通过websocket抓取斗鱼弹幕和礼物消息

小鱼儿 2022-09-12 01:55 334阅读 0赞

1.斗鱼弹幕协议

到斗鱼官方开放平台看斗鱼通讯协议，网址“[斗鱼开放平台][Link 1]”，登录后可查看

![d49b87191062cf54028e43ce79795892.png][]

所以根据斗鱼协议做编码函数：

![复制代码][48304ba5e6f9fe08f3fa1abda7d326ab.png]

1     def msg_encode(msg):
     2         #消息以 \0 结尾，并以utf-8编码
     3         msg = msg + '\0'
     4         msg_bytes = msg.encode('utf-8')
     5         #消息长度 + 头部长度8
     6         length_bytes = int.to_bytes(len(msg) + 8, 4, byteorder='little')
     7         #斗鱼客户端发送消息类型 689
     8         type = 689
     9         type_bytes = int.to_bytes(type, 2, byteorder='little')
    10         # 加密字段与保留字段，默认 0 长度各 1
    11         end_bytes = int.to_bytes(0, 1, byteorder='little')
    12         #按顺序相加  消息长度 + 消息长度 + 消息类型 + 加密字段 + 保留字段
    13         head_bytes = length_bytes + length_bytes + type_bytes + end_bytes + end_bytes
    14         #消息头部拼接消息内容
    15         data = head_bytes + msg_bytes
    16         return data

![复制代码][48304ba5e6f9fe08f3fa1abda7d326ab.png]

然后根据斗鱼协议做解码函数：

![复制代码][48304ba5e6f9fe08f3fa1abda7d326ab.png]

1     def msg_decode(msg_bytes):
     2         #定义一个游标位置
     3         cursor = 0
     4         msg = []
     5         while cursor < len(msg_bytes):
     6             #根据斗鱼协议，报文 前四位与第二个四位，都是消息长度，取前四位，转化成整型
     7             content_length = int.from_bytes(msg_bytes[cursor: (cursor + 4) - 1], byteorder='little')
     8             #报文长度不包含前4位，从第5位开始截取消息长度的字节流，并扣除前8位的协议头，取出正文，用utf-8编码成字符串
     9             content = msg_bytes[(cursor + 4) + 8:(cursor + 4) + content_length - 1].decode(encoding='utf-8',
    10                                                                                            errors='ignore')
    11             msg.append(content)
    12             cursor = (cursor + 4) + content_length
    13         # print(msg)
    14         return msg

![复制代码][48304ba5e6f9fe08f3fa1abda7d326ab.png]

解码后的消息需要反序列化：

![复制代码][48304ba5e6f9fe08f3fa1abda7d326ab.png]

1     def msg_format(msg_str):
     2         try:
     3             msg_dict = {}
     4             msg_list = msg_str.split('/')[0:-1]
     5             for msg in msg_list:
     6                 msg = msg.replace('@s', '/').replace('@A', '@')
     7                 msg_tmp = msg.split('@=')
     8                 msg_dict[msg_tmp[0]] = msg_tmp[1]
     9             return msg_dict
    10         except Exception as e:
    11             print(str(e))

![复制代码][48304ba5e6f9fe08f3fa1abda7d326ab.png]

2.抓包分析

建议通过chrome开发者工具抓包，或者使用fiddler抓包，网页的东西，直接用chrome更方便

为什么不用斗鱼开放平台提供的api直接用呢，因为api是面向注册的开发者的，api都需要申请使用，普通用户只能抓包分析通用接口

![53d7f77cf0b958c6b1ad90c1e37a5a33.png][]

分析：

wsproxy.douyu.com中交互消息可以获得多个弹幕服务器地址，我们直接取一个就行，不需要在程序中实现实时分析

然后分析danmuproxy.douyu.com连接中详细交互信息，实现与服务器交互

匿名登录时弹幕中礼物消息可能默认被屏蔽，需在登录时提交开启指令，或登录后提交开启指令

弹幕消息类型与指令非常丰富，这里举例简单的消息和指令，详细可以继续分析

结果如下：

\#\#匿名分配的username与uid可以从wsproxy.douyu.com中交互消息获取，也可直接自定义提交就行

退出登录：type@=logout/

心跳消息：type@=mrkl/

加入组消息：type@=joingroup/rid@=74751/gid@=-9999/           \#gid默认1，此处改成 - 9999 改成海量弹幕模式

屏蔽礼物消息：type@=dmfbdreq/dfl@=sn@AA=105@ASss@AA=1@AS@Ssn@AA=106@ASss@AA=1@AS@Ssn@AA=107@ASss@AA=1@AS@Ssn@AA=108@ASss@AA=1@AS@Ssn@AA=110@ASss@AA=1@AS@Ssn@AA=901@ASss@AA=1@AS@S/

开启礼物消息：type@=dmfbdreq/dfl@=sn@AA=105@ASss@AA=0@AS@Ssn@AA=106@ASss@AA=0@AS@Ssn@AA=107@ASss@AA=0@AS@Ssn@AA=108@ASss@AA=0@AS@Ssn@AA=110@ASss@AA=0@AS@Ssn@AA=901@ASss@AA=0@AS@S/

\#\#开启与屏蔽礼物可以详细控制，可以逐[ python教程][python]项尝试分析

![37f61ca1675aefd00ae28bdde31afaac.png][]

注：登录消息中有个字段/dfl@=，此处直接加入屏蔽礼物或开启礼物指令中相应dfl@=后面的内容，即可在登录时控制礼物屏蔽或开启

3.礼物消息处理

获取斗鱼礼物类型，需抓包分析来源网址

url1 = 'https://webconf.douyucdn.cn/resource/common/gift/flash/gift\_effect.json'

url2 = 'https://webconf.douyucdn.cn/resource/common/prop\_gift\_list/prop\_gift\_config.json'

将获取到的礼物json处理一下，合并且去除不需要信息，做成字典：

![复制代码][48304ba5e6f9fe08f3fa1abda7d326ab.png]

1     def get_gift_dict():
     2         gift_json = {}
     3         gift_json1 = requests.get('https://webconf.douyucdn.cn/resource/common/gift/flash/gift_effect.json').text
     4         gift_json2 = requests.get('https://webconf.douyucdn.cn/resource/common/prop_gift_list/prop_gift_config.json').text
     5         gift_json1=gift_json1.replace('DYConfigCallback(','')[0:-2]
     6         gift_json2=gift_json2.replace('DYConfigCallback(','')[0:-2]
     7         gift_json1 = json.loads(gift_json1)['data']['flashConfig']
     8         gift_json2= json.loads(gift_json2)['data']
     9         for gift in gift_json1:
    10             gift_json[gift] = gift_json1[gift]['name']
    11         for gift in gift_json2:
    12             gift_json[gift] = gift_json2[gift]['name']
    13         return gift_json

![复制代码][48304ba5e6f9fe08f3fa1abda7d326ab.png]

4.websocket连接

python环境先安装websocket包    pip3 install websocket-client

引入websocket包后，使用websecket.WebSocketApp建立websecket连接客户端，创建时在 on\_open、on\_error、on\_message、on\_close 4个参数中传入相应事件发生时需要处理的方法

顾名思义，4个参数很好理解了

websocket.WebSocketApp(url, on_open=on_open, on_error=on_error,on_message=on_message, on_close=on_close)

5.效果与源码

实际效果展示：

![65b858fb086aaf5aba4d9ff31fb8787a.png][]

源码（演示代码，异常判断，数据校验就不做了）：

![复制代码][48304ba5e6f9fe08f3fa1abda7d326ab.png]

1 __author__ = 'admin'
      2 import websocket
      3 import threading
      4 import time
      5 import requests
      6 import json
      7 
      8 
      9 class DyDanmu:
     10     def __init__(self, roomid, url):
     11         self.gift_dict = self.get_gift_dict()
     12         self.gift_dict_keys = self.gift_dict.keys()
     13         self.room_id = roomid
     14         self.client = websocket.WebSocketApp(url, on_open=self.on_open, on_error=self.on_error,
     15                                              on_message=self.on_message, on_close=self.on_close)
     16         self.heartbeat_thread = threading.Thread(target=self.heartbeat)
     17 
     18     def start(self):
     19         self.client.run_forever()
     20 
     21     def stop(self):
     22         self.logout()
     23         self.client.close()
     24 
     25     def on_open(self):
     26         self.login()
     27         self.join_group()
     28         self.heartbeat_thread.setDaemon(True)
     29         self.heartbeat_thread.start()
     30 
     31 
     32     def on_error(self, error):
     33         print(error)
     34 
     35     def on_close(self):
     36         print('close')
     37 
     38     def send_msg(self, msg):
     39         msg_bytes = self.msg_encode(msg)
     40         self.client.send(msg_bytes)
     41 
     42     def on_message(self, msg):
     43         message = self.msg_decode(msg)
     44         # print(message)
     45         for msg_str in message:
     46             msg_dict = self.msg_format(msg_str)
     47             if msg_dict['type'] == 'chatmsg':
     48                 print(msg_dict['nn'] + ':' + msg_dict['txt'])
     49             if msg_dict['type'] == 'dgb':
     50                 if msg_dict['gfid'] in self.gift_dict_keys:
     51                     print(msg_dict['nn'] + '\t送出\t' + msg_dict['gfcnt'] + '\t个\t' + self.gift_dict[msg_dict['gfid']])
     52                 else:
     53                     print(msg_dict['nn'] + '\t送出\t' + msg_dict['gfcnt'] + '\t个\t' + msg_dict['gfid'] + '\t未知礼物')
     54                     # print(msg_dict)
     55 
     56     # 发送登录信息
     57     def login(self):
     58         login_msg = 'type@=loginreq/roomid@=%s/' \
     59                     'dfl@=sn@AA=105@ASss@AA=0@AS@Ssn@AA=106@ASss@AA=0@AS@Ssn@AA=107@ASss@AA=0@AS@Ssn@AA=108@ASss@AA=0@AS@Ssn@AA=110@ASss@AA=0@AS@Ssn@AA=901@ASss@AA=0/' \
     60                     'username@=%s/uid@=%s/ltkid@=/biz@=/stk@=/devid@=8d8c22ce6093e6a7264f99da00021501/ct@=0/pt@=2/cvr@=0/tvr@=7/apd@=/rt@=1605498503/vk@=0afb8a90c2cb545e8459d60c760dc08b/' \
     61                     'ver@=20190610/aver@=218101901/dmbt@=chrome/dmbv@=78/' % (
     62                         self.room_id, 'visitor4444086', '1178849206'
     63                     )
     64         self.send_msg(login_msg)
     65 
     66     def logout(self):
     67         logout_msg = 'type@=logout/'
     68         self.send_msg(logout_msg)
     69 
     70     # 发送入组消息
     71     def join_group(self):
     72         join_group_msg = 'type@=joingroup/rid@=%s/gid@=-9999/' % (self.room_id)
     73         self.send_msg(join_group_msg)
     74 
     75     # 关闭礼物信息推送
     76     def close_gift(self):
     77         close_gift_msg = 'type@=dmfbdreq/dfl@=sn@AA=105@ASss@AA=1@AS@Ssn@AA=106@ASss@AA=1@AS@Ssn@AA=107@ASss@AA=1@AS@Ssn@AA=108@ASss@AA=1@AS@Ssn@AA=110@ASss@AA=1@AS@Ssn@AA=901@ASss@AA=1@AS@S/'
     78         self.send_msg(close_gift_msg)
     79 
     80     # 保持心跳线程
     81     def heartbeat(self):
     82         while True:
     83             # 45秒发送一个心跳包
     84             self.send_msg('type@=mrkl/')
     85             print('发送心跳')
     86             time.sleep(45)
     87 
     88 
     89     def msg_encode(self, msg):
     90         # 消息以 \0 结尾，并以utf-8编码
     91         msg = msg + '\0'
     92         msg_bytes = msg.encode('utf-8')
     93         #消息长度 + 头部长度8
     94         length_bytes = int.to_bytes(len(msg) + 8, 4, byteorder='little')
     95         #斗鱼客户端发送消息类型 689
     96         type = 689
     97         type_bytes = int.to_bytes(type, 2, byteorder='little')
     98         # 加密字段与保留字段，默认 0 长度各 1
     99         end_bytes = int.to_bytes(0, 1, byteorder='little')
    100         #按顺序相加  消息长度 + 消息长度 + 消息类型 + 加密字段 + 保留字段
    101         head_bytes = length_bytes + length_bytes + type_bytes + end_bytes + end_bytes
    102         #消息头部拼接消息内容
    103         data = head_bytes + msg_bytes
    104         return data
    105 
    106     def msg_decode(self, msg_bytes):
    107         # 定义一个游标位置
    108         cursor = 0
    109         msg = []
    110         while cursor < len(msg_bytes):
    111             #根据斗鱼协议，报文 前四位与第二个四位，都是消息长度，取前四位，转化成整型
    112             content_length = int.from_bytes(msg_bytes[cursor: (cursor + 4) - 1], byteorder='little')
    113             #报文长度不包含前4位，从第5位开始截取消息长度的字节流，并扣除前8位的协议头，取出正文，用utf-8编码成字符串
    114             content = msg_bytes[(cursor + 4) + 8:(cursor + 4) + content_length - 1].decode(encoding='utf-8',
    115                                                                                            errors='ignore')
    116             msg.append(content)
    117             cursor = (cursor + 4) + content_length
    118         # print(msg)
    119         return msg
    120 
    121     def msg_format(self, msg_str):
    122         try:
    123             msg_dict = {}
    124             msg_list = msg_str.split('/')[0:-1]
    125             for msg in msg_list:
    126                 msg = msg.replace('@s', '/').replace('@A', '@')
    127                 msg_tmp = msg.split('@=')
    128                 msg_dict[msg_tmp[0]] = msg_tmp[1]
    129             return msg_dict
    130         except Exception as e:
    131             print(str(e))
    132 
    133     def get_gift_dict(self):
    134         gift_json = {}
    135         gift_json1 = requests.get('https://webconf.douyucdn.cn/resource/common/gift/flash/gift_effect.json').text
    136         gift_json2 = requests.get(
    137             'https://webconf.douyucdn.cn/resource/common/prop_gift_list/prop_gift_config.json').text
    138         gift_json1 = gift_json1.replace('DYConfigCallback(', '')[0:-2]
    139         gift_json2 = gift_json2.replace('DYConfigCallback(', '')[0:-2]
    140         gift_json1 = json.loads(gift_json1)['data']['flashConfig']
    141         gift_json2 = json.loads(gift_json2)['data']
    142         for gift in gift_json1:
    143             gift_json[gift] = gift_json1[gift]['name']
    144         for gift in gift_json2:
    145             gift_json[gift] = gift_json2[gift]['name']
    146         return gift_json
    147 
    148 
    149 if __name__ == '__main__':
    150     roomid = '74751'
    151     url = 'wss://danmuproxy.douyu.com:8506/'
    152     dy = DyDanmu(roomid, url)
    153     dy.start()

![复制代码][48304ba5e6f9fe08f3fa1abda7d326ab.png]

[Link 1]: https://open.douyu.com/source/api/63
[d49b87191062cf54028e43ce79795892.png]: /images/20220828/752792e60158401e90c76e67ebd061c2.png
[48304ba5e6f9fe08f3fa1abda7d326ab.png]: /images/20220828/0422ddd4f0a0490e986b1fe6523fbd49.png
[53d7f77cf0b958c6b1ad90c1e37a5a33.png]: /images/20220828/d09e2abb57eb4df99c04a6023951a2ce.png
[python]: https://www.xin3721.com/eschool/pythonxin3721/
[37f61ca1675aefd00ae28bdde31afaac.png]: /images/20220828/2aed200fb91f4db4854370366ab2dbb5.png
[65b858fb086aaf5aba4d9ff31fb8787a.png]: /images/20220828/56ae2f4206fe470795c027c745069402.png