The Cryptowatch Python library provides a convenient access to the Cryptowatch API from applications written in the Python language.
It includes the following features:
- Auto-serialization of API responses into Python objects
- Websocket client with transparent handling of protobuf messages
- API credentials automatically read from your
~/.cw/credentials.yml
config file - Custom exceptions for API-specific issues (e.g.: Requests Allowance)
- Smart back-off retries in case of API connectivity loss
pip install cryptowatch-sdk
The cryptowatch
library is not related with Cryptowatch. If you installed it by mistake run pip uninstall cryptowatch
to remove it.
The correct library name is cryptowatch-sdk
.
Showing all Kraken markets that already gained at least 5% over the current weekly candle.
import cryptowatch as cw
from datetime import datetime, timedelta
# Get all Kraken markets
kraken = cw.markets.list("kraken")
# For each Kraken market...
for market in kraken.markets:
# Forge current market ticker, like KRAKEN:BTCUSD
ticker = "{}:{}".format(market.exchange, market.pair).upper()
# Request weekly candles for that market
candles = cw.markets.get(ticker, ohlc=True, periods=["1w"])
# Each candle is a list of [close_timestamp, open, high, low, close, volume, volume_quote]
# Get close_timestamp, open and close from the most recent weekly candle
close_ts, wkly_open, wkly_close = (
candles.of_1w[-1][0],
candles.of_1w[-1][1],
candles.of_1w[-1][4],
)
# Compute market performance, skip if open was 0
if wkly_open == 0:
continue
perf = (wkly_open - wkly_close) * 100 / wkly_open
# If the market performance was 5% or more, print it
if perf >= 5:
open_ts = datetime.utcfromtimestamp(close_ts) - timedelta(days=7)
print("{} gained {:.2f}% since {}".format(ticker, perf, open_ts))
- python v3.7+
- requests v0.8.8+
- marshmallow v3.2.2+
- pyyaml v5.1.2+
- websocket-client v0.56+
- protobuf v3.11.3+
Using a credentials file will allow you to authenticate your requests and grant you the API access of your Cryptowatch account.
Your account Credits will be consumed for the REST and WebSocket API. Specific Credit cost details can be found on the Pricing page.
-
Generate an Cryptowatch API key from your account
-
Create your credential file on your machine by running in order:
2.1
mkdir $HOME/.cw
2.2
echo "apikey: 123" > $HOME/.cw/credentials.yml
(where123
is your 20 digits public key) -
Verify with
cat $HOME/.cw/credentials.yml
that you see something like below (123
being your public key):
apikey: 123
The SDK will read your public key as soon as import cryptowatch
is ran in your script.
import cryptowatch as cw
# Set your API Key, it is by default read from your ~/.cw/credentials.yml file
cw.api_key = "123"
# Assets
cw.assets.list()
cw.assets.get("BTC")
# Exchanges
cw.exchanges.list()
cw.exchanges.get("KRAKEN")
# Instruments
cw.instruments.list()
cw.instruments.get("BTCUSD")
# Markets
cw.markets.list() # Returns list of all markets on all exchanges
cw.markets.list("BINANCE") # Returns all markets on Binance
# Returns market summary (last, high, low, change, volume)
cw.markets.get("KRAKEN:BTCUSD")
# Return market candlestick info (open, high, low, close, volume) on some timeframes
cw.markets.get("KRAKEN:BTCUSD", ohlc=True, periods=["4h", "1h", "1d"])
# Returns market last trades
cw.markets.get("KRAKEN:BTCUSD", trades=True)
# Return market current orderbook
cw.markets.get("KRAKEN:BTCUSD", orderbook=True)
# Return market current orderbook liquidity
cw.markets.get("KRAKEN:BTCUSD", liquidity=True)
You can access the raw HTTP response received via the _http_response
attribute which is a requests.Response
object:
import cryptowatch as cw
bitcoin = cw.assets.get('btc')
print(bitcoin._http_response)
import cryptowatch as cw
# Set your API Key
cw.api_key = "123"
# Subscribe to resources (https://docs.cryptowat.ch/websocket-api/data-subscriptions#resources)
cw.stream.subscriptions = ["markets:*:trades"]
# What to do on each trade update
def handle_trades_update(trade_update):
"""
trade_update follows Cryptowatch protocol buffer format:
https://github.com/cryptowatch/proto/blob/master/public/markets/market.proto
"""
market_msg = ">>> Market#{} Exchange#{} Pair#{}: {} New Trades".format(
trade_update.marketUpdate.market.marketId,
trade_update.marketUpdate.market.exchangeId,
trade_update.marketUpdate.market.currencyPairId,
len(trade_update.marketUpdate.tradesUpdate.trades),
)
print(market_msg)
for trade in trade_update.marketUpdate.tradesUpdate.trades:
trade_msg = "\tID:{} TIMESTAMP:{} TIMESTAMPNANO:{} PRICE:{} AMOUNT:{}".format(
trade.externalId,
trade.timestamp,
trade.timestampNano,
trade.priceStr,
trade.amountStr,
)
print(trade_msg)
cw.stream.on_trades_update = handle_trades_update
# Start receiving
cw.stream.connect()
# Call disconnect to close the stream connection
# cw.stream.disconnect()
See this script for more streaming example.
If you need to convert the protobuf message to JSON, you can do so with MessageToJson
. See the example below:
from google.protobuf.json_format import MessageToJson
import cryptowatch as cw
# Set your API Key
cw.api_key = "123"
# Subscribe to resources (https://docs.cryptowat.ch/websocket-api/data-subscriptions#resources)
cw.stream.subscriptions = ["markets:*:trades"]
# What to do on each trade update
def handle_trades_update(trade_update):
"""
trade_update follows Cryptowatch protocol buffer format:
https://github.com/cryptowatch/proto/blob/master/public/markets/market.proto
"""
MessageToJson(trade_update)
cw.stream.on_trades_update = handle_trades_update
# Start receiving
cw.stream.connect()
Logging can be enabled through Python's logging
module:
import logging
logging.basicConfig()
logging.getLogger("cryptowatch").setLevel(logging.DEBUG)
The library exposes a simple utility, named cryptowatch
, to return last market prices.
> cryptowatch
7425.0
> cryptowatch btceur
6758.1
> cryptowatch binance:ethbtc
0.020359
When the market doesn't exist a return code of 1
will be set (0
otherwise):
> cryptowatch binance:nosuchmarketusd
> echo $?
1
Unit tests are under the tests folder and use pytest
, run them all with:
make test
Integration tests sending real HTTP requests to the Cryptowatch API can be run with:
make test-http-real
Testing and developement dependencies are in the requirements.txt file, install them with:
pip install -r requirements.txt
The code base use the Black linter, run it with:
make lint