Skip to main content

Get ohlc from kraken web sockets.

Project description

aio-kraken-ws pipeline status coverage report PEP8

A module to collect ohlc candles from Kraken using WebSockets that is asyncio friendly! Looking for automated trading tools? Botcrypto may interest you.

Key features

  • Subscribe to kraken data using a single WebSocket.
  • Trigger a callback that is coroutine on each new closed candles from kraken.
  • Easy subscription/unsubscription to datasets, i.e. [(PAIR, Unit Time)].
  • Callback is regulary triggered at each end of the UT intervals, whatever is the number of data received by kraken.

Getting started

Install

pip install aio-kraken-ws

Usage

# check tests/learning/log_to_file.py for a complete example
async def callback(pair, interval, timestamp, o, h, l, c, v):
    """ A coroutine handling new candles.

    :param str pair:
    :param int interval: time in minutes
    :param int timestamp: candle open timestamp.
    :param float o: open price
    :param float h: high price
    :param float l: low price
    :param float c: close price
    :param float v: volume
    """
    with open("candles.txt", "a+") as file:
        file.write(f"[{pair}:{interval}]({timestamp},{o},{h},{l},{c},{v})\n")

kraken_ws = await KrakenWs.create(callback)
# subscribe to some datasets
kraken_ws.subscribe([("XBT/EUR", 1), ("ETH/EUR", 5)])

The callback function is called for each dataset at the end of each dataset's unit time interval.

E.g. if subscription start at 4h42.05 to the dataset ("XBT/EUR", 1), then callback is triggered at 4h43.00, at 4h44.00, at 4h45.00, etc... For ("XBT/EUR", 60), it would be at 5h00.00, at 6h00.00, etc... It's possible to get at most 10ms delay between the exact UT interval ending and the actual datetime of the call.

If no new data were received from Kraken during an interval, the callback is triggered with the latest known close price and v=0, as it's described in the following example.

E.g.

kraken_ws.subscribe([("XBT/EUR", 1)])
# time.time() = 120
await callback("XBT/EUR", 1, 60, 42.0, 57.0, 19.0, 24.0, 150.0)
# time.time() = 180
await callback("XBT/EUR", 1, 120, 19.0, 24.0, 8.0, 10.0, 13.0)
# time.time() = 240 : no data received in 60s, i.e. no activity
await callback("XBT/EUR", 1, 180, 10.0, 10.0, 10.0, 10.0, 0.0)

An exception raised by the callback will be logged and it wont stop the streams.

Warning

The callback should takes less than a minute to process. If the callback takes more than a minutes, a warning is emitted and you may lose market data.

Kraken WebSocket server seems to admit 20 subscriptions commands maximum. Further commands seems to be ignored by the server. Hopfully, aio-kraken-ws manage this limitation for you!

Moreover, after 24h a subscriptions seem to expire and no more market data is receive. To ensure we do not lose the stream of market data, aio-kraken-ws automatically reconnect and re-subscribe to the datasets every 30 minutes.

Tests

You can find a wroking example of KrakenWs in tests/learning/log_to_file.py.

Run tests locally

Clone the repo and install requirements

pip install -e .[test]

Run the suite tests

# unit tests - no call to kraken - fast
pytest --cov=aio_kraken_ws --cov-report= -v tests/unit

# integration tests - actual kraken subscription - slow
pytest --cov=aio_kraken_ws --cov-append -v -n 8 tests/integration

Changelog

See https://cdlr75.gitlab.io/aio-kraken-ws/CHANGELOG.html

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

aio-kraken-ws-0.2.4.tar.gz (17.2 kB view hashes)

Uploaded Source

Built Distribution

aio_kraken_ws-0.2.4-py3-none-any.whl (14.6 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page