import os

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

csv_file = "out_all.csv"
stability_file = "out_stability.csv"
degradation_file = "out_degradation.csv"
paper_names = {
    2: "Glossy Pro Platinum",
    3: "Photo Fine-Grained Luster",
    4: "Matte Photo",
    5: "Photo Premium Matte",
}
data_dir = r"\\gabor\Project\内科\舌診\TIAS\開発\カラーチャート\TCC2025-01\退色調査"


def make_all_data_csv() -> None:
    """Reads all CSV files."""  # noqa: D401

    print("Combining all data into a single CSV file...")
    files = [f for f in os.listdir(data_dir) if f.endswith(".csv")]
    all_data = pd.DataFrame()
    for i, file in enumerate(files):
        file_path = os.path.join(data_dir, file)
        paper = int(file[5:6])
        datestr = "2025/" + file[7:9] + "/" + file[9:11]
        iter = 0 if file[11:12] == "a" else 1
        # print(f"Reading {file_path} (Paper: {paper}, Date: {datestr}, Iteration: {iter})")
        df = pd.read_csv(file_path, encoding="utf-8")
        df["tcc"] = df["no"].astype("int")
        df["paper"] = paper
        df["date"] = datestr
        df["iter"] = iter
        # df.drop(columns=["no", "X", "Y", "Z"], inplace=True, errors="ignore")
        # df.drop(columns=[f"{c:.1f}" for c in range(400, 710, 10)], inplace=True, errors="ignore")
        all_data = pd.concat([all_data, df], ignore_index=True)
    # Save the combined DataFrame to a CSV file
    all_data[["date", "paper", "iter", "tcc", "L*", "a*", "b*"]].to_csv(
        csv_file, index=False, encoding="utf-8"
    )


def calc_repeat_stability():
    """Calculate repeat stability of color measurements."""
    if not os.path.exists(csv_file):
        make_all_data_csv()

    print("Calculating repeat stability...")
    df = pd.read_csv(csv_file, encoding="utf-8", parse_dates=["date"])

    df0 = df[df["iter"] == 0].drop(columns=["iter"]).copy()
    df0.columns = ["date", "paper", "tcc", "L0", "a0", "b0"]
    df1 = df[df["iter"] == 1].drop(columns=["iter"]).copy()
    df1.columns = ["date", "paper", "tcc", "L1", "a1", "b1"]
    dfm = pd.merge(df0, df1, on=["date", "paper", "tcc"])
    dfm["dL"] = dfm["L1"] - dfm["L0"]
    dfm["da"] = dfm["a1"] - dfm["a0"]
    dfm["db"] = dfm["b1"] - dfm["b0"]
    dfm["dE"] = np.sqrt(
        dfm["dL"] ** 2 + dfm["da"] ** 2 + dfm["db"] ** 2
    )  # Euclidean distance in Lab color space
    dfm["meanL"] = (dfm["L1"] + dfm["L0"]) / 2
    dfm["meana"] = (dfm["a1"] + dfm["a0"]) / 2
    dfm["meanb"] = (dfm["b1"] + dfm["b0"]) / 2

    dfm.to_csv(stability_file, index=False, encoding="utf-8")

    # print("Top 10 largest deltaE values:")
    # print(dfm.sort_values("dE", ascending=False).head(10))

    # deltaE values Histogram
    # plt.figure()
    # dfm["dE"].hist(bins=30)
    # plt.title("deltaE Histogram")
    # plt.xlabel("deltaE")
    # plt.ylabel("Frequency")
    # plt.show()


def color_degradation():
    """Analyze color degradation over time."""
    if not os.path.exists(stability_file):
        calc_repeat_stability()

    print("Calculating color degradation...")
    dfm = pd.read_csv(stability_file, encoding="utf-8", parse_dates=["date"])
    dfm = dfm[["date", "paper", "tcc", "meanL", "meana", "meanb"]]

    dfm["changeL"] = 0.0
    dfm["changea"] = 0.0
    dfm["changeb"] = 0.0
    dfm["changeDE"] = 0.0
    dfm["days"] = 0
    paper_list = dfm["paper"].drop_duplicates().sort_values()
    tcc_list = dfm["tcc"].drop_duplicates().sort_values()
    dfout = pd.DataFrame()
    for paper in paper_list:
        paper_data = dfm[dfm["paper"] == paper]
        begin_date = paper_data["date"].min()

        for tcc in tcc_list:
            tcc_data = paper_data[paper_data["tcc"] == tcc]
            begin_lab = (
                tcc_data[tcc_data["date"] == begin_date][["meanL", "meana", "meanb"]]
                .iloc[0]
                .to_list()
            )
            tcc_data.loc[:, "days"] = (tcc_data["date"] - begin_date).dt.days
            tcc_data.loc[:, "changeL"] = tcc_data["meanL"] - begin_lab[0]
            tcc_data.loc[:, "changea"] = tcc_data["meana"] - begin_lab[1]
            tcc_data.loc[:, "changeb"] = tcc_data["meanb"] - begin_lab[2]
            tcc_data.loc[:, "changeDE"] = np.sqrt(
                tcc_data["changeL"] ** 2 + tcc_data["changea"] ** 2 + tcc_data["changeb"] ** 2
            )
            dfout = pd.concat([dfout, tcc_data], ignore_index=True)
            # print(tcc_data.sort_values("date"))
            # print("begin_lab", begin_lab)
        #     break
        # break

    dfout = dfout.sort_values(by=["paper", "tcc", "date"]).reset_index(drop=True)
    dfout.to_csv(degradation_file, index=False, encoding="utf-8")


def visualize():
    """Visualize color degradation over time."""

    if not os.path.exists(degradation_file):
        color_degradation()

    print("Visualizing color degradation...")
    df = pd.read_csv(degradation_file, encoding="utf-8", parse_dates=["date"])
    paper_list = df["paper"].drop_duplicates().sort_values()

    fig = plt.figure()
    fig.set_size_inches(18, 12)
    fig.suptitle("Color Degradation Over Time (each TCC)", fontsize=16)
    fig.subplots_adjust(wspace=0.3, hspace=0.4)
    fig.subplots_adjust(top=0.9, bottom=0.05, left=0.05, right=0.95)
    for i in range(1, 25):
        ax = fig.add_subplot(4, 6, i)
        ax.set_title(f"TCC {i}")
        ax.set_xlabel("Days")
        ax.set_ylabel("Change in deltaE")
        for paper in paper_list:
            paper_data = df[df["paper"] == paper]
            if not paper_data.empty:
                ax.plot(
                    paper_data[paper_data["tcc"] == i]["days"],
                    paper_data[paper_data["tcc"] == i]["changeDE"],
                    marker="o",
                    markersize=4,
                    label=f"Paper {paper}",
                )
                ax.set_ylim(0, 7)  # Set y-axis limit for better visibility
        ax.legend()
    fig.savefig("out_degradation_eachTCC.png", dpi=300, bbox_inches="tight")

    paper_mean = df.groupby(["paper", "date"]).mean().reset_index()
    fig2 = plt.figure()
    fig2.set_size_inches(9, 6)
    fig2.suptitle("Color Degradation Over Time (mean TCC)", fontsize=16)
    for paper in paper_list:
        paper_data = paper_mean[paper_mean["paper"] == paper]
        if not paper_data.empty:
            plt.plot(
                paper_data["days"],
                paper_data["changeDE"],
                marker="o",
                markersize=6,
                label=f"Paper {paper} {paper_names[paper]}",
            )
    plt.legend(loc="upper left", borderaxespad=0)
    plt.xlabel("Days")
    plt.ylabel("Change in deltaE")
    fig2.savefig("out_degradation_meanTCC.png", dpi=300, bbox_inches="tight")

    # plt.show()


if __name__ == "__main__":
    visualize()
    print("All done.")