Function make_synthetic_futures_dataset

python/openquant/research.py:42–98

Build a deterministic synthetic multi-asset futures dataset. The first asset is treated as the primary traded instrument (e.g., crude oil), while the other assets provide cross-asset context for allocation/risk.

(
    n_bars: int = 192,
    seed: int = 7,
    asset_names: list[str] | None = None,
)

def make_synthetic_futures_dataset(
    n_bars: int = 192,
    seed: int = 7,
    asset_names: list[str] | None = None,
) -> ResearchDataset:
    """Build a deterministic synthetic multi-asset futures dataset.

    The first asset is treated as the primary traded instrument (e.g., crude oil),
    while the other assets provide cross-asset context for allocation/risk.
    """
    if n_bars < 32:
        raise ValueError("n_bars must be >= 32")
    rng = random.Random(seed)
    asset_names = asset_names or ["CL", "NG", "RB", "GC"]
    n_assets = len(asset_names)
    if n_assets < 2:
        raise ValueError("asset_names must contain at least 2 assets")

    start = datetime(2024, 1, 1, 9, 30, 0)
    timestamps = [(start + timedelta(minutes=i)).strftime("%Y-%m-%d %H:%M:%S") for i in range(n_bars)]

    base = 80.0
    close: list[float] = []
    for i in range(n_bars):
        seasonal = 0.45 * sin(i / 9.0) + 0.25 * sin(i / 17.0)
        drift = 0.006 * i
        noise = rng.uniform(-0.10, 0.10)
        price = base + drift + seasonal + noise
        close.append(max(price, 1.0))

    model_probabilities: list[float] = []
    model_sides: list[float] = []
    for i in range(n_bars):
        edge = 0.53 + 0.08 * sin(i / 13.0) + rng.uniform(-0.025, 0.025)
        p = min(max(edge, 0.05), 0.95)
        model_probabilities.append(p)
        model_sides.append(1.0 if sin(i / 11.0) >= 0.0 else -1.0)

    asset_prices: list[list[float]] = []
    for i in range(n_bars):
        row: list[float] = []
        for j in range(n_assets):
            lag = max(i - (j + 1), 0)
            spread = 0.45 + 0.07 * j
            px = close[lag] * (1.0 + 0.0015 * j) + spread * sin((i + 3 * j) / (9.5 + j))
            px += rng.uniform(-0.08, 0.08)
            row.append(max(px, 1.0))
        asset_prices.append(row)

    return ResearchDataset(
        timestamps=timestamps,
        close=close,
        model_probabilities=model_probabilities,
        model_sides=model_sides,
        asset_prices=asset_prices,
        asset_names=asset_names,
    )