Reactive Markets Platform SDK Python

The Reactive Platform SDK for Python.

Getting Started

Install

The package only support python3.7 and python3.8, not for lower versions.

Install the package in a python environment:

$ pip install reactive-platform

Uninstall the package:

$ pip uninstall reactive-platform

Using REST API

The REST API has endpoints for different type data, e.g. orders, analytics and references. At this moment, this package only supports querying reference data. To access the platform, a API token must be granted from platform UI.

Create a REST client

Create a client first:

from reactive.platform.rest.client import Client

key = 'xxx'
url = "https://api.platform.reactivemarkets.com"
rc = Client(url=url, key=key)

Reference Data

Reference data includes asset, instrument, venue and markets. Get those data:

asset_ref = rc.fetch_asset_ref()
# json format assets reference data
assets = asset_ref.to_json()
print("asset reference:\n", assets)
# dict format assets
assets_dict = asset_ref.to_dict()
print("assets dict:\n", assets)

# instr reference
instr_ref = rc.fetch_instr_ref()
instrs = instr_ref.to_json()
print("instrument reference:\n", instrs)

# venue reference
venue_ref = rc.fetch_venue_ref()
venues = venue_ref.to_json()
print("venue reference:\n", venues)

# market reference
market_ref = rc.fetch_market_ref()
markets = market_ref.to_json()
print("market reference:\n", venues)

If client uses pandas, the reference data can be viewed in a Dataframe table in jupyter-notebook:

import pandas as pd

df = pd.read_json(market_ref.to_json())
df

Feed Gateway WebSocket API

The websocket feed provides real-time level 2 market data snapshots and public trades via

wss://api.platform.reactivemarkets.com/feed

WebSocket Client

To access the feed gateway via web socket, create a Client which manages the web socket connection and provides methods to access the gateway. See the example in example/wsclient.py.

The message protocol via feed gateway is Flatbuffers, which provides an efficient serialization/deserializaton mechanism in terms of both processing and space requirements. The reactive-platform generated python classes Flatbuffer schema are located under the dependency reactive.papi.

Client object decodes into reactive.papi.Message.Message and allow user to use customised data_handler call back to handle the flatbuffer message.

Create a FeedClient

The platform-py also provides another option, FeedClient.

from reactive.platform.feed.client import FeedClient

TOKEN = 'xxx'
addr = "wss://api.platform.reactivemarkets.com/feed"
feed_client = FeedClient(addr=addr, key=TOKEN)

FeedClient provides methods subscribe and unsubscribe to control sending request to platform, and implement python classes for corresponding flatbuffer data types in reactive.papi. see a full example in example/marketdata.py or example/trade. The client can specify book view parameters in the request. Currently, the feed gateway supports, book depths (1, 5, 10, 20) and tick grouping (1, 50).