B站视频弹幕的爬取和绘制词云图

# 只需 shift+回车 运行本单元格，就可以让jupyter notebook宽屏显示
from IPython.core.display import display, HTML
display(HTML('<style>.container { width:100% !important; }</style>'))

1.爬虫

① url：网页链接

② 模拟浏览器的发送请求即响应

③ 解析网页内容

def BZDM(url,headers):
    # 获取网页信息,发送请求
    resp = requests.get(url,headers=headers)
    # 更改为"utf-8"编码，防止出现乱码
    resp.encoding = "utf-8"
    res = resp.text
    """
    正则表达式都需要先通过分析网页源代码，然后写表达式才能提取我们想要的信息
    这里写了3种表达式，都可以实现弹幕的提取
    """
    # 正则表达式提取每个人的弹幕，包括标点符号
    # ex = r"<d.*?>(.*?)</d>"
    ex = '[\u3002\uff1b\uff0c\uff1a\u201c\u201d\uff08\uff09\u3001\uff1f\u300a\u300b\u4e00-\u9fa5]+'
    #ex = '[\u4e00-\u9fa5]+' # 只提取汉字,不含标点符号
    # 提取
    ZM = re.findall(ex,res,re.S)
    return ZM

import requests
import re

# 头文件信息：模拟浏览器登录、访问
headers = {
        'user-agent':'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:79.0) Gecko/20100101 Firefox/79.0',
        "cookie":"_uuid=149B7425-9EC0-F418-E3D4-F05F11F52E7817158infoc; buvid3=7BCBA652-D987-4D86-903B-52710A2084A570379infoc; sid=9euuhyjd; DedeUserID=148925615; DedeUserID__ckMd5=805b9a52a946d4dc; SESSDATA=435a5c44%2C1611037459%2Ca06e4*71; bili_jct=acdb9ef8436f6e9edffb8acdfbefa73d; CURRENT_FNVAL=16; CURRENT_QUALITY=80; rpdid=|(um~u)lRm)k0J'ulmlR)Y|)k; PVID=1; bp_video_offset_148925615=420590814271723170; bp_t_offset_148925615=418367829385970886; bfe_id=393becc67cde8e85697ff111d724b3c8"
     }

url = "https://api.bilibili.com/x/v2/dm/history?type=1&oid=217870191&date=2020-08-03"
ZM = BZDM(url,headers)

④ 数据的存储

# 数据的存储
import csv # 储存csv库
b = []
for i in ZM:
    with open("B站弹幕.csv","a",newline="",encoding="utf-8-sig") as f:
        writer = csv.writer(f)
        writer.writerow([i])

2.绘制词云图

突出文章的关键词，词出现频率越大，在图中显示的就越大。

jieba中文分词地址

import jieba # 中文分词课，用于将句子拆分为词语
import wordcloud
import pyecharts.charts as pe
import matplotlib.pyplot as plt
import matplotlib.font_manager as fm
plt.rcParams['font.sans-serif']=['SimHei']  # 用来正常显示中文标签  
plt.rcParams['axes.unicode_minus']=False  # 用来正常显示负号

f = open("B站弹幕.csv",encoding="utf-8")
txt = f.read()

def make_wordcloud(text):
    # 把不想展示的词语删掉
    text1 = text.replace("哈哈哈", "")
    # 读取词云形状的图片
    # bg = plt.imread("zz.jpg")
    # 生成
    wc = wordcloud.WordCloud(# FFFAE3
        background_color="white",  # 设置背景为白色，默认为黑色
        width=1000,  # 设置图片的宽度
        height=600,  # 设置图片的高度
        #mask=bg,
        margin=1,  # 设置图片的边缘
        max_font_size=150,  # 显示的最大的字体大小
        random_state=10,  # 为每个单词返回一个PIL颜色
        font_path='simkai.ttf'  # 中文处理，用系统自带的字体
    ).generate_from_text(text1)
    # 为图片设置字体
    my_font = fm.FontProperties(fname='simkai.ttf')
    # 图片背景
    # bg_color = wordcloud.ImageColorGenerator(bg)
    # 设置图片大小和质量
    plt.figure(dpi=1000)
    # 显示
    plt.imshow(wc.recolor())
    # 去掉坐标轴
    plt.axis("off")
    
    # 保存云图
    #wc.to_file("CC.png")

make_wordcloud(txt)

def data_Pyechart(ZM):
    jieba.enable_paddle()# 启动paddle模式。
    Words = []
    for i in ZM:
        # seg_list = jieba.cut(i, cut_all=True) # 全模式
        # seg_list = jieba.cut(i, cut_all=False) # 精确模式
        seg_list = jieba.cut_for_search(i) # 搜索引擎
        for j in seg_list:
            Words.append(j)
    Words_set = set(Words)
    data = []
    for i in Words_set:
        n = 0
        for j in Words:
            if i == j:
                n += 1
        if n > 2:
            data.append((i,'%d' %n))
    return data

def Drow_Pyecharts_WordCloed(ZM,series_name,title):
    import pyecharts.options as opts
    from pyecharts.charts import WordCloud
    data = data_Pyechart(ZM)
    c = (
    WordCloud(init_opts=opts.InitOpts(width="1000px", height="600px"))
    .add(series_name=series_name, data_pair=data, word_size_range=[20, 50])
    .set_global_opts(
        title_opts=opts.TitleOpts(
            title=title, title_textstyle_opts=opts.TextStyleOpts(font_size=23)
        ),
        tooltip_opts=opts.TooltipOpts(is_show=True),
    )
    )
    return c.render_notebook()

Drow_Pyecharts_WordCloed(ZM,"AA","BB")

Paddle enabled successfully......