File size: 6,297 Bytes
3079197
484e5ab
3079197
 
 
 
 
 
 
 
 
 
 
 
 
3245107
d0db329
4c52eb9
cfd888e
1550520
cfd888e
d0db329
3245107
34b2ab3
 
 
3245107
 
 
 
 
34b2ab3
 
 
3245107
 
a8294f2
4c52eb9
cfd888e
4c52eb9
 
 
cfd888e
 
 
 
 
4c52eb9
cfd888e
3245107
 
9fe9fc4
 
 
 
 
c1bdfb8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9fe9fc4
34b2ab3
d0db329
34b2ab3
 
 
 
 
3245107
 
a8294f2
3245107
34b2ab3
d0db329
a8294f2
 
3245107
cfd888e
 
3245107
cfd888e
 
 
 
 
 
 
5e0a689
 
 
 
 
 
 
 
 
a8294f2
cfd888e
 
 
 
 
 
 
 
 
 
 
 
1550520
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
154
155
156
157
158
159
160
#
#  Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
#
#  Licensed under the Apache License, Version 2.0 (the "License");
#  you may not use this file except in compliance with the License.
#  You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
#  Unless required by applicable law or agreed to in writing, software
#  distributed under the License is distributed on an "AS IS" BASIS,
#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#  See the License for the specific language governing permissions and
#  limitations under the License.
#
from abc import ABC
from openai import OpenAI
import openai
from rag.nlp import is_english
from rag.utils import num_tokens_from_string


class Base(ABC):
    def __init__(self, key, model_name):
        pass

    def chat(self, system, history, gen_conf):
        raise NotImplementedError("Please implement encode method!")


class GptTurbo(Base):
    def __init__(self, key, model_name="gpt-3.5-turbo"):
        self.client = OpenAI(api_key=key)
        self.model_name = model_name

    def chat(self, system, history, gen_conf):
        if system: history.insert(0, {"role": "system", "content": system})
        try:
            response = self.client.chat.completions.create(
                model=self.model_name,
                messages=history,
                **gen_conf)
            ans = response.output.choices[0]['message']['content'].strip()
            if response.output.choices[0].get("finish_reason", "") == "length":
                ans += "...\nFor the content length reason, it stopped, continue?" if is_english(
                    [ans]) else "路路路路路路\n鐢变簬闀垮害鐨勫師鍥狅紝鍥炵瓟琚埅鏂簡锛岃缁х画鍚楋紵"
            return ans, response.usage.completion_tokens
        except openai.APIError as e:
            return "**ERROR**: "+str(e), 0


class MoonshotChat(GptTurbo):
    def __init__(self, key, model_name="moonshot-v1-8k"):
        self.client = OpenAI(api_key=key, base_url="https://api.moonshot.cn/v1",)
        self.model_name = model_name

    def chat(self, system, history, gen_conf):
        if system: history.insert(0, {"role": "system", "content": system})
        try:
            response = self.client.chat.completions.create(
                model=self.model_name,
                messages=history,
                **gen_conf)
            ans = response.choices[0].message.content.strip()
            if response.choices[0].finish_reason == "length":
                ans += "...\nFor the content length reason, it stopped, continue?" if is_english(
                    [ans]) else "路路路路路路\n鐢变簬闀垮害鐨勫師鍥狅紝鍥炵瓟琚埅鏂簡锛岃缁х画鍚楋紵"
            return ans, response.usage.completion_tokens
        except openai.APIError as e:
            return "**ERROR**: "+str(e), 0


from dashscope import Generation
class QWenChat(Base):
    def __init__(self, key, model_name=Generation.Models.qwen_turbo):
        import dashscope
        dashscope.api_key = key
        self.model_name = model_name

    def chat(self, system, history, gen_conf):
        from http import HTTPStatus
        if system: history.insert(0, {"role": "system", "content": system})
        response = Generation.call(
            self.model_name,
            messages=history,
            result_format='message',
            **gen_conf
        )
        ans = ""
        tk_count = 0
        if response.status_code == HTTPStatus.OK:
            ans += response.output.choices[0]['message']['content']
            tk_count += response.usage.output_tokens
            if response.output.choices[0].get("finish_reason", "") == "length":
                ans += "...\nFor the content length reason, it stopped, continue?" if is_english([ans]) else "路路路路路路\n鐢变簬闀垮害鐨勫師鍥狅紝鍥炵瓟琚埅鏂簡锛岃缁х画鍚楋紵"
            return ans, tk_count

        return "**ERROR**: " + response.message, tk_count


from zhipuai import ZhipuAI
class ZhipuChat(Base):
    def __init__(self, key, model_name="glm-3-turbo"):
        self.client = ZhipuAI(api_key=key)
        self.model_name = model_name

    def chat(self, system, history, gen_conf):
        if system: history.insert(0, {"role": "system", "content": system})
        try:
            response = self.client.chat.completions.create(
                self.model_name,
                messages=history,
                **gen_conf
            )
            ans = response.output.choices[0]['message']['content'].strip()
            if response.output.choices[0].get("finish_reason", "") == "length":
                ans += "...\nFor the content length reason, it stopped, continue?" if is_english(
                    [ans]) else "路路路路路路\n鐢变簬闀垮害鐨勫師鍥狅紝鍥炵瓟琚埅鏂簡锛岃缁х画鍚楋紵"
            return ans, response.usage.completion_tokens
        except Exception as e:
            return "**ERROR**: " + str(e), 0

class LocalLLM(Base):
    class RPCProxy:
        def __init__(self, host, port):
            self.host = host
            self.port = int(port)
            self.__conn()

        def __conn(self):
            from multiprocessing.connection import Client
            self._connection = Client((self.host, self.port), authkey=b'infiniflow-token4kevinhu')

        def __getattr__(self, name):
            import pickle
            def do_rpc(*args, **kwargs):
                for _ in range(3):
                    try:
                        self._connection.send(pickle.dumps((name, args, kwargs)))
                        return pickle.loads(self._connection.recv())
                    except Exception as e:
                        self.__conn()
                raise Exception("RPC connection lost!")

            return do_rpc

    def __init__(self, key, model_name="glm-3-turbo"):
        self.client = LocalLLM.RPCProxy("127.0.0.1", 7860)

    def chat(self, system, history, gen_conf):
        if system: history.insert(0, {"role": "system", "content": system})
        try:
            ans = self.client.chat(
                history,
                gen_conf
            )
            return ans, num_tokens_from_string(ans)
        except Exception as e:
            return "**ERROR**: " + str(e), 0