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）

---

当构建我们自己的代理程序时，我们需要为其提供一个工具列表，供其使用。除了被调用的实际函数外，工具还包括几个组件：

• name[str]：在提供给代理程序的工具集中必须是唯一的
• description[str. 可选]：可选的，但建议提供，因为代理程序使用它来确定工具的使用方式
• return_direct[bool]：默认为False
• args_schema[可选]：可选的，但建议提供，可用于提供更多信息（例如少量示例）或验证预期参数

有两种主要方式来定义工具，我们将在下面的示例中涵盖这两种方式。

导入通用需要的模块

from langchain import LLMMathChain, SerpAPIWrapper
from langchain.agents import AgentType, initialize_agent
from langchain.chat_models import ChatOpenAI
from langchain.tools import BaseTool, StructuredTool, Tool, tool
llm = ChatOpenAI(temperature=0)

字符串输入和输出工具

最简单的工具接受一个查询字符串，并返回一个字符串输出。如果我们的工具函数需要多个参数，我们可能希望跳转到下面的StructuredTool部分。有两种方法可以实现这一点：

• 使用Tool数据类
• 继承BasTool类

Tool数据类

Tool数据类封装了接受单个字符串输入并返回字符串输出的函数。

# Load the tool configs that are needed.
search = SerpAPIWrapper()
llm_math_chain = LLMMathChain(llm=llm, verbose=True)
tools = [
    Tool.from_function(
        func=search.run,
        name = "Search",
        description="useful for when you need to answer questions about current events"
        # coroutine= ... <- you can specify an async method if desired as well
    ),
]

我们还可以定义自定义的args_schema以提供有关输入的更多信息：

from pydantic import BaseModel, Field

class CalculatorInput(BaseModel):
    question: str = Field()
        

tools.append(
    Tool.from_function(
        func=llm_math_chain.run,
        name="Calculator",
        description="useful for when you need to answer questions about math",
        args_schema=CalculatorInput
        # coroutine= ... <- you can specify an async method if desired as well
    )
)
# Construct the agent. We will use the default agent type here.
# See documentation for a full list of options.
agent = initialize_agent(tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True)
agent.run("Who is Leo DiCaprio's girlfriend? What is her current age raised to the 0.43 power?")

日志输出：

Entering new AgentExecutor chain...
I need to find out Leo DiCaprio's girlfriend's name and her age
Action: Search
Action Input: "Leo DiCaprio girlfriend"
Observation: After rumours of a romance with Gigi Hadid, the Oscar winner has seemingly moved on. First being linked to the television personality in September 2022, it appears as if his "age bracket" has moved up. This follows his rumoured relationship with mere 19-year-old Eden Polani.
Thought:I still need to find out his current girlfriend's name and age
Action: Search
Action Input: "Leo DiCaprio current girlfriend"
Observation: Just Jared on Instagram: “Leonardo DiCaprio & girlfriend Camila Morrone couple up for a lunch date!
Thought:Now that I know his girlfriend's name is Camila Morrone, I need to find her current age
Action: Search
Action Input: "Camila Morrone age"
Observation: 25 years
Thought:Now that I have her age, I need to calculate her age raised to the 0.43 power
Action: Calculator
Action Input: 25^(0.43)

Entering new LLMMathChain chain...
25^(0.43)```text
25**(0.43)```

...numexpr.evaluate("25**(0.43)")...

Answer: 3.991298452658078
Finished chain.

Observation: Answer: 3.991298452658078
Thought:I now know the final answer
Final Answer: Camila Morrone's current age raised to the 0.43 power is approximately 3.99.

Finished chain.

输出：

"Camila Morrone's current age raised to the 0.43 power is approximately 3.99."

继承BaseTool类

我们还可以直接继承BaseTool类。这在我们希望对实例变量有更多控制或者希望将回调传递给嵌套链或其他工具时非常有用。

from typing import Optional, Type

from langchain.callbacks.manager import AsyncCallbackManagerForToolRun, CallbackManagerForToolRun

class CustomSearchTool(BaseTool):
    name = "custom_search"
    description = "useful for when you need to answer questions about current events"

    def _run(self, query: str, run_manager: Optional[CallbackManagerForToolRun] = None) -> str:
        """Use the tool."""
        return search.run(query)
    
    async def _arun(self, query: str, run_manager: Optional[AsyncCallbackManagerForToolRun] = None) -> str:
        """Use the tool asynchronously."""
        raise NotImplementedError("custom_search does not support async")
    
class CustomCalculatorTool(BaseTool):
    name = "Calculator"
    description = "useful for when you need to answer questions about math"
    args_schema: Type[BaseModel] = CalculatorInput

    def _run(self, query: str, run_manager: Optional[CallbackManagerForToolRun] = None) -> str:
        """Use the tool."""
        return llm_math_chain.run(query)
    
    async def _arun(self, query: str,  run_manager: Optional[AsyncCallbackManagerForToolRun] = None) -> str:
        """Use the tool asynchronously."""
        raise NotImplementedError("Calculator does not support async")
tools = [CustomSearchTool(), CustomCalculatorTool()]
agent = initialize_agent(tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True)
agent.run("Who is Leo DiCaprio's girlfriend? What is her current age raised to the 0.43 power?")

日志输出：

Entering new AgentExecutor chain...
I need to use custom_search to find out who Leo DiCaprio's girlfriend is, and then use the Calculator to raise her age to the 0.43 power.
Action: custom_search
Action Input: "Leo DiCaprio girlfriend"
Observation: After rumours of a romance with Gigi Hadid, the Oscar winner has seemingly moved on. First being linked to the television personality in September 2022, it appears as if his "age bracket" has moved up. This follows his rumoured relationship with mere 19-year-old Eden Polani.
Thought:I need to find out the current age of Eden Polani.
Action: custom_search
Action Input: "Eden Polani age"
Observation: 19 years old
Thought:Now I can use the Calculator to raise her age to the 0.43 power.
Action: Calculator
Action Input: 19 ^ 0.43

Entering new LLMMathChain chain...
19 ^ 0.43```text
19 ** 0.43```

...numexpr.evaluate("19 ** 0.43")...

Answer: 3.547023357958959
Finished chain.

Observation: Answer: 3.547023357958959
Thought:I now know the final answer.
Final Answer: 3.547023357958959

Finished chain.

输出：

'3.547023357958959'

使用tool装饰器

为了更容易定义自定义工具，提供了@tool装饰器。该装饰器可用于快速从简单函数创建一个Tool。默认情况下，装饰器使用函数名作为工具名，但可以通过将字符串作为第一个参数传递来覆盖此行为。此外，装饰器将使用函数的文档字符串作为工具的描述。

from langchain.tools import tool

@tool
def search_api(query: str) -> str:
    """Searches the API for the query."""
    return f"Results for query {query}"

search_api

我们还可以提供工具名和是否直接返回等参数。

@tool("search", return_direct=True)
def search_api(query: str) -> str:
    """Searches the API for the query."""
    return "Results"
search_api
Tool(name='search', description='search(query: str) -> str - Searches the API for the query.', args_schema=<class 'pydantic.main.SearchApi'>, return_direct=True, verbose=False, callback_manager=<langchain.callbacks.shared.SharedCallbackManager object at 0x12748c4c0>, func=<function search_api at 0x16bd66310>, coroutine=None)

我们还可以提供args_schema来提供有关参数的更多信息。

class SearchInput(BaseModel):
    query: str = Field(description="should be a search query")
        
@tool("search", return_direct=True, args_schema=SearchInput)
def search_api(query: str) -> str:
    """Searches the API for the query."""
    return "Results"
search_api
Tool(name='search', description='search(query: str) -> str - Searches the API for the query.', args_schema=<class '__main__.SearchInput'>, return_direct=True, verbose=False, callback_manager=<langchain.callbacks.shared.SharedCallbackManager object at 0x12748c4c0>, func=<function search_api at 0x16bcf0ee0>, coroutine=None)

自定义结构化工具

如果我们的函数需要更结构化的参数，我们可以直接使用StructuredTool类，或者仍然继承BaseTool类。

StructuredTool dataclass

要从给定的函数动态生成结构化工具，最快速的方法是使用StructuredTool.from_function()。

import requests
from langchain.tools import StructuredTool

def post_message(url: str, body: dict, parameters: Optional[dict] = None) -> str:
    """Sends a POST request to the given url with the given body and parameters."""
    result = requests.post(url, json=body, params=parameters)
    return f"Status: {result.status_code} - {result.text}"

tool = StructuredTool.from_function(post_message)

继承BaseTool类

BaseTool类会自动从_run方法的签名中推断出模式schema。

from typing import Optional, Type

from langchain.callbacks.manager import AsyncCallbackManagerForToolRun, CallbackManagerForToolRun
            
class CustomSearchTool(BaseTool):
    name = "custom_search"
    description = "useful for when you need to answer questions about current events"

    def _run(self, query: str, engine: str = "google", gl: str = "us", hl: str = "en", run_manager: Optional[CallbackManagerForToolRun] = None) -> str:
        """Use the tool."""
        search_wrapper = SerpAPIWrapper(params={"engine": engine, "gl": gl, "hl": hl})
        return search_wrapper.run(query)
    
    async def _arun(self, query: str,  engine: str = "google", gl: str = "us", hl: str = "en", run_manager: Optional[AsyncCallbackManagerForToolRun] = None) -> str:
        """Use the tool asynchronously."""
        raise NotImplementedError("custom_search does not support async")



# You can provide a custom args schema to add descriptions or custom validation

class SearchSchema(BaseModel):
    query: str = Field(description="should be a search query")
    engine: str = Field(description="should be a search engine")
    gl: str = Field(description="should be a country code")
    hl: str = Field(description="should be a language code")

class CustomSearchTool(BaseTool):
    name = "custom_search"
    description = "useful for when you need to answer questions about current events"
    args_schema: Type[SearchSchema] = SearchSchema

    def _run(self, query: str, engine: str = "google", gl: str = "us", hl: str = "en", run_manager: Optional[CallbackManagerForToolRun] = None) -> str:
        """Use the tool."""
        search_wrapper = SerpAPIWrapper(params={"engine": engine, "gl": gl, "hl": hl})
        return search_wrapper.run(query)
    
    async def _arun(self, query: str,  engine: str = "google", gl: str = "us", hl: str = "en", run_manager: Optional[AsyncCallbackManagerForToolRun] = None) -> str:
        """Use the tool asynchronously."""
        raise NotImplementedError("custom_search does not support async")
    

使用装饰器

tool装饰器会自动创建一个结构化工具，如果函数的签名具有多个参数。

import requests
from langchain.tools import tool

@tool
def post_message(url: str, body: dict, parameters: Optional[dict] = None) -> str:
    """Sends a POST request to the given url with the given body and parameters."""
    result = requests.post(url, json=body, params=parameters)
    return f"Status: {result.status_code} - {result.text}"

修改现有工具

现在，我们展示如何加载现有工具并直接修改它们。在下面的示例中，我们做一些非常简单的操作，将Search工具的名称更改为 Google Search。

from langchain.agents import load_tools
tools = load_tools(["serpapi", "llm-math"], llm=llm)
tools[0].name = "Google Search"
agent = initialize_agent(tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True)
agent.run("Who is Leo DiCaprio's girlfriend? What is her current age raised to the 0.43 power?")

日志输出：

Entering new AgentExecutor chain...
I need to find out Leo DiCaprio's girlfriend's name and her age.
Action: Google Search
Action Input: "Leo DiCaprio girlfriend"
Observation: After rumours of a romance with Gigi Hadid, the Oscar winner has seemingly moved on. First being linked to the television personality in September 2022, it appears as if his "age bracket" has moved up. This follows his rumoured relationship with mere 19-year-old Eden Polani.
Thought:I still need to find out his current girlfriend's name and her age.
Action: Google Search
Action Input: "Leo DiCaprio current girlfriend age"
Observation: Leonardo DiCaprio has been linked with 19-year-old model Eden Polani, continuing the rumour that he doesn't date any women over the age of ...
Thought:I need to find out the age of Eden Polani.
Action: Calculator
Action Input: 19^(0.43)
Observation: Answer: 3.547023357958959
Thought:I now know the final answer.
Final Answer: The age of Leo DiCaprio's girlfriend raised to the 0.43 power is approximately 3.55.

Finished chain.

输出：

"The age of Leo DiCaprio's girlfriend raised to the 0.43 power is approximately 3.55."

定义工具之间的优先级

当我们创建自定义工具时，我们可能希望Agent更多地使用自定义工具而不是普通工具。例如，我们创建了一个自定义工具，用于从数据库中获取有关音乐的信息。当用户需要有关歌曲的信息时，我们希望Agent更多地使用自定义工具而不是普通的搜索工具。但是Agent可能会优先考虑普通的搜索工具。为了实现这一目标，可以在描述中添加类似于如果问题是关于音乐的，例如“昨天的歌手是谁？”或“2022年最受欢迎的歌曲是什么？”的句子，优先使用这个工具而不是普通搜索工具的语句，下面是一个示例：

# Import things that are needed generically
from langchain.agents import initialize_agent, Tool
from langchain.agents import AgentType
from langchain.llms import OpenAI
from langchain import LLMMathChain, SerpAPIWrapper
search = SerpAPIWrapper()
tools = [
    Tool(
        name = "Search",
        func=search.run,
        description="useful for when you need to answer questions about current events"
    ),
    Tool(
        name="Music Search",
        func=lambda x: "'All I Want For Christmas Is You' by Mariah Carey.", #Mock Function
        description="A Music search engine. Use this more than the normal search if the question is about Music, like 'who is the singer of yesterday?' or 'what is the most popular song in 2022?'",
    )
]

agent = initialize_agent(tools, OpenAI(temperature=0), agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True)
agent.run("what is the most famous song of christmas")

日志输出：

Entering new AgentExecutor chain...
 I should use a music search engine to find the answer
Action: Music Search
Action Input: most famous song of christmas'All I Want For Christmas Is You' by Mariah Carey. I now know the final answer
Final Answer: 'All I Want For Christmas Is You' by Mariah Carey.

Finished chain.

输出：

"'All I Want For Christmas Is You' by Mariah Carey."

使用直接返回的工具

通常情况下，如果一个工具被调用，将其输出直接返回给用户可能是很有意义的。我们可以通过将工具的return_direct标志设置为True来轻松实现这一点。

llm_math_chain = LLMMathChain(llm=llm)
tools = [
    Tool(
        name="Calculator",
        func=llm_math_chain.run,
        description="useful for when you need to answer questions about math",
        return_direct=True
    )
]
llm = OpenAI(temperature=0)
agent = initialize_agent(tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True)
agent.run("whats 2**.12")

日志输出：

Entering new AgentExecutor chain...
 I need to calculate this
Action: Calculator
Action Input: 2**.12Answer: 1.086734862526058

Finished chain.

输出：

'Answer: 1.086734862526058'

处理工具错误

当工具遇到错误且异常未被捕获时，Agent将停止执行。如果我们希望Agent继续执行，可以引发ToolException并相应地设置handle_tool_error。当抛出ToolException时，Agent将不会停止工作，而是根据工具的handle_tool_error变量处理异常，并将处理结果作为观察返回给Agent，并以红色打印出来。我们可以将handle_tool_error设置为True，设置为统一的字符串值，或者设置为函数。如果设置为函数，该函数应该以ToolException作为参数，并返回一个str。需要注意的是，仅仅引发ToolException是不会生效的。我们需要先设置工具的handle_tool_error，因为它的默认值是False。

from langchain.schema import ToolException

from langchain import SerpAPIWrapper
from langchain.agents import AgentType, initialize_agent
from langchain.chat_models import ChatOpenAI
from langchain.tools import Tool

from langchain.chat_models import ChatOpenAI

def _handle_error(error:ToolException) -> str:
    return  "The following errors occurred during tool execution:" + error.args[0]+ "Please try another tool."
def search_tool1(s: str):raise ToolException("The search tool1 is not available.")
def search_tool2(s: str):raise ToolException("The search tool2 is not available.")
search_tool3 = SerpAPIWrapper()
description="useful for when you need to answer questions about current events.You should give priority to using it."
tools = [
    Tool.from_function(
        func=search_tool1,
        name="Search_tool1",
        description=description,
        handle_tool_error=True,
    ),
    Tool.from_function(
        func=search_tool2,
        name="Search_tool2",
        description=description,
        handle_tool_error=_handle_error,
    ),
    Tool.from_function(
        func=search_tool3.run,
        name="Search_tool3",
        description="useful for when you need to answer questions about current events",
    ),
]

agent = initialize_agent(
    tools,
    ChatOpenAI(temperature=0),
    agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
    verbose=True,
)

agent.run("Who is Leo DiCaprio's girlfriend?")

日志输出：

Entering new AgentExecutor chain...
I should use Search_tool1 to find recent news articles about Leo DiCaprio's personal life.
Action: Search_tool1
Action Input: "Leo DiCaprio girlfriend"
Observation: The search tool1 is not available.
Thought:I should try using Search_tool2 instead.
Action: Search_tool2
Action Input: "Leo DiCaprio girlfriend"
Observation: The following errors occurred during tool execution:The search tool2 is not available.Please try another tool.
Thought:I should try using Search_tool3 as a last resort.
Action: Search_tool3
Action Input: "Leo DiCaprio girlfriend"
Observation: Leonardo DiCaprio and Gigi Hadid were recently spotted at a pre-Oscars party, sparking interest once again in their rumored romance. The Revenant actor and the model first made headlines when they were spotted together at a New York Fashion Week afterparty in September 2022.
Thought:Based on the information from Search_tool3, it seems that Gigi Hadid is currently rumored to be Leo DiCaprio's girlfriend.
Final Answer: Gigi Hadid is currently rumored to be Leo DiCaprio's girlfriend.

Finished chain.

输出：

"Gigi Hadid is currently rumored to be Leo DiCaprio's girlfriend."

参考文献：
[1] LangChain官方网站：https://www.langchain.com/
[2] LangChain ️ 中文网，跟着LangChain一起学LLM/GPT开发：https://www.langchain.com.cn/
[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://www.cnlangchain.com/