Add RealizedVolatility indicator (Parkinson high-low estimator)

[rairishi764]

What this PR adds

• Indicators/RealizedVolatility.cs — a new indicator implementing the Parkinson (1980) high-low range estimator of realized volatility, as BarIndicator + IIndicatorWarmUpPeriodProvider, following the same shape as RogersSatchellVolatility.
• Tests/Indicators/RealizedVolatilityTests.cs — extends CommonIndicatorTests<IBaseDataBar>, comparing against an independently-computed reference CSV (Tests/TestData/spy_with_rv.csv), following the same convention as RogersSatchellVolatilityTests/spy_with_rsv.csv.
• Algorithm/QCAlgorithm.Indicators.cs — the RV(...) helper method, mirroring RSV(...) for RogersSatchellVolatility.
• Tests/QuantConnect.Tests.csproj — wires up the new test fixture CSV.

What this PR does not change

• No changes to existing indicators or the Python API surface.
• No new external package references.

Public API surface

public class RealizedVolatility : BarIndicator, IIndicatorWarmUpPeriodProvider
{
    public int WarmUpPeriod { get; }
    public RealizedVolatility(string name, int period);
    public RealizedVolatility(int period);
}

QCAlgorithm.RV(symbol, period, resolution, selector) creates and registers it, mirroring RSV(...).

Why Parkinson, and why now

LEAN has several volatility indicators (AverageTrueRange, RogersSatchellVolatility, ImpliedVolatility, Keltner/Bollinger) but no realized volatility indicator — the standard backward-looking measure of historical price variation. Parkinson is the simplest high-low estimator (needs only High/Low, unlike Garman-Klass/Yang-Zhang which also need Open/Close), making it a clean first addition and a template for those follow-ups.

Testing

• dotnet build clean on Common, Algorithm, and the full Tests project (net10.0).
• RealizedVolatilityTests: 16/17 passing locally. The 1 failure (WarmUpIndicatorProducesConsistentResults) is a pre-existing Python runtime DLL gap in my local dev environment, not a regression — I confirmed the identical failure occurs when running the same test against the already-merged RogersSatchellVolatility indicator on this machine.
• The reference CSV (spy_with_rv.csv) was generated independently in Python from the existing spy_with_rsv.csv OHLC data, not derived from my C# implementation.

Closes #9863