"""Project 02 — Python Data Explorer starter script."""

from pathlib import Path

import matplotlib.pyplot as plt
import pandas as pd


DATA_FILE = Path(__file__).parent.parent / "data" / "customer-data.csv"
OUTPUT_DIR = Path(__file__).parent.parent / "outputs"


def main():
    try:
        df = pd.read_csv(DATA_FILE)
        print("Dataset loaded successfully.")
    except FileNotFoundError:
        print(f"Dataset not found: {DATA_FILE}")
        return

    print("\nFirst five rows:")
    print(df.head())
    print("\nShape:", df.shape)
    print("\nColumns:", list(df.columns))
    print("\nData types:")
    df.info()

    print("\nMissing values:")
    print(df.isnull().sum())
    print("\nDuplicate rows:", df.duplicated().sum())

    df = df.drop_duplicates().copy()
    df["Monthly Spending"] = df["Monthly Spending"].fillna(
        df["Monthly Spending"].median()
    )

    print("\nNumerical summary:")
    print(df.describe())
    print("\nAverage monthly spending:", round(df["Monthly Spending"].mean(), 2))
    print("Average number of purchases:", round(df["Number of Purchases"].mean(), 2))
    print("\nMembership counts:")
    print(df["Membership Type"].value_counts())

    high_spending_customers = df[df["Monthly Spending"] > 100]
    print("\nHigh-spending customers:")
    print(high_spending_customers.head())

    membership_summary = (
        df.groupby("Membership Type")["Monthly Spending"]
        .mean()
        .sort_values(ascending=False)
    )
    print("\nAverage spending by membership:")
    print(membership_summary)

    OUTPUT_DIR.mkdir(exist_ok=True)
    df.to_csv(OUTPUT_DIR / "cleaned-customer-data.csv", index=False)
    (OUTPUT_DIR / "data-summary.txt").write_text(
        "Python Data Explorer summary\n\n" + df.describe(include="all").to_string(),
        encoding="utf-8",
    )

    membership_summary.plot(kind="bar", color="#f6b518")
    plt.title("Average Spending by Membership Type")
    plt.xlabel("Membership Type")
    plt.ylabel("Average Monthly Spending")
    plt.tight_layout()
    plt.savefig(OUTPUT_DIR / "membership-spending.png", dpi=160)
    plt.show()


if __name__ == "__main__":
    main()