Starstruk Logo MarkStarstruk

v1.174.0

nautechsystems/nautilus_traderv1.174.0May 20, 2023by github-actions[bot]

AI Summary

This release improves backtest stability by handling account blow-up scenarios and reorganizes the namespace for order book data.

Key Highlights

  • Improved handling for backtest account blow-up scenarios (negative balance or margin exceeded).
  • Moved order book data from `model.orderbook.data` into the `model.data.book` namespace.
  • Parquet schemas now shift towards catalog v2.

Breaking Changes

  • Parquet schemas are now shifting towards catalog v2 (legacy support dropped).
  • Moved order book data from `model.orderbook.data` into the `model.data.book` namespace.

New Features

  • Added `AccountMarginExceeded` exception.
  • Refined `AccountBalanceNegative` exception.
  • Various improvements to `Binance` clients error handling and logging.

Full Release Notes

# NautilusTrader 1.174.0 Beta

Released on 19th May 2023 (UTC).

### Breaking Changes
- Parquet schemas are now shifting towards catalog v2 (we recommend you don't upgrade if using legacy catalog)
- Moved order book data from `model.orderbook.data` into the `model.data.book` namespace

### Enhancements
- Improved handling for backtest account blow-up scenarios (balance negative or margin exceeded)
- Added `AccountMarginExceeded` exception and refined `AccountBalanceNegative`
- Various improvements to `Binance` clients error handling and logging
- Improve Binance HTTP error messages

### Fixes
- Fixed handling of emulated order contingencies (not based on status of spawned algorithm orders)
- Fixed sending execution algorithm commands from strategy
- Fixed `OrderEmulator` releasing of already closed orders
- Fixed `MatchingEngine` processing of reduce only for child contingency orders
- Fixed `MatchingEngine` position ID assignment for child contingency orders
- Fixed `Actor` handling of historical data from requests (will now call `on_historical_data` regardless of state), thanks for reporting @miller-moore
- Fixed pyarrow schema dictionary index keys being too narrow (int8 -> int16), thanks for reporting @rterbush