import os
import tempfile
import shutil

import gradio as gr
import matplotlib.cm as cm
import pandas as pd
import japanize_matplotlib
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

def all_likelihood_plot(csv_file_name, tmpdir):
   df = pd.read_csv(csv_file_name, header=[1, 2])
   df = df.drop(df.columns[[0]], axis=1)
   columns = df.columns.droplevel(1)

   # 重複を削除
   columns = columns.drop_duplicates()
   likelihood = [df[x]["likelihood"] for x in columns]
   a = pd.DataFrame(likelihood, index=columns).T

   #平均値を求める
   point_average = a.mean()
   parts = ["指節1", "指節2", "指節3", "指節4", "指節5", "指節6", "指節7", "指節8", "指節9",
            "指節10", "指節11", "指節12", "指節13", "指節14", "触角(左)", "触角(右)", "頭部", "腹尾節"]

   # カラーマップの設定
   a.columns = parts
   cmap = plt.get_cmap('rainbow')
   # バイオリン図のプロット
   sns.set(style="whitegrid",font="IPAexGothic")
   fig, ax = plt.subplots()

   # データをバイオリンプロットで描画
   sns.violinplot(data=a, palette=[cmap(i)
                  for i in np.linspace(0, 1, len(columns))], ax=ax,inner=None)

   # 横軸のラベルを重ならないように
   plt.xticks(rotation=65)

   ax.set_title('付属肢別の尤度')
   ax.set_xlabel('付属肢')
   ax.set_ylabel('尤度')

   #それぞれの要素の平均値をプロット
   plt.scatter(x=parts, y=point_average, color='black', marker='x')

   # 最大値を1に
   plt.ylim(0, 1)

   # fig.set_figwidth(10)

   #ラベルがはみ出ないように
   plt.tight_layout()

   # グラフを表示
   plt.savefig(f"likelihood.png", dpi=300)


def main(csv_file):
   with tempfile.TemporaryDirectory(dir=".") as tmpdir:
      all_likelihood_plot(csv_file, tmpdir)
      return f"likelihood.png"


iface = gr.Interface(fn=main, inputs="file",
                     outputs="image", title="尤度のグラフを作成します。")
iface.launch()