RollingMax window

Here is the corrected RollingMax indicator for QuantConnect Python. This version implements all abstract methods required by IndicatorBase, uses snake_case naming, and updates the RollingWindow with the correct logic for maximum calculation:

from AlgorithmImports import *

class RollingMax(IndicatorBase[IndicatorDataPoint]):
    def __init__(self, name: str, period: int) -> None:
        super().__init__(name)
        self._period: int = max(1, int(period))
        self._window: RollingWindow[float] = RollingWindow[float](self._period)

    def is_ready(self) -> bool:  # type: ignore[override]
        return self._window.is_ready

    def update(self, input: IndicatorDataPoint) -> bool:  # type: ignore[override]
        return super().update(input)

    def compute_next_value(self, input: IndicatorDataPoint) -> float:  # type: ignore[override]
        self._window.add(float(input.value))
        if not self._window.is_ready:
            return float('nan')
        return float(max(self._window))

You can now instantiate and use this RollingMax in your algorithm without getting the abstract class error!

Sources: Maximum - QuantConnect.com, Rolling Window - QuantConnect.com, Rolling Window - QuantConnect.com, Rolling Window Algorithm, Custom Indicator Algorithm