File size: 6,267 Bytes
20048e3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
161
162
163
164
165
166
167
168
169
170
171
172
173
from typing import Dict, Any, List

from langchain.memory import ConversationBufferWindowMemory
from langchain.schema import (
    BaseMessage,
    HumanMessage,
    AIMessage,
)

from .registry import BaseParent


class ChatGLMConversationBufferWindowMemory(ConversationBufferWindowMemory):

    human_prefix: str = "问"
    ai_prefix: str = "答"

    def load_memory_variables(self, inputs: Dict[str, Any]) -> Dict[str, str]:
        """Return history buffer."""

        if self.return_messages:
            buffer: Any = self.buffer[-self.k * 2:]
        else:
            buffer = self.get_buffer_string(
                self.buffer[-self.k * 2:],
                human_prefix=self.human_prefix,
                ai_prefix=self.ai_prefix,
            )
        return {self.memory_key: buffer}

    @staticmethod
    def get_buffer_string(
        messages: List[BaseMessage], human_prefix: str = "问", ai_prefix: str = "答"
    ) -> str:
        """Get buffer string of messages."""
        string_messages, i = [], 0
        for m in messages:
            if isinstance(m, HumanMessage):
                role = human_prefix
                string_messages.append(f"[Round {i}]\n{role}{m.content}")
                i += 1
            elif isinstance(m, AIMessage):
                role = ai_prefix
                string_messages.append(f"{role}{m.content}")
            else:
                raise ValueError(f"Got unsupported message type: {m}")

        return "\n".join(string_messages) + f"\n[Round {i}]"


class ChineseAlpacaConversationBufferWindowMemory(ChatGLMConversationBufferWindowMemory):

    human_prefix: str = "### Instruction"
    ai_prefix: str = "### Response"

    @staticmethod
    def get_buffer_string(
        messages: List[BaseMessage], human_prefix: str = "### Instruction", ai_prefix: str = "### Response"
    ) -> str:
        """Get buffer string of messages."""
        string_messages = []
        for m in messages:
            if isinstance(m, HumanMessage):
                role = human_prefix
                string_messages.append(f"{role}:\n\n{m.content}")
            elif isinstance(m, AIMessage):
                role = ai_prefix
                string_messages.append(f"{role}:\n\n{m.content}")
            else:
                raise ValueError(f"Got unsupported message type: {m}")

        return "\n\n".join(string_messages)


class FireFlyConversationBufferWindowMemory(ChatGLMConversationBufferWindowMemory):

    human_prefix: str = "Human"
    ai_prefix: str = "Assistant"

    @staticmethod
    def get_buffer_string(
        messages: List[BaseMessage], human_prefix: str = "Human", ai_prefix: str = "Assistant"
    ) -> str:
        """Get buffer string of messages."""
        string_messages = []
        for m in messages:
            if isinstance(m, HumanMessage):
                string_messages.append(f"<s>{m.content}</s>")
            elif isinstance(m, AIMessage):
                string_messages.append(f"</s>{m.content}</s>")
            else:
                raise ValueError(f"Got unsupported message type: {m}")

        return "".join(string_messages)


class PhoenixConversationBufferWindowMemory(ChatGLMConversationBufferWindowMemory):

    human_prefix: str = "Human"
    ai_prefix: str = "Assistant"

    @staticmethod
    def get_buffer_string(
        messages: List[BaseMessage], human_prefix: str = "Human", ai_prefix: str = "Assistant"
    ) -> str:
        """Get buffer string of messages."""
        string_messages = []
        for m in messages:
            if isinstance(m, HumanMessage):
                role = human_prefix
            elif isinstance(m, AIMessage):
                role = ai_prefix
            else:
                raise ValueError(f"Got unsupported message type: {m}")
            string_messages.append(f"{role}: <s>{m.content}</s>")
        return "".join(string_messages)


class MossConversationBufferWindowMemory(ChatGLMConversationBufferWindowMemory):

    human_prefix: str = "<|Human|>"
    ai_prefix: str = "<|MOSS|>"

    @staticmethod
    def get_buffer_string(
        messages: List[BaseMessage], human_prefix: str = "<|Human|>", ai_prefix: str = "<|MOSS|>"
    ) -> str:
        """Get buffer string of messages."""
        string_messages = []
        for m in messages:
            if isinstance(m, HumanMessage):
                string_messages.append(f"{human_prefix}: {m.content}<eoh>")
            elif isinstance(m, AIMessage):
                string_messages.append(f"{ai_prefix}: {m.content}<eom>")
            else:
                raise ValueError(f"Got unsupported message type: {m}")
        return "\n".join(string_messages)


class GuanacoConversationBufferWindowMemory(ChatGLMConversationBufferWindowMemory):

    human_prefix: str = "### Human"
    ai_prefix: str = "### Assistant"

    @staticmethod
    def get_buffer_string(
        messages: List[BaseMessage], human_prefix: str = "### Human", ai_prefix: str = "### Assistant"
    ) -> str:
        """Get buffer string of messages."""
        string_messages = []
        for m in messages:
            if isinstance(m, HumanMessage):
                string_messages.append(f"{human_prefix}: {m.content}")
            elif isinstance(m, AIMessage):
                string_messages.append(f"{ai_prefix}: {m.content}")
            else:
                raise ValueError(f"Got unsupported message type: {m}")
        return "\n".join(string_messages)


class CustomConversationBufferWindowMemory(BaseParent):

    registry = {}


CustomConversationBufferWindowMemory.add_to_registry("gpt-3.5-turbo", ConversationBufferWindowMemory)
CustomConversationBufferWindowMemory.add_to_registry("chatglm", ChatGLMConversationBufferWindowMemory)
CustomConversationBufferWindowMemory.add_to_registry("chinese-llama-alpaca", ChineseAlpacaConversationBufferWindowMemory)
CustomConversationBufferWindowMemory.add_to_registry("firefly", FireFlyConversationBufferWindowMemory)
CustomConversationBufferWindowMemory.add_to_registry("phoenix", PhoenixConversationBufferWindowMemory)
CustomConversationBufferWindowMemory.add_to_registry("moss", MossConversationBufferWindowMemory)
CustomConversationBufferWindowMemory.add_to_registry("guanaco", GuanacoConversationBufferWindowMemory)