purgedcv implements purging and embargoing of overlapping labels, walk-forward validation, purged and group-purged k-fold, Combinatorial Purged Cross-Validation (CPCV) with backtest-path reconstruction, and the Probabilistic Sharpe Ratio, Deflated Sharpe Ratio, Minimum Track Record Length, and Minimum Backtest Length. Every splitter follows the scikit-learn splitter protocol.