tardis-machine
Locally installable high level API server & caching layer for tardis.dev API - historical tick-level crypto markets data replay service.
Provides easy to use, high level streaming HTTP and WebSocket endpoints that allow replaying supported historical crypto markets WebSocket data feeds from any past point in time.
Check out tardis-client
as well if you're using Node.js.
HTTP endpoint
Accessible via /replay/:<EXCHANGE>?from=<FROM_DATE>&to=<TO_DATE>&filters=<FILTERS>
Allows replaying historical crypto exchange WebSocket data feed via streaming HTTP response (new line delimited JSON).
Returns up to 400 000 messages per second (depending on the machine set-up and local cache ratio).
Example requests:
http://localhost:8000/replay/bitmex?from=2019-05-01&to=2019-05-02&filters=[{"channel":"trade"}]
- returns all trades that happened on BitMEX on 1st of May 2019 in ndjson format.http://localhost:8000/replay/bitmex?from=2019-04-01&to=2019-04-02&filters=[{"channel":"trade","symbols":["XBTUSD","ETHUSD"]},{"channel":"orderBookL2","symbols":["XBTUSD"]}]
- returns all trades that happened on BitMEX on 1st of April 2019 for XBTUSD
and ETHUSD
as well as orderbookL2
snapshot and updates for XBTUSD
in ndjson format.
Available params:
name | type | default value | description |
---|
exchange | string | - | requested exchange name. Check out allowed echanges |
from | string | - | replay period start date (UTC) - eg: 2019-04-05 or 2019-05-05T00:00:00.000Z |
to | string | - | replay period end date (UTC) - eg: 2019-04-05 or 2019-05-05T00:00:00.000Z |
filters (optional) | Url encoded JSON string with {channel:string, symbols?: string[]}[] structure | undefined | optional filters for requested data feed. Check out allowed channels for each exchange, up to date list of accepted symbols can be found via https://tardis.dev/api/v1/exchanges/<EXCHANGE_NAME> API call |
Example response snippet:
{"localTimestamp":"2019-05-01T00:09:42.2760012Z","message":{"table":"orderBookL2","action":"update","data":[{"symbol":"XBTUSD","id":8799473750,"side":"Buy","size":2333935}]}}
{"localTimestamp":"2019-05-01T00:09:42.2932826Z","message":{"table":"orderBookL2","action":"update","data":[{"symbol":"XBTUSD","id":8799474250,"side":"Buy","size":227485}]}}
{"localTimestamp":"2019-05-01T00:09:42.4249304Z","message":{"table":"trade","action":"insert","data":[{"timestamp":"2019-05-01T00:09:42.407Z","symbol":"XBTUSD","side":"Buy","size":1500,"price":5263,"tickDirection":"ZeroPlusTick","trdMatchID":"29d7de7f-27b6-9574-48d1-3ee9874831cc","grossValue":28501500,"homeNotional":0.285015,"foreignNotional":1500}]}}
{"localTimestamp":"2019-05-01T00:09:42.4249403Z","message":{"table":"orderBookL2","action":"update","data":[{"symbol":"XBTUSD","id":8799473700,"side":"Sell","size":454261}]}}
{"localTimestamp":"2019-05-01T00:09:42.4583155Z","message":{"table":"orderBookL2","action":"update","data":[{"symbol":"XBTUSD","id":8799473750,"side":"Buy","size":2333838},{"symbol":"XBTUSD","id":8799473800,"side":"Buy","size":547746}]}}
Response structure (single line):
localTimestamp
is a date when message has been received in ISO 8601 format with 100 nano second resolution.
message
is and JSON object/array with exactly the same structure as provided by particular exchange.
Check out tests to see how such messages stream can be easily consumed using Node.js streams and async iterators.
WebSocket endpoint
Accessible via /ws-replay/?exchange=<EXCHANGE>?from=<FROM_DATE>&to=<TO_DATE>
Exchanges & various 3rd party data providers WebSocket APIs allows subscribing only real-time data feeds and there is no way to subscribe to and "replay" market from any point in the past. Using tardis-machine
that is no longer the case.
Example:
Example below shows how to subscribe to BitMEX historical data feed (from 2019-06-01 to 2019-06-02) to trade:XBTUSD
and orderBookL2:XBTUSD
channels and replay such stream as if it was a real-time one. Subscribe
request messages format as well as received messages format is exactly the same as in native exchanges WebSocket APIs.
In many cases such websocket historical data feeds can be consumed using already available WebSocket clients for various exchanges. Imagine having single 'data pipeline' for real-time trading and backtesting.
const ws = new WebSocket('ws://localhost:8000/ws-replay?exchange=bitmex&from=2019-06-01&to=2019-06-02')
ws.onmessage = message => {
console.log(message)
}
ws.onopen = () => {
const subscribePayload = {
op: 'subscribe',
args: ['trade:XBTUSD', 'orderBookL2:XBTUSD']
}
ws.send(JSON.stringify(subscribePayload))
}
Check out tests for more examples.
Available query string params:
In contrast to HTTP API for WebSocket API filters aren't provided explicitly as those are created based on subscribe
messages over WebSocket received in first 5 seconds of the connection.
name | type | default value | description |
---|
exchange | string | - | requested exchange name. Currently supported: bitmex, coinbase, deribit, cryptofacilities, bitstamp, okex |
from | string | - | replay period start date (UTC) - eg: 2019-04-05 or 2019-05-05T00:00:00.000Z |
to | string | - | replay period start date (UTC) - eg: 2019-04-05 or 2019-05-05T00:00:00.000Z |
Installation
-
npx
(requires Node.js v12 installed on host machine)
That will start tardis-machine server running on port 8000
by default (port can be changes --port
)
npx tardis-machine --api-key=YOUR_API_KEY
If you'd like to test it out on sample data (first full day of each month) simply run the command without --api-key
option.
npx tardis-machine
Run npx tardis-machine --help
to see all available options (setting custom cache dir etc.)
-
npm
(requires Node.js v12 installed on host machine)
Installs tardis-machine
globally.
npm install -g tardis-machine
tardis-machine --api-key=YOUR_API_KEY
If you'd like to test it out on sample data (first full day of each month) simply run the command without --api-key
option.
tardis-machine
Run tardis-machine --help
to see all available options (setting custom cache dir etc.)
-
Docker
docker run -v ./host-cache-dir:/.cache -p 8000:8000 -e "TM_API-KEY=YOUR_API_KEY" -d tardisdev/tardis-machine
Command above will pull and run latest version of tardisdev/tardis-machine
image. Local proxy will be available on host via 8000
port (eg http://localhost:8000/replay/...
).
It will also pass YOUR_API_KEY
for accessing the data (otherwise only first day of each month is accessible).
Finally it will use ./host-cache-dir
as persistent volume (bind mount) cache dir (otherwise each docker restart will result in loosing local cache)
In order to debug issues/see detailed logs one can run:
docker run -v ./host-cache-dir:/.cache -p 8000:8000 -e "TM_API-KEY=YOUR_API_KEY" -e="DEBUG=tardis*" -d tardisdev/tardis-machine
Setting DEBUG=tardis*
wil cause to print various useful debug logs to stdout.
If using volumes causes issues (Windows?) it's perfectly fine to run docker without it (with the caveat of potentially poor local cache ratio after container restart)
docker run -p 8000:8000 -e "TM_API-KEY=YOUR_API_KEY" -d tardisdev/tardis-machine
tardisdev/tardis-machine
docker image can be used in K8S setup as well (just set up proper env variables and persisten volumes in analogue way).
FAQ
Order book snapshots
Order book snapshots are available at the beginning of the day ( 00:00 UTC). (TODO: that requires more info)
How to debug it if something went wrong?
This lib uses debug package for verbose logging and debugging purposes that can be enabled via DEBUG
environment variable set to tardis*
.
License
MPL-2.0