"""
core/regression_auditor.py
Automated rank shift regression tracking against historic snapshots.
Generates an audit report of movements exceeding the 100-position threshold.
"""

from pathlib import Path
from datetime import datetime, timezone
import pandas as pd


def audit_ranking_regression(
    current_csv_path: str = "data/hiffi_multiplatform_500.csv",
    snapshot_dir: str = "data/snapshots",
    threshold: int = 100
) -> pd.DataFrame:
    cur_path = Path(current_csv_path)
    snap_path = Path(snapshot_dir)
    snap_path.mkdir(parents=True, exist_ok=True)

    df_current = pd.read_csv(cur_path)

    # Find most recent previous snapshot
    snapshots = sorted(snap_path.glob("hiffi_ranking_*.csv"))
    
    if not snapshots:
        # Save initial baseline snapshot if none exists
        today_str = datetime.now(timezone.utc).strftime("%Y%m%d")
        initial_snap = snap_path / f"hiffi_ranking_{today_str}.csv"
        df_current.to_csv(initial_snap, index=False)
        print(f"[RegressionAuditor] Initial snapshot saved to {initial_snap}. No prior history to compare.")
        return pd.DataFrame()

    latest_snapshot = snapshots[-1]
    print(f"[RegressionAuditor] Comparing current dataset against snapshot: {latest_snapshot.name}")
    df_prev = pd.read_csv(latest_snapshot)

    # Compare published ranks
    cur_ranks = df_current.set_index("name")["hiffi_xp_published_rank"]
    prev_ranks = df_prev.set_index("name")["hiffi_xp_published_rank"]

    comparison = pd.DataFrame({
        "previous_published_rank": prev_ranks,
        "current_published_rank": cur_ranks
    }).dropna()

    comparison["rank_shift"] = comparison["previous_published_rank"] - comparison["current_published_rank"]
    flagged = comparison[comparison["rank_shift"].abs() > threshold].sort_values(by="rank_shift", ascending=False)

    print(f"[RegressionAuditor] Total positions analyzed: {len(comparison)} | Movements > {threshold}: {len(flagged)}")
    return flagged


if __name__ == "__main__":
    audit_ranking_regression()