Engine Reference¶

Backtrader-style advanced event-driven API.

Backtrader-compatible facade backed by tradelearn's backtest engine.

Analyzer ¶

Base analyzer class attached through Cerebro.addanalyzer.

on_bar ¶

on_bar(bar: Any) -> None

Deprecated for Post Analyzers. Kept for Streaming Analyzers.

Cerebro ¶

Main orchestrator for backtesting (Backtrader facade).

setcommission ¶

setcommission(commission: float = 0.0, margin: float = 0.0, mult: float = 1.0, comminfo: Any = None, name: str | None = None) -> None

Set commission parameters or a custom CommInfo object.

plot ¶

plot(*args: Any, **kwargs: Any) -> list[Any]

Return market replay charts for the most recent run.

report ¶

report(path: str = 'report.html', benchmark: Any | None = None, sections: Any | None = None) -> Any

Write a Tradelearn report for the most recent run.

OptReturn ¶

Lightweight optimization result returned by Cerebro.run(optreturn=True).

GridSearchResult `dataclass` ¶

One grid-search run result.

Indicator ¶

Base class for Engine custom indicators.

Built-in indicators are intentionally not exposed from tradelearn.engine. Use bt.pta / bt.talib / bt.tdx / bt.tv for built-in vector indicators, or subclass this class for strategy-specific custom indicators.

Observer ¶

Minimal Backtrader observer base class.

Signal ¶

Wraps any single-line indicator as a signal source.

SignalStrategy ¶

Strategy subclass that auto-trades based on registered signals.

Signals are indicator lines whose current value (> 0 or < 0) drives entry/exit decisions. Users can still define next() for custom logic executed after signal processing.

Usage via cerebro.add_signal::

cerebro.add_signal(SIGNAL_LONG, MySignalIndicator, period=20)

Or directly inside __init__::

class MySigStrategy(SignalStrategy):
    def __init__(self):
        sma = bt.talib.SMA(self.data.close, period=20)
        sig = Signal(sma - self.data.close)
        self.signal_add(SIGNAL_LONGSHORT, sig)

Sizer ¶

Backtrader-style Sizer with params support.

CommInfoBase ¶

Base class for commission schemes.

DataFeed ¶

Backtrader-style DataFeed with LineSeries magic.

LineSeries ¶

to_series ¶

to_series() -> pd.Series

Return the full line as a pandas Series for vector indicators.

get ¶

get(ago: int = 0, size: int = 1) -> np.ndarray

Return a Backtrader-style trailing window ending at ago.

line.get(size=n) follows Backtrader's strategy helper semantics and returns a trailing window ending at the addressed cursor. Missing leading values are omitted.

wrap_indicator ¶

wrap_indicator(values: Any, name: str | None = None) -> Any

Wrap vector indicator output back into Engine line objects.

Order `dataclass` ¶

from_request `classmethod` ¶

from_request(ref: int, request: OrderRequest, *, data: Any = None) -> Order

Create a backtest runtime order from a broker-neutral request.

to_fill ¶

to_fill(*, qty: float, price: float, commission: float, ts: Timestamp | None = None, broker_oid: str | None = None) -> Fill

Emit a broker-neutral fill event for this backtest order.

Params ¶

Backtrader-style parameter storage object.

Position `dataclass` ¶

call ¶

__call__(*args: Any, **kwargs: Any) -> Position

Return self so facade position objects can be used as position().

Strategy ¶

Backtrader-compatible Strategy that maintains data context.

datetime `property` ¶

datetime

Shortcut for self.data.datetime to match backtrader behavior.

history_panel ¶

history_panel(lookback: int | None = None)

Return recent multi-data OHLCV bars as a timestamp/symbol panel.

grid_search ¶

grid_search(strategy: type[Strategy], data: DataFrame, param_grid: dict[str, list[Any]], *, mlflow: dict[str, Any] | None = None) -> list[GridSearchResult]

Run a simple strategy parameter grid with nested MLflow runs.

num2date ¶

num2date(num)

Convert numeric timestamp back to datetime object.

date2num ¶

date2num(dt)

Convert datetime object to numeric timestamp.