# 설치가 필요하다면 아래 문장으로 라이브러리 설치
# !pip pandas numpy matplotlib wordcloud bar_chart_race

import pandas as pd
import numpy as np
import matplotlib as mpl
import matplotlib.pyplot as plt
import seaborn as sns
import wordcloud
import bar_chart_race as bcr
import warnings
warnings.filterwarnings("ignore")

import matplotlib.font_manager as fm

font_list = fm.findSystemFonts(fontpaths = None, fontext = 'ttf')

font_list
# print([i for i in font_list if 'kr' in i])

font_location = '/Users/subinium/Library/Fonts/NotoSansCJKkr-Light.otf' 
font_name = fm.FontProperties(fname=font_location).get_name()
# print(f'시각화에 사용한 폰트는 {font_name} 입니다.')
mpl.rc('font', family=font_name)

# 시각화를 위한 설정
pd.options.display.max_columns = 999
mpl.rcParams['figure.dpi'] = 300
mpl.rcParams['axes.spines.top'] = False
mpl.rcParams['axes.spines.right'] = False

month_rank = data.pivot(columns='search_rank_nm', index='base_ym', values='srchwrd_nm').iloc[:,0:20]

# 연속 구간 추출 함수
def continuous_seq(kwd):
    arr = np.array(np.where(month_rank.values==kwd))
    _, length = arr.shape
    ret, st = [], 0
    for i in range(1, length):
        if arr[0, i-1]+1 != arr[0, i]: 
            ret.append(arr[:, st:i])
            st = i            
    ret.append(arr[:, st:])
    return ret 

# 키워드 별 rank plot
def highlight(kwd, color, ax):
    lsts = continuous_seq(kwd)
    for lst in lsts:
        ax.plot(lst[0,:], lst[1,:], 
                marker='o', ms=15, lw=2.5,
                color=color, zorder=0)
        
        ax.scatter(lst[0,:], lst[1,:], 
                marker='o', s=110,
                color='white', zorder=10)
    return ax

# 전체 rank plot
def ranking(kwds):
    fig, axes = plt.subplots(2, 1, figsize=(30, 23))
    
    ax = axes[0]
    total_kwds = pd.Series(month_rank.values.flatten()).unique()

    for kwd in total_kwds:
        highlight(kwd, '#E0E0E0', ax)

        
    color_tot = sns.color_palette("twilight_shifted", 14)[::2]
    kwd_color = {k:c for k, c in zip(kwds, color_tot)}    
    for kwd in kwds:
        highlight(kwd, kwd_color[kwd], ax)

    ax.invert_yaxis()
    ax.set_xticks(range(23))
    ax.set_xticklabels(pd.to_datetime(month_rank.index, format='%Y%m').strftime('%m\n%Y'), fontweight='bold', fontsize=15)
    ax.set_xlim(-0.2, 22.2)
    ax.set_yticks(range(20))
    ax.set_yticklabels([f'Rank {i}' for i in range(1, 21)], fontweight='bold', fontsize=15, color='gray')
    ax.spines['left'].set_visible(False)
    ax.spines['bottom'].set_visible(False)
    title_tmp = ['\"'+k+'\"' for k in kwds]
    ax.set_title(f'월별 {",".join(title_tmp)} 키워드 순위', weight='bold', fontsize=25, fontweight='bold')
    
    ax = axes[1]
    stack = pd.Series(np.ones(23))
    for kwd in kwds:
        b1 = ax.bar(range(23), data_rank_percentage[kwd], bottom=stack, color=kwd_color[kwd], width=0.6, label=kwd)
        ax.bar_label(b1, label_type='center', color='white', fontsize=13, fmt='%.1f%%')
        stack += data_rank_percentage[kwd].values
    ax.set_xticks(range(23))
    ax.set_xticklabels(pd.to_datetime(month_rank.index, format='%Y%m').strftime('%m\n%Y'), fontweight='bold', fontsize=15)
    ax.set_xlim(-0.4, 22.4)
    ax.set_yticks([])
    ax.legend(fontsize=15)
    ax.spines['left'].set_visible(False)
    ax.set_title(f'월별 {",".join(title_tmp)} 키워드 비율', weight='bold', fontsize=25, fontweight='bold')
    
    plt.show()