Python告诉你：从《入海》到《消愁》毛不易的歌里都在唱些什么？

# 导入所需包
import requests
import json
import pandas as pd
import time


def get_bili_comment_one(url):
    """
    功能：定义函数，获取一页的信息
    """
    # 添加headers
    headers = {
        'Host': 'api.bilibili.com',
        'Referer': 'https://www.bilibili.com/video/BV1YZ4y1j7s5',
        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/80.0.3987.149 Safari/537.36'
    }

    # 添加cookies
    cookies = {
        "cookie": "复制您的浏览器cookie信息"
    }

    # 发起请求
    try:
        r = requests.get(url, headers=headers, cookies=cookies, timeout=3)
    except Exception as e:
        print(e)
        r = requests.get(url, headers=headers, cookies=cookies, timeout=3)

    # 解析为字典
    r_json = json.loads(r.text)

    # 提取信息
    replies_data = r_json['data']['replies']

    # 用户名
    user_name = [i['member'].get('uname') for i in replies_data]
    # 性别
    sex = [i['member'].get('sex') for i in replies_data]
    # 签名
    sign = [i['member'].get('sign') for i in replies_data]
    # 用户等级
    current_level = [i['member']['level_info'].get('current_level') for i in replies_data]
    # 评论内容
    content = [i['content'].get('message') for i in replies_data]
    # 用户设备
    device = [i['content'].get('device') for i in replies_data]
    # 评论时间
    content_time = [i.get('ctime') for i in replies_data]
    # 回复数
    reply_count = [i['rcount'] for i in replies_data]

    # 存储数据
    df = pd.DataFrame({
        'user_name': user_name,
        'sex': sex,
        'sign': sign,
        'current_level': current_level,
        'content': content,
        'device': device,
        'content_time': content_time,
        'reply_count': reply_count
    })

    return df


def get_bili_comment_all(oid, num):
    """
    功能：定义函数，获取B站视频指定页评论信息
    """
    # 循环构建URL
    df_all = pd.DataFrame()

    for page_num in range(1, num):
        try:
            # 构建URL
            url = 'https://api.bilibili.com/x/v2/reply?&pn={}&type=1&oid={}&sort=2'.format(page_num, oid)
            # 调用函数
            df = get_bili_comment_one(url)
            # 判断
            if df.shape[0] == 0:
                break
            else:
                # 循环追加
                df_all = df_all.append(df, ignore_index=True)
                # 打印进度
                print('我正在获取第{}页的信息'.format(page_num))
        except:
            break

    # 休眠一秒
    time.sleep(0.5)

    return df_all


# 《入海》bilibili X 毛不易 | 跃入人海，各有风雨灿烂
df = get_bili_comment_all(oid='795637027', num=973)

# 读入数据
df.head()  

df.info() 

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 19099 entries, 0 to 19098
Data columns (total 8 columns):
user_name        19099 non-null object
sex              19099 non-null object
sign             9896 non-null object
current_level    19099 non-null int64
content          19099 non-null object
device           4159 non-null object
content_time     19099 non-null int64
reply_count      19099 non-null int64
dtypes: int64(3), object(5)
memory usage: 1.2+ MB

# 导入包
import numpy as np 
import pandas as pd

# 读入数据
df = pd.read_excel('../data/B站评论数据-入海5.23.xlsx')

# 去重
df = df.drop_duplicates()

# 转换类型
df['content'] = [str(i) for i in df.content] 

# 定义转换时间
def transform_timestamp(time_second):
    timeArray = time.localtime(time_second)
    otherStyleTime = time.strftime('%Y-%m-%d %H:%M:%S', timeArray) 
    return otherStyleTime

# 提取时间
df['content_time'] = df['content_time'].apply(lambda x:transform_timestamp(x)) 

from pyecharts.charts import Bar, Pie, Line, WordCloud, Page
from pyecharts import options as opts 
from pyecharts.globals import SymbolType

import stylecloud
from IPython.display import Image

# 总体评分分布
sex_num = df['sex'].value_counts()
sex_num.drop('保密', inplace=True) 

# 绘制饼图
data_pair =  [list(z) for z in zip(sex_num.index.tolist(), sex_num.values.tolist())]

# 绘制饼图
pie1 = Pie(init_opts=opts.InitOpts(width='1350px', height='750px'))
pie1.add('', data_pair, radius=['35%', '60%'])
pie1.set_global_opts(title_opts=opts.TitleOpts(title='评论用户性别占比'), 
                     legend_opts=opts.LegendOpts(orient='vertical', pos_top='15%', pos_left='2%'))
pie1.set_series_opts(label_opts=opts.LabelOpts(formatter="{b}:{d}%"))
pie1.set_colors(['#EF9050', '#3B7BA9', '#6FB27C'])
pie1.render()

device_num = df.device.value_counts(ascending=True) 

# 柱形图
bar1 = Bar(init_opts=opts.InitOpts(width='1350px', height='750px'))
bar1.add_xaxis(device_num.index.tolist())
bar1.add_yaxis('', device_num.values.tolist(), 
               label_opts=opts.LabelOpts(position='right'))
bar1.set_global_opts(title_opts=opts.TitleOpts(title='评论客户端分布'), 
                     visualmap_opts=opts.VisualMapOpts(max_=3000))
bar1.reversal_axis()
bar1.render() 

# 用户等级
level_num = df.current_level.value_counts()

data_pair2 =  [list(z) for z in zip(['LV' + i for i in level_num.index.astype('str').tolist()] , level_num.values.tolist())]

# 绘制饼图
pie2 = Pie(init_opts=opts.InitOpts(width='1350px', height='750px'))
pie2.add('', data_pair=data_pair2, radius=['35%', '60%'])
pie2.set_global_opts(title_opts=opts.TitleOpts(title='评论用户等级分布'), 
                     legend_opts=opts.LegendOpts(orient='vertical', pos_top='15%', pos_left='2%'))
pie2.set_series_opts(label_opts=opts.LabelOpts(formatter="{b}:{c}({d}%)"))
pie2.set_colors(['#EF9050', '#3B7BA9', '#6FB27C', '#FFAF34'])
pie2.render()

# 时间数据处理
df['time'] = df.content_time.str.split('-').str[1] + '-' + df.content_time.str.split('-').str[2]
df['time'] = df.time.str.split(':').str[0]
time_num = df.time.value_counts().sort_index()

# 产生数据
x1_line1 = time_num.index.values.astype('str').tolist()
y1_line1 = time_num.values.tolist() 

# 绘制面积图
line1 = Line(init_opts=opts.InitOpts(width='1350px', height='750px'))
line1.add_xaxis(x1_line1)
line1.add_yaxis('', y1_line1, areastyle_opts=opts.AreaStyleOpts(opacity=0.3),
                markpoint_opts=opts.MarkPointOpts(data=[
                    opts.MarkPointItem(type_='max', name='最大值')
                ])) 
line1.set_global_opts(title_opts=opts.TitleOpts('各个时段评论人数'), 
                      xaxis_opts=opts.AxisOpts(axislabel_opts=opts.LabelOpts(rotate='30'))
                     ) 
line1.set_series_opts(label_opts=opts.LabelOpts(is_show=False), 
                      axisline_opts=opts.AxisLineOpts()
                     )
line1.render()

import stylecloud
from IPython.display import Image # 用于在jupyter lab中显示本地图片

stylecloud.gen_stylecloud(text=' '.join(word_num),  # txt需要传入str格式 
                          collocations=False,
                          font_path=r'‪C:\Windows\Fonts\msyh.ttc',
                          icon_name='fas fa-graduation-cap',
                          size=768,
                          output_name='B站评论词云图.png')
Image(filename='B站评论词云图.png')