Kraken Report

tribulnation-kraken, venue name kraken. Keyed: Balance needs Funds permissions - Query, Ledgers needs Data - Query ledger entries.

Snapshot

snapshot() is one Balance call: every asset the account holds, keyed by Kraken's internal asset id (XXBT, XETH, ZUSD, USDC), in a single unnamed subaccount. Funds in Earn appear under suffixed ids of their own — XXBT.F for Kraken Rewards, .B for bonded products — and are reported as the venue names them. Spot has no positions, so none are reported.

History

Kraken has no unified history endpoint, but its ledger is close: every balance-affecting event is one Ledgers row per asset leg, which is the SDK's per-observation model. history() pages the ledger newest first, 50 rows at a time, and yields one HistoryRecord per row, its provenance id the ledger id.

Ledger `type`	Observation
`trade`	`TradeLeg` (`event_type = spot_trade`, `trade_id` the row's `refid`, `label` its `subtype`)
`deposit`	`FiatDeposit` for a fiat asset, `CryptoDeposit` otherwise
`withdrawal`	`FiatWithdrawal` for a fiat asset, `CryptoWithdrawal` otherwise
`staking`	`Yield`
`transfer`	`Transfer`
`reward`	`Bonus`
anything else	`UnknownObservation`

A row carrying a non-zero fee adds a FeeLeg in the same asset, pointing at the row.

Two things are transcribed rather than reconstructed:

Trades are legs, not trades. A spot trade is two trade rows — the base leg and the quote leg — sharing a refid, and each is its own record. Joining them back into one SpotTrade would need TradesHistory for the price and side; the ledger row is what this surface reads.
Fiat is a fixed list. Kraken's asset catalogue classes fiat and crypto alike as currency, so fiat deposits and withdrawals are told apart by the asset id (ZUSD, ZEUR, ZGBP, ZCAD, ZJPY, ZAUD, ZKRW, CHF, MXN, BRL).

Example

from datetime import datetime, timedelta, timezone

from dotenv import load_dotenv
from tribulnation.sdk import ReportSDK, accounts

load_dotenv()

report = ReportSDK({'kraken': accounts.Kraken()}).venue('kraken')
end = datetime.now(timezone.utc)
async with report:
  print((await report.snapshot()).snapshot.balances)
  async for record in report.history(end - timedelta(days=30), end):
    for observation in record.observations:
      print(observation)