summaryrefslogtreecommitdiffstats
path: root/g4f/Provider/Liaobots.py
blob: b16b44d3f47febafd008fb60f4c7a432f025352d (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
from __future__ import annotations

import uuid

from aiohttp import ClientSession, BaseConnector

from ..typing import AsyncResult, Messages
from .base_provider import AsyncGeneratorProvider, ProviderModelMixin

models = {
    "gpt-4": {
        "id": "gpt-4",
        "name": "GPT-4",
        "maxLength": 24000,
        "tokenLimit": 8000,
    },
    "gpt-4-0613": {
        "id": "gpt-4-0613",
        "name": "GPT-4",
        "maxLength": 32000,
        "tokenLimit": 8000,
    },
    "gpt-3.5-turbo": {
        "id": "gpt-3.5-turbo",
        "name": "GPT-3.5-Turbo",
        "maxLength": 48000,
        "tokenLimit": 14000,
        "context": "16K",
    },
    "gpt-3.5-turbo-16k": {
        "id": "gpt-3.5-turbo-16k",
        "name": "GPT-3.5-16k",
        "maxLength": 48000,
        "tokenLimit": 16000,
    },
    "gpt-4-1106-preview": {
        "id": "gpt-4-1106-preview",
        "name": "GPT-4-Turbo",
        "maxLength": 260000,
        "tokenLimit": 126000,
        "context": "128K",
    },
    "gpt-4-plus": {
        "id": "gpt-4-plus",
        "name": "GPT-4-Plus",
        "maxLength": 130000,
        "tokenLimit": 31000,
        "context": "32K",
    },
    "gemini-pro": {
        "id": "gemini-pro",
        "name": "Gemini-Pro",
        "maxLength": 120000,
        "tokenLimit": 30000,
        "context": "32K",
    },
    "claude-2": {
        "id": "claude-2",
        "name": "Claude-2-200k",
        "maxLength": 800000,
        "tokenLimit": 200000,
        "context": "200K",
    },
    "claude-instant-1": {
        "id": "claude-instant-1",
        "name": "Claude-instant-1",
        "maxLength": 400000,
        "tokenLimit": 100000,
        "context": "100K",
    }
}

class Liaobots(AsyncGeneratorProvider, ProviderModelMixin):
    url = "https://liaobots.site"
    working = True
    supports_message_history = True
    supports_gpt_35_turbo = True
    supports_gpt_4 = True
    default_model = "gpt-3.5-turbo"
    models = [m for m in models]
    model_aliases = {
        "claude-v2": "claude-2"
    }
    _auth_code = None
    _cookie_jar = None

    @classmethod
    async def create_async_generator(
        cls,
        model: str,
        messages: Messages,
        auth: str = None,
        proxy: str = None,
        connector: BaseConnector = None,
        **kwargs
    ) -> AsyncResult:
        headers = {
            "authority": "liaobots.com",
            "content-type": "application/json",
            "origin": cls.url,
            "referer": f"{cls.url}/",
            "user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/112.0.0.0 Safari/537.36",
        }
        if proxy and not connector:
            try:
                from aiohttp_socks import ProxyConnector
                connector = ProxyConnector.from_url(proxy)
            except ImportError:
                raise RuntimeError('Install "aiohttp_socks" package for proxy support')
        async with ClientSession(
            headers=headers,
            cookie_jar=cls._cookie_jar,
            connector=connector
        ) as session:
            cls._auth_code = auth if isinstance(auth, str) else cls._auth_code
            if not cls._auth_code:
                async with session.post(
                    "https://liaobots.work/recaptcha/api/login",
                    proxy=proxy,
                    data={"token": "abcdefghijklmnopqrst"},
                    verify_ssl=False
                ) as response:
                    response.raise_for_status()
                async with session.post(
                    "https://liaobots.work/api/user",
                    proxy=proxy,
                    json={"authcode": ""},
                    verify_ssl=False
                ) as response:
                    response.raise_for_status()
                    cls._auth_code = (await response.json(content_type=None))["authCode"]
                    cls._cookie_jar = session.cookie_jar
                    
            data = {
                "conversationId": str(uuid.uuid4()),
                "model": models[cls.get_model(model)],
                "messages": messages,
                "key": "",
                "prompt": kwargs.get("system_message", "You are ChatGPT, a large language model trained by OpenAI. Follow the user's instructions carefully."),
            }
            async with session.post(
                "https://liaobots.work/api/chat",
                proxy=proxy,
                json=data,
                headers={"x-auth-code": cls._auth_code},
                verify_ssl=False
            ) as response:
                response.raise_for_status()
                async for chunk in response.content.iter_any():
                    if b"<html coupert-item=" in chunk:
                        raise RuntimeError("Invalid session")
                    if chunk:
                        yield chunk.decode()