lemon.markets Python SDK

lemon.markets Python SDK facilitates communication with the lemon.markets API for Python programs. The library implements all calls to the endpoints defined in the Market Data API and Trading API. We currently do not support asynchronous calls.

Documentation

See the API docs.

Installation

You can install library using pip:

pip install lemon

Requirements:

• Python 3.7+
• requests
• pytz
• typing-extensions

Usage

SDK client

To create and configure our SDK client you will need to have separate API tokens for Market Data API and Trading API. You also need to choose which environment you want to use for trading: paper or live, defaults to paper. For how to obtain and use our api token, see here. The snippet below shows how to properly create a SDK client:

from lemon import api

client = api.create(
    market_data_api_token='your-market-data-api-token',
    trading_api_token='your-trading-api-token',
    env='paper'  # or env='live'
)

lemon.api.create method allows also to configure:

• timeout - default timeout for requests
• retry_count - default number of retries for requests
• retry_backoff_factor - default retry backoff factor for retries

The SDK client consists of three parts:

• market_data - let's you access the Market Data API endpoints
• streaming - let's you retrieve an authentication token that you can use to stream live data
• trading - let's you access the Trading API endpoints. Choose the desired target environment (paper or live) in the client configuration.

Market Data API usage

from lemon import api
from datetime import datetime, timezone

client = api.create(
    market_data_api_token='your-market-data-api-token',
    trading_api_token='your-trading-api-token',
)

# get venues
response = client.market_data.venues.get()
print(response.results)

# get instruments
response = client.market_data.instruments.get(
    isin=["US88160R1014", "US0231351067"]
)
response = client.market_data.instruments.get(
    tradable=True,
)
# automatically iterate over all pages. auto_iter() is available on all list responses
for instrument in response.auto_iter():
    print(instrument)

response = client.market_data.instruments.get(
    type=['stock', 'etf'],
    currency=['EUR'],
    limit=10,
    sorting='asc'
)

# get latest ohlc
response = client.market_data.ohlc.get(
    isin=['US88160R1014'],
    period='m1',
    from_='latest',
    epoch=True,
    decimals=True
)

# get ohlc
response = client.market_data.ohlc.get(
    isin=['US88160R1014'],
    period='d1',
    from_=datetime(2021, 1, 2)
)

# get latest quotes
response = client.market_data.quotes.get_latest(
    isin=['US88160R1014', 'US0231351067'],
    epoch=True,
    sorting='asc'
)

# get historical quotes
# for period <timestamp, timestamp + 1 day)
response = client.market_data.quotes.get(
    isin="US88160R1014",
    from_=datetime(2022, 10, 5, tzinfo=timezone.utc),
)
# for period <timestamp - 1 day, timestamp)
response = client.market_data.quotes.get(
    isin="US88160R1014",
    to=datetime(2022, 10, 5, tzinfo=timezone.utc),
)
# for specific period <from, to) - timedelta has to be <= 1 day
response = client.market_data.quotes.get(
    isin="US88160R1014",
    from_=datetime(2022, 10, 5, tzinfo=timezone.utc),
    to=datetime(2022, 10, 5, 15,  tzinfo=timezone.utc),
)
# if you don't provide from/to - endpoint works the same as 'client.market_data.quotes.get_latest'
response = client.market_data.quotes.get(isin="US88160R1014")

# get trades
response = client.market_data.trades.get_latest(
    isin=['US88160R1014', 'US0231351067'],
    decimals=True
)

# get historical trades
# for period <timestamp, timestamp + 1 day)
response = client.market_data.trades.get(
    isin="US88160R1014",
    from_=datetime(2022, 10, 5, tzinfo=timezone.utc),
)
# for period <timestamp - 1 day, timestamp)
response = client.market_data.trades.get(
    isin="US88160R1014",
    to=datetime(2022, 10, 5, tzinfo=timezone.utc),
)
# for specific period <from, to) - timedelta has to be <= 1 day
response = client.market_data.trades.get(
    isin="US88160R1014",
    from_=datetime(2022, 10, 5, tzinfo=timezone.utc),
    to=datetime(2022, 10, 5, 15,  tzinfo=timezone.utc),
)
# if you don't provide from/to - endpoint works the same as 'client.market_data.trades.get_latest'
response = client.market_data.trades.get(isin="US88160R1014")

Streaming API Usage

from lemon import api

client = api.create(...)

# get live streaming authentication token
response = client.streaming.authenticate()

Streaming API example

This example relies on that you have both this SDK installed as well as paho-mqtt package.

Below is an example usage of live streaming quotes through alby mqtt broker using paho mqtt client. When connecting to the broker the on_connect callback will be triggered. This in return will trigger the on_subscribe callback where we can let the broker know what ISINS we are interested in There is a limitation to only have 4 channels connected at once. You may be able to create more than 4 channels - however we then may close any one of them at any time.

After that we will simply get all the quote updates through the on_message callback.

from lemon import api
from lemon.market_data.model import Quote
import paho.mqtt.client as mqtt
import json

client = api.create(...)

# get live streaming authentication token
response = client.streaming.authenticate()

def on_connect(mqtt_client, userdata, flags, rc):
    mqtt_client.subscribe(response.user_id)

def on_subscribe(mqtt_client, userdata, level, buff):
    mqtt_client.publish(f"{response.user_id}.subscriptions", "US88160R1014,US0231351067")

def on_message(mqtt_client, userdata, msg):
    data = json.loads(msg.payload)
    quote = Quote._from_data(data, int, int)

# initiate mqtt- client for streaming
mqtt_client = mqtt.Client("Ably_Client")
mqtt_client.username_pw_set(username=response.token)
mqtt_client.on_connect = on_connect # callbck to handle connect
mqtt_client.on_subscribe = on_subscribe # callbck to handle subscribe
mqtt_client.on_message = on_message # callbck to handle message

mqtt_client.connect("mqtt.ably.io")
mqtt_client.loop_forever() # start the mqtt client and loop forever

Trading API usage

from lemon import api

client = api.create(...)

# create buy order
response = client.trading.orders.create(
    isin='US88160R1014',
    side='buy',
    quantity=1,
)
order_id = response.results.id

# activate buy order
response = client.trading.orders.activate(order_id=order_id)

# get buy order status
response = client.trading.orders.get_order(order_id=order_id)

# get orders
response = client.trading.orders.get()

# iterate over all orders of all pages
for order in response.auto_iter():
    print(order)

# cancel order
# create an order first
response = client.trading.orders.create(
    isin='US88160R1014',
    side='buy',
    quantity=1,
)
# cancel the order via order_id
response = client.trading.orders.cancel(order_id=response.results.id)

# update account
response = client.trading.account.update(
    address_street='Ritterstrasse',
    address_city="Berlin"
)

# withdraw money from account
response = client.trading.account.withdraw(
    amount=100000,
    pin="1234"
)

# get bank statements
response = client.trading.account.get_bank_statements(
    type='eod_balance',
    from_="beginning"
)

# get documents
response = client.trading.account.get_documents()

# get document
response = client.trading.account.get_document(document_id='doc_xyz')

# get user
response = client.trading.user.get()

# get positions
response = client.trading.positions.get(isin='US88160R1014')

# get statements
response = client.trading.positions.get_statements()

# get performance
response = client.trading.positions.get_performance()

Direct API calls

It is possible to call Market Data API/Trading API endpoints directly by providing path to the endpoint and optionally - query parameters and payload to be sent. SDK currently supports GET|POST|PUT|DELETE methods.

from lemon import api

client = api.create(...)

response = client.trading.put(url="account", json={"address_street": 'New street name'})
response = client.market_data.get(url="instruments", params={"search": "t*sla"})

Error handling

Error structure available in the SDK is presented below

• BaseLemonError - base class for all errors thrown within the SDK
  • LemonError - base class of errors returned by API. It contains information directly from it, such as status, error_code, error_message, ...
    • InvalidQueryError - HTTP request validation error
    • AuthenticationError - authentication error
    • InternalServerError - internal API error
    • BusinessLogicError - a business logic error caused by specific conditions of the system preventing from fulfilling the request
  • APIError - error thrown in case when it's impossible to decode the API response

Please note that errors raised from requests module are passed as they are.

from lemon import api, errors

client = api.create(...)

try:
    response = client.trading.orders.create(isin='...', side='buy', quantity=1)
except errors.InvalidQueryError:
    ...
except errors.AuthenticationError:
    ...
except errors.InternalServerError:
    ...
except errors.BusinessLogicError as err:
    if err.error_code == 'some_specific_error':
        print(err.error_message)
        ... # do something
    ...
except errors.LemonError:  # InvalidRequestError/AuthenticationError/InternalServerError/BusinessLogicError
    ...
except errors.APIError:
    ...
except errors.BaseLemonError:  # catches all errors defined above
    ...
except:  # other errors, including errors from `requests` module
    ...

Model serialization

Every response(or response nested structure) can be serialized to python dictionary or JSON:

from lemon import api

client = api.create(
    market_data_api_token='your-market-data-api-token',
    trading_api_token='your-trading-api-token',
)
response = client.market_data.instruments.get()

print(response.dict())
print(response.json())
print(response.results[0].dict())
print(response.results[0].json())