Fix Issue 1251: covariance calc for insufficient data points

[jensnesten]

Fixes #1251

The code was generating runtime warnings when calculating covariance with insufficient data points:

RuntimeWarning: Degrees of freedom <= 0 for slice
cov_matrix = np.cov(equity_log_returns, market_log_returns)

Added length checks before covariance calculation to handle cases where:

• There are no trades executed
• The data series is too short
• There are NaN values in the data

The fix returns np.nan for both Alpha and Beta metrics when there isn't enough data for a valid covariance calculation.

[kernc]

Applied with slight modification. Thanks for figuring it out!