Native Order Book

quantpylib.hft.orderbook exposes the native order-book implementation. It applies quantpylib.standards.models.BookUpdate events and keeps live bid/ask state in fast, native cpp implementation and exposed via nanobind to Python.

Prices and sizes are fixed-point int64 values. PRICE_SCALE and SIZE_SCALE are both 1_000_000_000, so decoded values are stored_int / 1_000_000_000. Level arrays are returned as NumPy int64 arrays with shape (n, 2), where each row is [price, size]. Bid levels are returned best bid first; ask levels are returned best ask first.

Implementations

OrderBook is the default public alias for OrderBookMap. OrderBookMap and OrderBookVec expose the same Python API:

Class	Storage model	Typical use
`OrderBook`	Alias for `OrderBookMap`	Default materialized order-book state.
`OrderBookMap`	Native ordered-map using `std::map`	General-purpose order-book materialization.
`OrderBookVec`	Native vector using `std:vector`	Reverse-vector-backed materialization.

API Reference

`OrderBook`

Default native order book.

OrderBook is the default runtime alias for quantpylib.hft.orderbook.OrderBookMap. It is the order-book type created by quantpylib.hft.materializers.OrderbookMaterializerSink when no existing book is supplied.

`Side`

Order-book side enum used by native order-book methods.

Attributes:

Name	Type	Description
`Bid`	`Side`	Bid side.
`Ask`	`Side`	Ask side.

`OrderBookMap`

Map-backed native order book.

OrderBookMap stores aggregate levels in native ordered maps and exposes the same Python API as quantpylib.hft.orderbook.OrderBookVec. OrderBook is the default runtime alias for this implementation.

Parameters:

Name	Type	Description	Default
`reserve_orders`	`int`	Initial native capacity for order-level storage.	`1024`

`add_order(id, side, price, size)`

Add one order-level entry and update aggregate levels.

`apply(update)`

Apply one native BookUpdate snapshot or delta.

`apply_mbp_level(side, price, size)`

Set or delete one aggregate market-by-price level.

`best_ask()`

Return (price, size) for the best ask, or None.

`best_bid()`

Return (price, size) for the best bid, or None.

`cancel_order(id)`

Cancel one order-level entry and update aggregate levels.

`clear()`

Clear all book state.

`get_asks()`

Return all ask levels as a best-to-worst int64 array.

`get_bids()`

Return all bid levels as a best-to-worst int64 array.

`get_top_asks(depth)`

Return up to depth best ask levels.

`get_top_bids(depth)`

Return up to depth best bid levels.

`has_order(id)`

Return whether an order id is present.

`mid()`

Return the midpoint price, or None when a side is missing.

`modify_order(id, new_size)`

Modify one order-level entry and update aggregate levels.

`order(id)`

Return one order-level entry as a dictionary.

`OrderBookVec`

Vector-backed native order book.

OrderBookVec maintains aggregate L2 levels and optional order-level state in contiguous native storage. It accepts quantpylib.standards.models.BookUpdate through apply(), supports direct MBP level mutation through apply_mbp_level(), and exposes best-to-worst bid/ask levels as NumPy int64 arrays with shape (n, 2).