分类目录:《大模型从入门到应用》总目录

LangChain系列文章:

  • 基础知识
  • 快速入门
    • 安装与环境配置
    • 链(Chains)、代理(Agent:)和记忆(Memory)
    • 快速开发聊天模型
  • 模型(Models)
    • 基础知识
    • 大型语言模型(LLMs)
      • 基础知识
      • LLM的异步API、自定义LLM包装器、虚假LLM和人类输入LLM(Human Input LLM)
      • 缓存LLM的调用结果
      • 加载与保存LLM类、流式传输LLM与Chat Model响应和跟踪tokens使用情况
    • 聊天模型(Chat Models)
      • 基础知识
      • 使用少量示例和响应流式传输
    • 文本嵌入模型
      • Aleph Alpha、Amazon Bedrock、Azure OpenAI、Cohere等
      • Embaas、Fake Embeddings、Google Vertex AI PaLM等
  • 提示(Prompts)
    • 基础知识
    • 提示模板
      • 基础知识
      • 连接到特征存储
      • 创建自定义提示模板和含有Few-Shot示例的提示模板
      • 部分填充的提示模板和提示合成
      • 序列化提示信息
    • 示例选择器(Example Selectors)
    • 输出解析器(Output Parsers)
  • 记忆(Memory)
    • 基础知识
    • 记忆的类型
      • 会话缓存记忆、会话缓存窗口记忆和实体记忆
      • 对话知识图谱记忆、对话摘要记忆和会话摘要缓冲记忆
      • 对话令牌缓冲存储器和基于向量存储的记忆
    • 将记忆添加到LangChain组件中
    • 自定义对话记忆与自定义记忆类
    • 聊天消息记录
    • 记忆的存储与应用
  • 索引(Indexes)
    • 基础知识
    • 文档加载器(Document Loaders)
    • 文本分割器(Text Splitters)
    • 向量存储器(Vectorstores)
    • 检索器(Retrievers)
  • 链(Chains)
    • 基础知识
    • 通用功能
      • 自定义Chain和Chain的异步API
      • LLMChain和RouterChain
      • SequentialChain和TransformationChain
      • 链的保存(序列化)与加载(反序列化)
    • 链与索引
      • 文档分析和基于文档的聊天
      • 问答的基础知识
      • 图问答(Graph QA)和带来源的问答(Q&A with Sources)
      • 检索式问答
      • 文本摘要(Summarization)、HyDE和向量数据库的文本生成
  • 代理(Agents)
    • 基础知识
    • 代理类型
    • 自定义代理(Custom Agent)
    • 自定义MRKL代理
    • 带有ChatModel的LLM聊天自定义代理和自定义多操作代理(Custom MultiAction Agent)
    • 工具
      • 基础知识
      • 自定义工具(Custom Tools)
      • 多输入工具和工具输入模式
      • 人工确认工具验证和Tools作为OpenAI函数
    • 工具包(Toolkit)
    • 代理执行器(Agent Executor)
      • 结合使用Agent和VectorStore
      • 使用Agents的异步API和创建ChatGPT克隆
      • 处理解析错误、访问中间步骤和限制最大迭代次数
      • 为代理程序设置超时时间和限制最大迭代次数和为代理程序和其工具添加共享内存
    • 计划与执行
  • 回调函数(Callbacks)

----### 多输入工具
本节演示如何在Agent中使用需要多个输入的工具,推荐的方法是使用StructuredTool类。

import os
os.environ["LANGCHAIN_TRACING"] = "true"
from langchain import OpenAI
from langchain.agents import initialize_agent, AgentType

llm = OpenAI(temperature=0)
from langchain.tools import StructuredTool

def multiplier(a: float, b: float) -> float:
    """Multiply the provided floats."""
    return a * b

tool = StructuredTool.from_function(multiplier)
# Structured tools are compatible with the STRUCTURED_CHAT_ZERO_SHOT_REACT_DESCRIPTION agent type. 
agent_executor = initialize_agent([tool], llm, agent=AgentType.STRUCTURED_CHAT_ZERO_SHOT_REACT_DESCRIPTION, verbose=True)
agent_executor.run("What is 3 times 4")

日志输出:

Entering new AgentExecutor chain...

Thought: I need to multiply 3 and 4
Action:

{
“action”: “multiplier”,
“action_input”: {“a”: 3, “b”: 4}
}


Observation: 12
Thought: I know what to respond
Action:

{
“action”: “Final Answer”,
“action_input”: “3 times 4 is 12”
}


Finished chain.

输出:

'3 times 4 is 12'
使用字符串格式的多输入工具

与结构化工具相比,另一种方法是使用常规的Tool类,并接受单个字符串作为输入。然后,工具必须处理解析逻辑,从文本中提取相关值,这将使工具表示与Agent提示紧密耦合。如果底层语言模型无法可靠地生成结构化模式,这仍然很有用。

让我们以乘法函数为例。为了使用它,我们将告诉Agent生成以逗号分隔的长度为两个的Action Input。然后,我们编写一个简单的包装器,将字符串分割成两部分(以逗号为界),并将两个解析后的值作为整数传递给乘法函数。

from langchain.llms import OpenAI
from langchain.agents import initialize_agent, Tool
from langchain.agents import AgentType

以下是乘法函数及其字符串解析器的示例:

def multiplier(a, b):
    return a * b

def parsing_multiplier(string):
    a, b = string.split(",")
    return multiplier(int(a), int(b))
llm = OpenAI(temperature=0)
tools = [
    Tool(
        name = "Multiplier",
        func=parsing_multiplier,
        description="useful for when you need to multiply two numbers together. The input to this tool should be a comma separated list of numbers of length two, representing the two numbers you want to multiply together. For example, `1,2` would be the input if you wanted to multiply 1 by 2."
    )
]
mrkl = initialize_agent(tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True)
mrkl.run("What is 3 times 4")

日志输出:

Entering new AgentExecutor chain...
 I need to multiply two numbers
Action: Multiplier
Action Input: 3,4
Observation: 12
Thought: I now know the final answer
Final Answer: 3 times 4 is 12

Finished chain.

输出:

'3 times 4 is 12'

工具输入模式

默认情况下,工具通过检查函数签名来推断参数模式。对于更严格的要求,可以指定自定义输入模式,以及自定义的验证逻辑。

from typing import Any, Dict

from langchain.agents import AgentType, initialize_agent
from langchain.llms import OpenAI
from langchain.tools.requests.tool import RequestsGetTool, TextRequestsWrapper
from pydantic import BaseModel, Field, root_validator

llm = OpenAI(temperature=0)

我们还需安装tldextract

!pip install tldextract > /dev/null
notice A new release of pip is available: 23.0.1 -> 23.1
notice To update, run: pip install --upgrade pip

输入:

import tldextract

_APPROVED_DOMAINS = {
    "langchain",
    "wikipedia",
}

class ToolInputSchema(BaseModel):

    url: str = Field(...)
    
    @root_validator
    def validate_query(cls, values: Dict[str, Any]) -> Dict:
        url = values["url"]
        domain = tldextract.extract(url).domain
        if domain not in _APPROVED_DOMAINS:
            raise ValueError(f"Domain {domain} is not on the approved list:"
                             f" {sorted(_APPROVED_DOMAINS)}")
        return values
    
tool = RequestsGetTool(args_schema=ToolInputSchema, requests_wrapper=TextRequestsWrapper())
agent = initialize_agent([tool], llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=False)
# This will succeed, since there aren't any arguments that will be triggered during validation
answer = agent.run("What's the main title on langchain?")
print(answer)

输出:

The main title of langchain is "LANG CHAIN 🦜️🔗 Official Home Page"

将会报错的输入:

agent.run("What's the main title on google?")

输出:

---------------------------------------------------------------------------

ValidationError                           Traceback (most recent call last)

Cell In[7], line 1
----> 1 agent.run("What's the main title on google?")


File ~/code/lc/lckg/langchain/chains/base.py:213, in Chain.run(self, *args, **kwargs)
    211     if len(args) != 1:
    212         raise ValueError("`run` supports only one positional argument.")
--> 213     return self(args[0])[self.output_keys[0]]
    215 if kwargs and not args:
    216     return self(kwargs)[self.output_keys[0]]


File ~/code/lc/lckg/langchain/chains/base.py:116, in Chain.__call__(self, inputs, return_only_outputs)
    114 except (KeyboardInterrupt, Exception) as e:
    115     self.callback_manager.on_chain_error(e, verbose=self.verbose)
--> 116     raise e
    117 self.callback_manager.on_chain_end(outputs, verbose=self.verbose)
    118 return self.prep_outputs(inputs, outputs, return_only_outputs)


File ~/code/lc/lckg/langchain/chains/base.py:113, in Chain.__call__(self, inputs, return_only_outputs)
    107 self.callback_manager.on_chain_start(
    108     {"name": self.__class__.__name__},
    109     inputs,
    110     verbose=self.verbose,
    111 )
    112 try:
--> 113     outputs = self._call(inputs)
    114 except (KeyboardInterrupt, Exception) as e:
    115     self.callback_manager.on_chain_error(e, verbose=self.verbose)


File ~/code/lc/lckg/langchain/agents/agent.py:792, in AgentExecutor._call(self, inputs)
    790 # We now enter the agent loop (until it returns something).
    791 while self._should_continue(iterations, time_elapsed):
--> 792     next_step_output = self._take_next_step(
    793         name_to_tool_map, color_mapping, inputs, intermediate_steps
    794     )
    795     if isinstance(next_step_output, AgentFinish):
    796         return self._return(next_step_output, intermediate_steps)


File ~/code/lc/lckg/langchain/agents/agent.py:695, in AgentExecutor._take_next_step(self, name_to_tool_map, color_mapping, inputs, intermediate_steps)
    693         tool_run_kwargs["llm_prefix"] = ""
    694     # We then call the tool on the tool input to get an observation
--> 695     observation = tool.run(
    696         agent_action.tool_input,
    697         verbose=self.verbose,
    698         color=color,
    699         **tool_run_kwargs,
    700     )
    701 else:
    702     tool_run_kwargs = self.agent.tool_run_logging_kwargs()


File ~/code/lc/lckg/langchain/tools/base.py:110, in BaseTool.run(self, tool_input, verbose, start_color, color, **kwargs)
    101 def run(
    102     self,
    103     tool_input: Union[str, Dict],
   (...)
    107     **kwargs: Any,
    108 ) -> str:
    109     """Run the tool."""
--> 110     run_input = self._parse_input(tool_input)
    111     if not self.verbose and verbose is not None:
    112         verbose_ = verbose


File ~/code/lc/lckg/langchain/tools/base.py:71, in BaseTool._parse_input(self, tool_input)
     69 if issubclass(input_args, BaseModel):
     70     key_ = next(iter(input_args.__fields__.keys()))
---> 71     input_args.parse_obj({key_: tool_input})
     72 # Passing as a positional argument is more straightforward for
     73 # backwards compatability
     74 return tool_input


File ~/code/lc/lckg/.venv/lib/python3.11/site-packages/pydantic/main.py:526, in pydantic.main.BaseModel.parse_obj()


File ~/code/lc/lckg/.venv/lib/python3.11/site-packages/pydantic/main.py:341, in pydantic.main.BaseModel.__init__()


ValidationError: 1 validation error for ToolInputSchema
__root__
  Domain google is not on the approved list: ['langchain', 'wikipedia'] (type=value_error)

参考文献:
[1] LangChain官方网站:https://www.langchain/
[2] LangChain 🦜️🔗 中文网,跟着LangChain一起学LLM/GPT开发:https://www.langchain/
[3] LangChain中文网 - LangChain 是一个用于开发由语言模型驱动的应用程序的框架:http://wwwlangchain/