Data Models (dccd.models)

Pydantic models for validating exchange API responses.

All raw exchange payloads are validated through Pydantic v2 models before being stored. If a response field has an unexpected type (e.g. a non-numeric price string) the model raises a ValidationError immediately, so corrupt data never silently reaches the Parquet files.

Models at a glance

OHLCBar — one OHLCV candle returned by REST endpoints:

Field	Type	Notes
date	float	Unix timestamp in seconds.
open, high, low, close	float	OHLC prices.
volume	float	Base-asset volume.
quoteVolume	float	Quote-asset volume (price × volume).
weightedAverage	float | None	VWAP. Populated by Kraken only; None for all other exchanges.

Trade — one individual trade from REST history or WebSocket streams:

Field	Type	Notes
tid	int | None	Exchange trade ID. None when the exchange does not provide an integer ID (e.g. Kraken, Bybit).
timestamp	float	Unix timestamp in seconds.
price	float	Trade price.
amount	float	Trade size in base asset.
type	str | None	'buy' or 'sell'. None when the side is not reported.

OrderBookEntry — one price level from a snapshot or diff stream:

Field	Type	Notes
side	str	'bid' or 'ask'.
price	str	Price level stored as a string to preserve full exchange precision.
amount	float	Total quantity available at this level.
count	int | None	Number of open orders. None when not provided (e.g. Binance, Kraken).

API reference

OHLCBar(*, date, open, high, low, close, ...)	OHLCV bar returned by exchange REST APIs.
Trade(*[, tid, type])	Individual trade from WebSocket streams or REST history endpoints.
OrderBookEntry(*, side, price, amount[, count])	Single order book level from REST snapshot or WebSocket streams.