from pathlib import Path

import matplotlib.pyplot as plt
import matplotlib.ticker as mticker
import polars as pl
import seaborn as sns

sns.set_theme(context="paper", style="ticks", palette="deep", color_codes=True)
plt.rcParams["figure.autolayout"] = True
plt.rcParams["figure.dpi"] = 300

match_results = pl.read_csv(
    Path("./../data/match_results.csv"),
).cast(
    {
        "DaysInPost": pl.Int64,
        "Goals": pl.Int64,
        "Date": pl.Date,
    }
)

match_results.head()

head_coach = pl.read_csv(Path("./../data/head_coach.csv")).cast(
    {"Appointed": pl.Date, "EndDate": pl.Date}
)
head_coach.head()

first_season = match_results.get_column("Date").dt.year().min() + 1
last_season = match_results.get_column("Date").dt.year().max()
leagues = ", ".join(match_results.get_column("League").unique())
n_match = round(len(match_results) / 2)
n_team = match_results.get_column("Team").n_unique()
n_team_no_coach = (
    match_results.group_by("Team")
    .agg(pl.col("HeadCoach").is_null().all().alias("all_null"))
    .filter(pl.col("all_null"))
    .height
)
perc_match_no_coach = match_results.get_column("HeadCoach").is_null().sum() / len(
    match_results
)
n_coach = head_coach.get_column("HeadCoach").n_unique()
n_appointments = len(head_coach)

match_summary = (
    match_results.group_by(["League", "Country"])
    .agg(
        [
            pl.col("Team").n_unique().alias("Number of Teams"),
            pl.col("Date").len().alias("Number of Matches"),
            (
                (pl.col("HeadCoach").is_null().sum() / pl.col("HeadCoach").len() * 100)
                .round(2)
                .cast(pl.String)
                + "%"
            ).alias("Percentage of Matches without Head Coach"),
        ]
    )
    .sort("Number of Matches", descending=True)
)

match_summary

# Plot distribution of match
plt.figure()
sns.histplot(
    match_results.get_column("Date").dt.month(),
    stat="density",
    discrete=True,
    alpha=1,
)
plt.title("Monthly Distribution of Matches")
plt.xlabel("Month")
plt.ylabel("Density")
plt.gca().yaxis.set_major_formatter(mticker.PercentFormatter(xmax=1))
plt.gca().set_xticks(range(1, 13))
plt.gca().set_xticklabels(
    ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"]
)
plt.show()

mean_home_goals = match_results.filter(pl.col("isHome")).get_column("Goals").mean()
mean_away_goals = match_results.filter(~pl.col("isHome")).get_column("Goals").mean()
diff_goal_perc = (mean_home_goals - mean_away_goals) / mean_away_goals

match_result_home_away = (
    match_results.group_by(["isHome", "Result"])
    .len()
    .with_columns(proportion=pl.col("len") / pl.col("len").sum().over("isHome"))
    .select(
        pl.col("isHome")
        .replace_strict({True: "Home", False: "Away"})
        .alias("Team venue"),
        pl.col("Result").str.to_titlecase(),
        pl.format("{} %", (pl.col("proportion") * 100).round(1)).alias(
            "Proportion (%)"
        ),
    )
    .sort("Team venue", "Result", descending=True)
)

match_result_home_away

home_stats = match_result_home_away.filter(pl.col("Team venue") == "Home").get_column(
    "Proportion (%)"
)
away_stats = match_result_home_away.filter(pl.col("Team venue") == "Away").get_column(
    "Proportion (%)"
)

perc_home_win = home_stats.item(0)
perc_home_loss = home_stats.item(1)
perc_away_win = away_stats.item(0)
perc_away_loss = away_stats.item(1)
perc_draw = home_stats.item(2)