Python爬虫 | 如何获取网易云音乐评论

"""
不按照大佬写的加密方法，只能获取第一页的评论/(ㄒoㄒ)/~~
"""

import requests
from bs4 import BeautifulSoup
import json
import time

def get_song_html(url):
        """获取网页HTML"""
        headers={
                "User-Agent":"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/73.0.3683.86 Safari/537.36",
                "Host":"music.163.com",
                "Upgrade-Insecure-Requests":'1'
        }
        data={
                "params": "wFDYU3SUWyoCUekCRy6S6oPwnDHv/2Cvd8zMLZ1TCZexhtvcOGdiZdCw+UC7Y5QKC+7KdMMOZqc2eHjTDFfqVEPwuajKbFwywKKBuxe2gfkYBTNiC02rbZM5OxMM22qhVrZRPZMzAxWZz3t213Ts8A==",
                "encSecKey": "b6c67b2c848ef79a5cc0bc0261b6b88b75209276f9f1050091d64731398809b2e0d03081618d9c1a3d442ae1367e7e1a1f54224a6e94fed8eddc3bb337017d0b9f3bb8a274fbf58e8142020b7cbd909f9addf68c674f0232811fa18bf7a1dd90030a5f607ff2c488f20e2aab37dbab1bedff5cfa6684f6e49b69bfc727e943c1"
        }
        song_id=url.split("=")[1]
        """已知的网易云音乐网页链接"""
        #歌曲
        url_so="http://music.163.com/weapi/v1/resource/comments/R_SO_4_{}?csrf_token=".format(song_id)
        #专辑
        url_al="https://music.163.com/weapi/v1/resource/comments/R_AL_3_{}?csrf_token=".format(song_id)
        #电台
        url_dj="https://music.163.com/weapi/v1/resource/comments/A_DJ_1_{}?csrf_token=".format(song_id)
        urls=[url_so,url_al,url_dj]#常用三个url
        answer=input("获取热门评论：1\n获取全部评论：2\n请输入： ")
        if answer=='1':
                for url in urls:
                        try:
                                html=requests.post(url,headers=headers,data=data)
                                html.raise_for_status()
                        except:
                                return "爬取失败！"
                        else:
                                get_song_hot_comments(html)
        elif answer=='2':
                for url in urls:
                        try:
                                html=requests.post(url,headers=headers,data=data)
                                html.raise_for_status()
                        except:
                                return "爬取失败！"
                        else:
                                get_song_comments(html)
        else:
                print("输入错误！")

def get_song_hot_comments(html):
        comments = json.loads(html.text)
        hot_comments = comments['hotComments']
        if hot_comments:
                try:
                        with open('网易云音乐热门评论.txt','w',encoding='utf-8') as f:
                                for C_list in hot_comments:
                                        f.write("User: "+C_list['user']['nickname']+'\n')
                                        f.write("Comment: \n"+C_list['content']+'\n')
                                        f.write("\n"+"-"*30+"\n\n")
                except:
                        print("保存热门评论失败！")
                else:
                        print("保存热门评论成功！")

def get_song_comments(html):
        comments = json.loads(html.text)
        comments = comments['comments']
        if comments:
                try:
                        with open('网易云音乐评论.txt','w',encoding='utf-8') as f:
                                for C_list in comments:
                                        f.write("User: "+C_list['user']['nickname']+'\n')
                                        f.write("Comment: \n"+C_list['content']+'\n')
                                        f.write("\n"+"-"*30+"\n\n")
                except:
                        print("保存全部评论失败！")
                else:
                        print("保存全部评论成功！")

def main():
        url=input("请输入需要获取的音乐网址（仅网易云音乐）: ")
        get_song_html(url)

if __name__ == "__main__":
        main()

#单曲{id}{limit}{offset}
url_so="http://music.163.com/api/v1/resource/comments/R_SO_4_{}?limit={}&offset={}"
#专辑
url_al="http://music.163.com/api/v1/resource/comments/R_AL_3_{}?limit={}&offset={}"
#电台
url_dj="http://music.163.com/api/v1/resource/comments/A_DJ_1_{}?limit={}&offset={}"
#视频
url_vi="http://music.163.com/api/v1/resource/comments/R_VI_62_{}?limit={}&offset={}"
#MV
url_mv="http://music.163.com/api/v1/resource/comments/R_MV_5_{}?limit={}&offset={}"
#歌单
url_pl"http://music.163.com/api/v1/resource/comments/A_PL_0_{}?limit={}&offset={}"
#事件（这个好像比较特殊，有id和uid）
url_ev="http://music.163.com/api/v1/resource/comments/A_EV_2_{}_{}?limit={}&offset={}"