import os
import re
import pandas as pd

DATA_DIR = "data"
INPUT_PATH = os.path.join(DATA_DIR, "master_artists_intelligence_17k.csv")
FINAL_OUTPUT_PATH = os.path.join(DATA_DIR, "master_artists_intelligence_17k_final.csv")


def normalize_name(text: str) -> str:
    if not text:
        return ""
    cleaned = re.sub(r"\(.*?\)|\[.*?\]", "", str(text))
    cleaned = re.sub(r"[^\w\s]", "", cleaned)
    return cleaned.strip().lower()


def run_final_qa():
    print(f"Reading dataset: {INPUT_PATH}")
    df = pd.read_csv(INPUT_PATH)
    initial_count = len(df)
    print(f"Initial total artists: {initial_count:,}")

    # 1. Deduplication by normalized name and artist_id
    df["norm_key"] = df["artist_name"].apply(normalize_name)
    
    # Sort so verified records take priority over unverified during deduplication
    tier_order = {
        "TIER_1_GLOBAL_SUPERSTAR_CSPC": 1,
        "TIER_2A_MAJOR_COMMERCIAL_CORE": 2,
        "TIER_2B_CERTIFIED_CATALOG": 3,
        "TIER_3A_VERIFIED_DIGITAL_CATALOG": 4,
        "TIER_3B_ACTIVE_MONTHLY_STREAMERS": 5,
        "TIER_3_UNAUDITED_DIGITAL_CATALOG": 6,
    }
    df["tier_rank"] = df["catalog_intelligence_tier"].map(tier_order).fillna(99)
    df.sort_values(by=["tier_rank", "audited_sales_units", "spotify_total_streams", "spotify_monthly_listeners"], 
                   ascending=[True, False, False, False], inplace=True)

    # Drop duplicate normalized names, retaining highest-ranked tier
    df.drop_duplicates(subset=["norm_key"], keep="first", inplace=True)
    deduped_count = len(df)
    print(f"Deduplicated count: {deduped_count:,} (Removed {initial_count - deduped_count:,} duplicate variants)")

    # 2. Add Unified Sorting Metric Columns for Downstream UI
    def get_primary_metric(row):
        tier = row["catalog_intelligence_tier"]
        if tier in ["TIER_1_GLOBAL_SUPERSTAR_CSPC", "TIER_2A_MAJOR_COMMERCIAL_CORE", "TIER_2B_CERTIFIED_CATALOG"]:
            return pd.Series(["AUDITED_SALES_UNITS", row["audited_sales_units"]])
        elif tier == "TIER_3A_VERIFIED_DIGITAL_CATALOG":
            return pd.Series(["LIFETIME_STREAMS", row["spotify_total_streams"]])
        elif tier == "TIER_3B_ACTIVE_MONTHLY_STREAMERS":
            return pd.Series(["MONTHLY_LISTENERS", row["spotify_monthly_listeners"]])
        return pd.Series(["UNAUDITED", None])

    df[["primary_metric_type", "primary_metric_value"]] = df.apply(get_primary_metric, axis=1)

    # Clean temporary helper columns
    df.drop(columns=["norm_key", "tier_rank"], inplace=True)

    # Sort final export: Tiered priority first, then primary metric descending
    df.sort_values(by=["catalog_intelligence_tier", "primary_metric_value"], 
                   ascending=[True, False], inplace=True)

    df.to_csv(FINAL_OUTPUT_PATH, index=False)
    print(f"\nFinal hardened intelligence catalog exported to: {FINAL_OUTPUT_PATH}")
    print("\nFinal Tier Breakdown:")
    print(df["catalog_intelligence_tier"].value_counts())


if __name__ == "__main__":
    run_final_qa()