for paginated endpoints in the form of generators which provide a clean interface for iteration but may make multiple HTTP requests behind the scenes. The pagination options before, after, and limit may be supplied as keyword arguments if desired, but aren't necessary for typical use cases.

fills_gen = auth_client.get_fills()
# Get all fills (will possibly make multiple HTTP requests)
all_fills = list(fills_gen)

One use case for pagination parameters worth pointing out is retrieving only new data since the previous request. For the case of get_fills(), the trade_id is the parameter used for indexing. By passing before=some_trade_id, only fills more recent than that trade_id will be returned. Note that when using before, a maximum of 100 entries will be returned - this is a limitation of CB Pro.

from itertools import islice
# Get 5 most recent fills
recent_fills = islice(auth_client.get_fills(), 5)
# Only fetch new fills since last call by utilizing `before` parameter.
new_fills = auth_client.get_fills(before=recent_fills[0]['trade_id'])

AuthenticatedClient Methods

WebsocketClient

If you would like to receive real-time market updates, you must subscribe to the websocket feed.

Subscribe to a single product

import cbpro
# Paramters are optional
wsClient = cbpro.WebsocketClient(url="wss://ws-feed.pro.coinbase.com", products="BTC-USD")
# Do other stuff...
wsClient.close()

Subscribe to multiple products

import cbpro
# Paramaters are optional
wsClient = cbpro.WebsocketClient(url="wss://ws-feed.pro.coinbase.com",
                                products=["BTC-USD", "ETH-USD"])
# Do other stuff...
wsClient.close()

WebsocketClient + Mongodb

The WebsocketClient now supports data gathering via MongoDB. Given a MongoDB collection, the WebsocketClient will stream results directly into the database collection.

# import PyMongo and connect to a local, running Mongo instance
from pymongo import MongoClient
import cbpro
mongo_client = MongoClient('mongodb://localhost:27017/')

# specify the database and collection
db = mongo_client.cryptocurrency_database
BTC_collection = db.BTC_collection

# instantiate a WebsocketClient instance, with a Mongo collection as a parameter
wsClient = cbpro.WebsocketClient(url="wss://ws-feed.pro.coinbase.com", products="BTC-USD",
    mongo_collection=BTC_collection, should_print=False)
wsClient.start()

WebsocketClient Methods

The WebsocketClient subscribes in a separate thread upon initialization. There are three methods which you could overwrite (before initialization) so it can react to the data streaming in. The current client is a template used for illustration purposes only.

wsClient = myWebsocketClient() wsClient.start() print(wsClient.url, wsClient.products) while (wsClient.message_count < 500): print ("\nmessage_count =", "{} \n".format(wsClient.message_count)) time.sleep(1) wsClient.close()


### Real-time OrderBook
The ```OrderBook``` subscribes to a websocket and keeps a real-time record of
the orderbook for the product_id input.  Please provide your feedback for future
improvements.

```python
import cbpro, time
order_book = cbpro.OrderBook(product_id='BTC-USD')
order_book.start()
time.sleep(10)
order_book.close()

Development

Development Environment

A development environment can be created and activated with the commands

make venv && . venv/bin/activate

Automated Testing

Unit test framework uses pytest, coverage, and tox. Tests are a work in progress, please contribute to the coverage.

make test

Pre-commit

This project uses pre-commit to enforce coding style. You can run pre-commit with the following command (after activating the venv)

pre-commit run --files [FILES]

You can also run

pre-commit run --all-files

make test also runs pre-commit as a part of its testing routine.