本帖最后由 天空宫阙 于 2020-2-9 16:28 编辑
先说结果吧，直接用浏览器浏览发现，网易云音乐的评论现在只公开最新和最旧各1000条左右，即前后各50页左右
所以要抓取一首歌的全部评论目前来说无解，app使用的接口是否公开所有评论未知。
但想着写都写了就发出来吧。
功能是按song_id抓取最多2000条评论，存入mongodb（一开始想爬40万条的所以选择数据库）
可利用mongodb的查询语句查询定 userId 的评论
{"userId":123} # 精确查找userId为123的评论
{"content":{"$regex":"123"}} #模糊查找评论内容包含123字符的评论
[Python] 纯文本查看 复制代码 from Crypto.Cipher import AES
import base64
import hashlib
import json
import requests
import pymongo
import time
from requests import exceptions
# 连接数据库
client = pymongo.MongoClient()
db = client['Netease_2']
def save_to_mongo(data):
""" 存入mongodb数据库 """
# 按commentId更新
if db['comments'].update({'commentId': data['commentId']}, {'$set': data}, True):
print('Saved to Mongo', data["nickname"],data["content"])
else:
print('Saved to Mongo Failed', data['commentId'])
class Netease:
def __init__(self):
self.key1 = b'0CoJUm6Qyw8W8jud'
self.key2 = b"ihPb70NS86eAsg9M"
self.encSecKey = "3bc924d00a13c5c27664d436967a26bb13cca16e576f53ed642298406b26ffd3340805b2c563969cb6462d2c93c41c50e6684e1dce8fb63f6e3fd5c95e8e8375f727ebe195530c4989bcc897e3eec4107850ee126b6f95a184432d0eaf6cba1c6e0e37e86d83bf7e3f670f7790d3d8ded5f7391f746554311f31cb9b430fff10"
self.login_url_phone = 'http://music.163.com/weapi/login/cellphone'
self.base_comments_API = "https://music.163.com/weapi/v1/resource/comments/R_SO_4_"
self.csrf_token = ''
self.headers = {
'Accept':'*/*',
'Accept-Language':'zh-CN,zh;q=0.8,gl;q=0.6,zh-TW;q=0.4',
'Connection':'keep-alive',
'Content-Type':'application/x-www-form-urlencoded',
'Referer':'http://music.163.com',
'Host':'music.163.com',
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/72.0.3626.119 Safari/537.36'
}
self.session = requests.Session()
self.session.headers.update(self.headers)
def _encrypt(self,data):
""" 两次aes加密 """
data_string = json.dumps(data)
temp = self._ase_encrypt(self.key1,data_string)
return self._ase_encrypt(self.key2,temp),self.encSecKey
def _ase_encrypt(self,key,content):
pad = lambda s: s + (16 - len(s) % 16) * chr(16 - len(s) % 16)
my_ase = AES.new(key=key,mode=AES.MODE_CBC,IV=b'0102030405060708')
return base64.b64encode(my_ase.encrypt(pad(content).encode())).decode()
def login(self,phone,password):
""" 登录，此处仅支持手机号，邮箱另一个接口 """
md5 = hashlib.md5()
md5.update(password.encode('utf-8'))
password = md5.hexdigest()
data = {
'phone':phone,
'password': password,
'rememberLogin': True
}
params,encSecKey = self._encrypt(data)
post_data = {
'params': params,
'encSecKey': encSecKey,
}
response = self.session.post(self.login_url_phone,data=post_data)
if response.status_code ==200:
res_json = response.json()
# 登录成功
if res_json['code'] ==200:
print('login successfully')
ckjar = self.session.cookies
ck_dict = requests.utils.dict_from_cookiejar(ckjar)
self.csrf_token = ck_dict['__csrf']
# print(self.csrf_token)
elif(res_json['code'] == 400) or (res_json['code'] == 502):
print('fail to login, username or password error')
else:
print(res_json.get('msg'))
def get_one_page(self,song_id,page):
""" 抓取一页 """
print('current page:',page)
data = {"rid":f"R_SO_4_{str(song_id)}","offset":f"{str((page-1)*20)}","total":"false","limit":"20","csrf_token":f"{self.csrf_token}"}
params,encSecKey = self._encrypt(data)
post_data = {
'params': params,
'encSecKey': encSecKey,
}
try:
response = self.session.post(url=self.base_comments_API+str(song_id),data=post_data)
if response.status_code ==200:
# print(response.json())
return response.json()
else:
print(response.status_code)
except exceptions.ProxyError:
time.sleep(3)
print('try again...')
return self.get_one_page(song_id,page)
def get_last_50_comments(self,song_id,total):
""" 抓取最后50页 """
total_page = int(int(total)/20)+1
for page in range(total_page-50,total_page+1):
res_json = self.get_one_page(song_id,page)
more = res_json['more']
comments = res_json['comments']
for comment in comments:
data = self.parse_each_comment(comment)
data['song_id'] = song_id
save_to_mongo(data)
time.sleep(3)
if not more:
break
def parse_each_comment(self,comment):
nickname = comment["user"]["nickname"]
userId = comment["user"]["userId"]
content = comment["content"]
time = comment["time"]
likedCount = comment["likedCount"]
commentId = comment["commentId"]
return {
"userId":userId,
"nickname":nickname,
"content":content,
"time":time,
"likedCount":likedCount,
"commentId":commentId
}
def run(self):
# 登录没用
# self.login('','')
# 抓取指定song_id下的最多约2000条评论
song_id = 436514312
for page in range(1,101):
res_json = self.get_one_page(song_id,page)
more = res_json['more']
comments = res_json['comments']
for comment in comments:
data = self.parse_each_comment(comment)
data['song_id'] = song_id
save_to_mongo(data)
time.sleep(3)
if not more:
if page <100:
"""需要抓取最后50页 """
total = self.get_one_page(song_id,1)['total']
self.get_last_50_comments(song_id,total)
break
if __name__ == "__main__":
t = Netease()
t.run()
""" MongoDB的query """
# {"userId":123} # 精确查找userId为123的评论
# {"content":{"$regex":"123"}} #模糊查找评论内容包含123字符的评论
可扩展为查询某用户的评论若该用户近期有评论，被评论歌曲评论总数未超过2000，大概率可以查到，反之亦反
（自己扩展吧song_id可从，用户动态里分享的歌曲，听歌排行榜歌曲（周榜及总榜），收藏的歌曲及创建歌单内的歌曲，收藏歌单的歌曲中获取） | 
发表于 2020-2-9 16:23
