mirror of
https://github.com/RYDE-WORK/Langchain-Chatchat.git
synced 2026-01-26 00:33:35 +08:00
9b62b1c72b
* dev分支解决pydantic版本冲突问题，增加ollama配置，支持ollama会话和向量接口 1、因dev版本的pydantic升级到了v2版本，由于在class History(BaseModel)中使用了from server.pydantic_v1，而fastapi的引用已变为pydantic的v2版本，所以fastapi用v2版本去校验用v1版本定义的对象，当会话历史histtory不为空的时候，会报错：TypeError: BaseModel.validate() takes 2 positional arguments but 3 were given。经测试，解方法为在class History(BaseModel)中也使用v2版本即可; 2、配置文件参照其它平台配置，增加了ollama平台相关配置，会话模型用户可根据实际情况自行添加，向量模型目前支持nomic-embed-text（必须升级ollama到0.1.29以上）。 3、因ollama官方只在会话部分对openai api做了兼容，向量api暂未适配，好在langchain官方库支持OllamaEmbeddings，因而在get_Embeddings方法中添加了相关支持代码。 * 修复 pydantic 升级到 v2 后 DocumentWithVsID 和 /v1/embeddings 兼容性问题 --------- Co-authored-by: srszzw <srszzw@163.com> Co-authored-by: liunux4odoo <liunux@qq.com>
47 lines
1.4 KiB
Python
47 lines
1.4 KiB
Python
from functools import lru_cache
|
||
from server.pydantic_v2 import BaseModel, Field
|
||
from langchain.prompts.chat import ChatMessagePromptTemplate
|
||
from configs import logger, log_verbose
|
||
from typing import List, Tuple, Dict, Union
|
||
|
||
|
||
class History(BaseModel):
|
||
"""
|
||
对话历史
|
||
可从dict生成，如
|
||
h = History(**{"role":"user","content":"你好"})
|
||
也可转换为tuple，如
|
||
h.to_msy_tuple = ("human", "你好")
|
||
"""
|
||
role: str = Field(...)
|
||
content: str = Field(...)
|
||
|
||
def to_msg_tuple(self):
|
||
return "ai" if self.role=="assistant" else "human", self.content
|
||
|
||
def to_msg_template(self, is_raw=True) -> ChatMessagePromptTemplate:
|
||
role_maps = {
|
||
"ai": "assistant",
|
||
"human": "user",
|
||
}
|
||
role = role_maps.get(self.role, self.role)
|
||
if is_raw: # 当前默认历史消息都是没有input_variable的文本。
|
||
content = "{% raw %}" + self.content + "{% endraw %}"
|
||
else:
|
||
content = self.content
|
||
|
||
return ChatMessagePromptTemplate.from_template(
|
||
content,
|
||
"jinja2",
|
||
role=role,
|
||
)
|
||
|
||
@classmethod
|
||
def from_data(cls, h: Union[List, Tuple, Dict]) -> "History":
|
||
if isinstance(h, (list,tuple)) and len(h) >= 2:
|
||
h = cls(role=h[0], content=h[1])
|
||
elif isinstance(h, dict):
|
||
h = cls(**h)
|
||
|
||
return h
|