import pandas as pd
from pyecharts.charts import Bar
from pyecharts import options as opts
from pyecharts.globals import ThemeType


df = pd.read_csv('./JD.csv', encoding='gbk')
# print(f"==>> df: {df}")

# 价格去掉无用字符，然后把价格转换成浮点类型数据
df['价格'] = df['价格'].str.replace('￥', '').str.replace(',', '').astype(float)
df['评论数'] = df['评论数'].str.replace('+', '').str.replace('万', '0000').astype(float)



# print(f"==>> df: {df.head(5)}")
describe = df.describe()
print(f"==>> describe: {describe}")

# 价格区间，和标签
bins = [0, 100, 300, 500, 800, 1000, 1300]
labels = ['100以下', '100-300', '300-500', '500-800', '800-1000', '1000以上']

# 把价格区间分为6个区间
df['价格区间'] = pd.cut(df['价格'], bins=bins, labels=labels, include_lowest=True)
# 统计区间内的价格
df_price_count = df['价格区间'].value_counts(sort=False)
print(f"==>> df_price_count: {df_price_count}")

# 画直方图
hist = (
    Bar(init_opts=opts.InitOpts(theme=ThemeType.LIGHT))
    .add_xaxis(df_price_count.index.tolist())
    .add_yaxis('价格区间', df_price_count.values.tolist())
    .set_global_opts(title_opts=opts.TitleOpts(title='价格区间分布直方图'))
)

hist.render('./价格区间分布直方图.html')


# 把评论数去重，排序，然后取前十
df_sorted = df.sort_values(by='评论数', ascending=False).drop_duplicates(subset='评论数').head(10)
# print(f"==>> df_sorted: {df_sorted}")

# 画条形图
bar_reverse = (
    Bar(init_opts=opts.InitOpts(theme=ThemeType.LIGHT, width='1200px', height='800px'))
    .add_xaxis([i for i in reversed(df_sorted['商品'].tolist())])
    # .add_xaxis([i for i in reversed(df_sorted['店铺'].tolist())])
    .add_yaxis('评论数', [i for i in reversed(df_sorted['评论数'].values.tolist())])
    .reversal_axis()
    .set_series_opts(label_opts=opts.LabelOpts(position='right'))
    .set_global_opts(
            yaxis_opts=opts.AxisOpts(
                axislabel_opts=opts.LabelOpts(font_size=10, rotate=60)  # 调整字体大小并倾斜标签
            )
    )
)
bar_reverse.render('./评论数前十条直方图.html')