Part 3: Building Powerful Chains and Agents in LangChain

0368

LangChian (4 Part Series)

1 Mastering LangChain: Part 1 – Introduction to LangChain and Its Key Components
2 Part 2: Mastering Prompts and Language Models with LangChain
3 Part 3: Building Powerful Chains and Agents in LangChain
4 LangChain Part 4 – Leveraging Memory and Storage in LangChain: A Comprehensive Guide

Building Powerful Chains and Agents in LangChain

In this comprehensive guide, we’ll dive deep into the world of LangChain, focusing on constructing powerful chains and agents. We’ll cover everything from understanding the fundamentals of chains to combining them with large language models (LLMs) and introducing sophisticated agents for autonomous decision-making.

1. Understanding Chains

1.1 What are Chains in LangChain?

Chains in LangChain are sequences of operations or tasks that process data in a specific order. They allow for modular and reusable workflows, making it easier to handle complex data processing and language tasks. Chains are the building blocks for creating sophisticated AI-driven systems.

1.2 Types of Chains

LangChain offers several types of chains, each suited for different scenarios:

  1. Sequential Chains: These chains process data in a linear order, where the output of one step serves as the input for the next. They’re ideal for straightforward, step-by-step processes.

  2. Map/Reduce Chains: These chains involve mapping a function over a set of data and then reducing the results to a single output. They’re great for parallel processing of large datasets.

  3. Router Chains: These chains direct inputs to different sub-chains based on certain conditions, allowing for more complex, branching workflows.

1.3 Creating Custom Chains

Creating custom chains involves defining specific operations or functions that will be part of the chain. Here’s an example of a custom sequential chain: 

<span>from</span> <span>langchain.chains</span> <span>import</span> <span>LLMChain</span>
<span>from</span> <span>langchain.llms</span> <span>import</span> <span>OpenAI</span>
<span>from</span> <span>langchain.prompts</span> <span>import</span> <span>PromptTemplate</span>
<span>class</span> <span>CustomChain</span><span>:</span>
    <span>def</span> <span>__init__</span><span>(</span><span>self</span><span>,</span> <span>llm</span><span>):</span>
        <span>self</span><span>.</span><span>llm</span> <span>=</span> <span>llm</span>
        <span>self</span><span>.</span><span>steps</span> <span>=</span> <span>[]</span>
    <span>def</span> <span>add_step</span><span>(</span><span>self</span><span>,</span> <span>prompt_template</span><span>):</span>
        <span>prompt</span> <span>=</span> <span>PromptTemplate</span><span>(</span><span>template</span><span>=</span><span>prompt_template</span><span>,</span> <span>input_variables</span><span>=</span><span>[</span><span>"</span><span>input</span><span>"</span><span>])</span>
        <span>chain</span> <span>=</span> <span>LLMChain</span><span>(</span><span>llm</span><span>=</span><span>self</span><span>.</span><span>llm</span><span>,</span> <span>prompt</span><span>=</span><span>prompt</span><span>)</span>
        <span>self</span><span>.</span><span>steps</span><span>.</span><span>append</span><span>(</span><span>chain</span><span>)</span>
    <span>def</span> <span>execute</span><span>(</span><span>self</span><span>,</span> <span>input_text</span><span>):</span>
        <span>for</span> <span>step</span> <span>in</span> <span>self</span><span>.</span><span>steps</span><span>:</span>
            <span>input_text</span> <span>=</span> <span>step</span><span>.</span><span>run</span><span>(</span><span>input_text</span><span>)</span>
        <span>return</span> <span>input_text</span>
<span># Initialize the chain </span><span>llm</span> <span>=</span> <span>OpenAI</span><span>(</span><span>temperature</span><span>=</span><span>0.7</span><span>)</span>
<span>chain</span> <span>=</span> <span>CustomChain</span><span>(</span><span>llm</span><span>)</span>
<span># Add steps to the chain </span><span>chain</span><span>.</span><span>add_step</span><span>(</span><span>"</span><span>Summarize the following text in one sentence: {input}</span><span>"</span><span>)</span>
<span>chain</span><span>.</span><span>add_step</span><span>(</span><span>"</span><span>Translate the following English text to French: {input}</span><span>"</span><span>)</span>
<span># Execute the chain </span><span>result</span> <span>=</span> <span>chain</span><span>.</span><span>execute</span><span>(</span><span>"</span><span>LangChain is a powerful framework for building AI applications.</span><span>"</span><span>)</span>
<span>print</span><span>(</span><span>result</span><span>)</span>
<span>from</span> <span>langchain.chains</span> <span>import</span> <span>LLMChain</span>
<span>from</span> <span>langchain.llms</span> <span>import</span> <span>OpenAI</span>
<span>from</span> <span>langchain.prompts</span> <span>import</span> <span>PromptTemplate</span>

<span>class</span> <span>CustomChain</span><span>:</span>
    <span>def</span> <span>__init__</span><span>(</span><span>self</span><span>,</span> <span>llm</span><span>):</span>
        <span>self</span><span>.</span><span>llm</span> <span>=</span> <span>llm</span>
        <span>self</span><span>.</span><span>steps</span> <span>=</span> <span>[]</span>

    <span>def</span> <span>add_step</span><span>(</span><span>self</span><span>,</span> <span>prompt_template</span><span>):</span>
        <span>prompt</span> <span>=</span> <span>PromptTemplate</span><span>(</span><span>template</span><span>=</span><span>prompt_template</span><span>,</span> <span>input_variables</span><span>=</span><span>[</span><span>"</span><span>input</span><span>"</span><span>])</span>
        <span>chain</span> <span>=</span> <span>LLMChain</span><span>(</span><span>llm</span><span>=</span><span>self</span><span>.</span><span>llm</span><span>,</span> <span>prompt</span><span>=</span><span>prompt</span><span>)</span>
        <span>self</span><span>.</span><span>steps</span><span>.</span><span>append</span><span>(</span><span>chain</span><span>)</span>

    <span>def</span> <span>execute</span><span>(</span><span>self</span><span>,</span> <span>input_text</span><span>):</span>
        <span>for</span> <span>step</span> <span>in</span> <span>self</span><span>.</span><span>steps</span><span>:</span>
            <span>input_text</span> <span>=</span> <span>step</span><span>.</span><span>run</span><span>(</span><span>input_text</span><span>)</span>
        <span>return</span> <span>input_text</span>

<span># Initialize the chain </span><span>llm</span> <span>=</span> <span>OpenAI</span><span>(</span><span>temperature</span><span>=</span><span>0.7</span><span>)</span>
<span>chain</span> <span>=</span> <span>CustomChain</span><span>(</span><span>llm</span><span>)</span>

<span># Add steps to the chain </span><span>chain</span><span>.</span><span>add_step</span><span>(</span><span>"</span><span>Summarize the following text in one sentence: {input}</span><span>"</span><span>)</span>
<span>chain</span><span>.</span><span>add_step</span><span>(</span><span>"</span><span>Translate the following English text to French: {input}</span><span>"</span><span>)</span>

<span># Execute the chain </span><span>result</span> <span>=</span> <span>chain</span><span>.</span><span>execute</span><span>(</span><span>"</span><span>LangChain is a powerful framework for building AI applications.</span><span>"</span><span>)</span>
<span>print</span><span>(</span><span>result</span><span>)</span>
from langchain.chains import LLMChain
from langchain.llms import OpenAI
from langchain.prompts import PromptTemplate

class CustomChain:
    def __init__(self, llm):
        self.llm = llm
        self.steps = []

    def add_step(self, prompt_template):
        prompt = PromptTemplate(template=prompt_template, input_variables=["input"])
        chain = LLMChain(llm=self.llm, prompt=prompt)
        self.steps.append(chain)

    def execute(self, input_text):
        for step in self.steps:
            input_text = step.run(input_text)
        return input_text

# Initialize the chain llm = OpenAI(temperature=0.7)
chain = CustomChain(llm)

# Add steps to the chain chain.add_step("Summarize the following text in one sentence: {input}")
chain.add_step("Translate the following English text to French: {input}")

# Execute the chain result = chain.execute("LangChain is a powerful framework for building AI applications.")
print(result)

Enter fullscreen mode Exit fullscreen mode

This example creates a custom chain that first summarizes an input text and then translates it to French.

2. Combining Chains and LLMs

2.1 Integrating Chains with Prompts and LLMs

Chains can be seamlessly integrated with prompts and LLMs to create more powerful and flexible systems. Here’s an example: 

<span>from</span> <span>langchain</span> <span>import</span> <span>PromptTemplate</span><span>,</span> <span>LLMChain</span>
<span>from</span> <span>langchain.llms</span> <span>import</span> <span>OpenAI</span>
<span>from</span> <span>langchain.chains</span> <span>import</span> <span>SimpleSequentialChain</span>
<span>llm</span> <span>=</span> <span>OpenAI</span><span>(</span><span>temperature</span><span>=</span><span>0.7</span><span>)</span>
<span># First chain: Generate a topic </span><span>first_prompt</span> <span>=</span> <span>PromptTemplate</span><span>(</span>
    <span>input_variables</span><span>=</span><span>[</span><span>"</span><span>subject</span><span>"</span><span>],</span>
    <span>template</span><span>=</span><span>"</span><span>Generate a random {subject} topic:</span><span>"</span>
<span>)</span>
<span>first_chain</span> <span>=</span> <span>LLMChain</span><span>(</span><span>llm</span><span>=</span><span>llm</span><span>,</span> <span>prompt</span><span>=</span><span>first_prompt</span><span>)</span>
<span># Second chain: Write a paragraph about the topic </span><span>second_prompt</span> <span>=</span> <span>PromptTemplate</span><span>(</span>
    <span>input_variables</span><span>=</span><span>[</span><span>"</span><span>topic</span><span>"</span><span>],</span>
    <span>template</span><span>=</span><span>"</span><span>Write a short paragraph about {topic}:</span><span>"</span>
<span>)</span>
<span>second_chain</span> <span>=</span> <span>LLMChain</span><span>(</span><span>llm</span><span>=</span><span>llm</span><span>,</span> <span>prompt</span><span>=</span><span>second_prompt</span><span>)</span>
<span># Combine the chains </span><span>overall_chain</span> <span>=</span> <span>SimpleSequentialChain</span><span>(</span><span>chains</span><span>=</span><span>[</span><span>first_chain</span><span>,</span> <span>second_chain</span><span>],</span> <span>verbose</span><span>=</span><span>True</span><span>)</span>
<span># Run the chain </span><span>result</span> <span>=</span> <span>overall_chain</span><span>.</span><span>run</span><span>(</span><span>"</span><span>science</span><span>"</span><span>)</span>
<span>print</span><span>(</span><span>result</span><span>)</span>
<span>from</span> <span>langchain</span> <span>import</span> <span>PromptTemplate</span><span>,</span> <span>LLMChain</span>
<span>from</span> <span>langchain.llms</span> <span>import</span> <span>OpenAI</span>
<span>from</span> <span>langchain.chains</span> <span>import</span> <span>SimpleSequentialChain</span>

<span>llm</span> <span>=</span> <span>OpenAI</span><span>(</span><span>temperature</span><span>=</span><span>0.7</span><span>)</span>

<span># First chain: Generate a topic </span><span>first_prompt</span> <span>=</span> <span>PromptTemplate</span><span>(</span>
    <span>input_variables</span><span>=</span><span>[</span><span>"</span><span>subject</span><span>"</span><span>],</span>
    <span>template</span><span>=</span><span>"</span><span>Generate a random {subject} topic:</span><span>"</span>
<span>)</span>
<span>first_chain</span> <span>=</span> <span>LLMChain</span><span>(</span><span>llm</span><span>=</span><span>llm</span><span>,</span> <span>prompt</span><span>=</span><span>first_prompt</span><span>)</span>

<span># Second chain: Write a paragraph about the topic </span><span>second_prompt</span> <span>=</span> <span>PromptTemplate</span><span>(</span>
    <span>input_variables</span><span>=</span><span>[</span><span>"</span><span>topic</span><span>"</span><span>],</span>
    <span>template</span><span>=</span><span>"</span><span>Write a short paragraph about {topic}:</span><span>"</span>
<span>)</span>
<span>second_chain</span> <span>=</span> <span>LLMChain</span><span>(</span><span>llm</span><span>=</span><span>llm</span><span>,</span> <span>prompt</span><span>=</span><span>second_prompt</span><span>)</span>

<span># Combine the chains </span><span>overall_chain</span> <span>=</span> <span>SimpleSequentialChain</span><span>(</span><span>chains</span><span>=</span><span>[</span><span>first_chain</span><span>,</span> <span>second_chain</span><span>],</span> <span>verbose</span><span>=</span><span>True</span><span>)</span>

<span># Run the chain </span><span>result</span> <span>=</span> <span>overall_chain</span><span>.</span><span>run</span><span>(</span><span>"</span><span>science</span><span>"</span><span>)</span>
<span>print</span><span>(</span><span>result</span><span>)</span>
from langchain import PromptTemplate, LLMChain
from langchain.llms import OpenAI
from langchain.chains import SimpleSequentialChain

llm = OpenAI(temperature=0.7)

# First chain: Generate a topic first_prompt = PromptTemplate(
    input_variables=["subject"],
    template="Generate a random {subject} topic:"
)
first_chain = LLMChain(llm=llm, prompt=first_prompt)

# Second chain: Write a paragraph about the topic second_prompt = PromptTemplate(
    input_variables=["topic"],
    template="Write a short paragraph about {topic}:"
)
second_chain = LLMChain(llm=llm, prompt=second_prompt)

# Combine the chains overall_chain = SimpleSequentialChain(chains=[first_chain, second_chain], verbose=True)

# Run the chain result = overall_chain.run("science")
print(result)

Enter fullscreen mode Exit fullscreen mode

This example creates a chain that generates a random science topic and then writes a paragraph about it.

2.2 Debugging and Optimizing Chain-LLM Interactions

To debug and optimize chain-LLM interactions, you can use the verbose parameter and custom callbacks: 

<span>from</span> <span>langchain.callbacks</span> <span>import</span> <span>StdOutCallbackHandler</span>
<span>from</span> <span>langchain.chains</span> <span>import</span> <span>LLMChain</span>
<span>from</span> <span>langchain.llms</span> <span>import</span> <span>OpenAI</span>
<span>from</span> <span>langchain.prompts</span> <span>import</span> <span>PromptTemplate</span>
<span>class</span> <span>CustomHandler</span><span>(</span><span>StdOutCallbackHandler</span><span>):</span>
    <span>def</span> <span>on_llm_start</span><span>(</span><span>self</span><span>,</span> <span>serialized</span><span>,</span> <span>prompts</span><span>,</span> <span>**</span><span>kwargs</span><span>):</span>
        <span>print</span><span>(</span><span>f</span><span>"</span><span>LLM started with prompt: </span><span>{</span><span>prompts</span><span>[</span><span>0</span><span>]</span><span>}</span><span>"</span><span>)</span>
    <span>def</span> <span>on_llm_end</span><span>(</span><span>self</span><span>,</span> <span>response</span><span>,</span> <span>**</span><span>kwargs</span><span>):</span>
        <span>print</span><span>(</span><span>f</span><span>"</span><span>LLM finished with response: </span><span>{</span><span>response</span><span>.</span><span>generations</span><span>[</span><span>0</span><span>][</span><span>0</span><span>].</span><span>text</span><span>}</span><span>"</span><span>)</span>
<span>llm</span> <span>=</span> <span>OpenAI</span><span>(</span><span>temperature</span><span>=</span><span>0.7</span><span>,</span> <span>callbacks</span><span>=</span><span>[</span><span>CustomHandler</span><span>()])</span>
<span>template</span> <span>=</span> <span>"</span><span>Tell me a {adjective} joke about {subject}.</span><span>"</span>
<span>prompt</span> <span>=</span> <span>PromptTemplate</span><span>(</span><span>input_variables</span><span>=</span><span>[</span><span>"</span><span>adjective</span><span>"</span><span>,</span> <span>"</span><span>subject</span><span>"</span><span>],</span> <span>template</span><span>=</span><span>template</span><span>)</span>
<span>chain</span> <span>=</span> <span>LLMChain</span><span>(</span><span>llm</span><span>=</span><span>llm</span><span>,</span> <span>prompt</span><span>=</span><span>prompt</span><span>,</span> <span>verbose</span><span>=</span><span>True</span><span>)</span>
<span>result</span> <span>=</span> <span>chain</span><span>.</span><span>run</span><span>(</span><span>adjective</span><span>=</span><span>"</span><span>funny</span><span>"</span><span>,</span> <span>subject</span><span>=</span><span>"</span><span>programming</span><span>"</span><span>)</span>
<span>print</span><span>(</span><span>result</span><span>)</span>
<span>from</span> <span>langchain.callbacks</span> <span>import</span> <span>StdOutCallbackHandler</span>
<span>from</span> <span>langchain.chains</span> <span>import</span> <span>LLMChain</span>
<span>from</span> <span>langchain.llms</span> <span>import</span> <span>OpenAI</span>
<span>from</span> <span>langchain.prompts</span> <span>import</span> <span>PromptTemplate</span>

<span>class</span> <span>CustomHandler</span><span>(</span><span>StdOutCallbackHandler</span><span>):</span>
    <span>def</span> <span>on_llm_start</span><span>(</span><span>self</span><span>,</span> <span>serialized</span><span>,</span> <span>prompts</span><span>,</span> <span>**</span><span>kwargs</span><span>):</span>
        <span>print</span><span>(</span><span>f</span><span>"</span><span>LLM started with prompt: </span><span>{</span><span>prompts</span><span>[</span><span>0</span><span>]</span><span>}</span><span>"</span><span>)</span>

    <span>def</span> <span>on_llm_end</span><span>(</span><span>self</span><span>,</span> <span>response</span><span>,</span> <span>**</span><span>kwargs</span><span>):</span>
        <span>print</span><span>(</span><span>f</span><span>"</span><span>LLM finished with response: </span><span>{</span><span>response</span><span>.</span><span>generations</span><span>[</span><span>0</span><span>][</span><span>0</span><span>].</span><span>text</span><span>}</span><span>"</span><span>)</span>

<span>llm</span> <span>=</span> <span>OpenAI</span><span>(</span><span>temperature</span><span>=</span><span>0.7</span><span>,</span> <span>callbacks</span><span>=</span><span>[</span><span>CustomHandler</span><span>()])</span>
<span>template</span> <span>=</span> <span>"</span><span>Tell me a {adjective} joke about {subject}.</span><span>"</span>
<span>prompt</span> <span>=</span> <span>PromptTemplate</span><span>(</span><span>input_variables</span><span>=</span><span>[</span><span>"</span><span>adjective</span><span>"</span><span>,</span> <span>"</span><span>subject</span><span>"</span><span>],</span> <span>template</span><span>=</span><span>template</span><span>)</span>
<span>chain</span> <span>=</span> <span>LLMChain</span><span>(</span><span>llm</span><span>=</span><span>llm</span><span>,</span> <span>prompt</span><span>=</span><span>prompt</span><span>,</span> <span>verbose</span><span>=</span><span>True</span><span>)</span>

<span>result</span> <span>=</span> <span>chain</span><span>.</span><span>run</span><span>(</span><span>adjective</span><span>=</span><span>"</span><span>funny</span><span>"</span><span>,</span> <span>subject</span><span>=</span><span>"</span><span>programming</span><span>"</span><span>)</span>
<span>print</span><span>(</span><span>result</span><span>)</span>
from langchain.callbacks import StdOutCallbackHandler
from langchain.chains import LLMChain
from langchain.llms import OpenAI
from langchain.prompts import PromptTemplate

class CustomHandler(StdOutCallbackHandler):
    def on_llm_start(self, serialized, prompts, **kwargs):
        print(f"LLM started with prompt: {prompts[0]}")

    def on_llm_end(self, response, **kwargs):
        print(f"LLM finished with response: {response.generations[0][0].text}")

llm = OpenAI(temperature=0.7, callbacks=[CustomHandler()])
template = "Tell me a {adjective} joke about {subject}."
prompt = PromptTemplate(input_variables=["adjective", "subject"], template=template)
chain = LLMChain(llm=llm, prompt=prompt, verbose=True)

result = chain.run(adjective="funny", subject="programming")
print(result)

Enter fullscreen mode Exit fullscreen mode

This example uses a custom callback handler to provide detailed information about the LLM’s input and output.

3. Introducing Agents

3.1 What are Agents in LangChain?

Agents in LangChain are autonomous entities that can use tools and make decisions to accomplish tasks. They combine LLMs with external tools to solve complex problems, allowing for more dynamic and adaptable AI systems.

3.2 Built-in Agents and Their Capabilities

LangChain provides several built-in agents, such as the zero-shot-react-description agent: 

<span>from</span> <span>langchain.agents</span> <span>import</span> <span>load_tools</span><span>,</span> <span>initialize_agent</span><span>,</span> <span>AgentType</span>
<span>from</span> <span>langchain.llms</span> <span>import</span> <span>OpenAI</span>
<span>llm</span> <span>=</span> <span>OpenAI</span><span>(</span><span>temperature</span><span>=</span><span>0</span><span>)</span>
<span>tools</span> <span>=</span> <span>load_tools</span><span>([</span><span>"</span><span>wikipedia</span><span>"</span><span>,</span> <span>"</span><span>llm-math</span><span>"</span><span>],</span> <span>llm</span><span>=</span><span>llm</span><span>)</span>
<span>agent</span> <span>=</span> <span>initialize_agent</span><span>(</span>
    <span>tools</span><span>,</span> 
    <span>llm</span><span>,</span> 
    <span>agent</span><span>=</span><span>AgentType</span><span>.</span><span>ZERO_SHOT_REACT_DESCRIPTION</span><span>,</span>
    <span>verbose</span><span>=</span><span>True</span>
<span>)</span>
<span>result</span> <span>=</span> <span>agent</span><span>.</span><span>run</span><span>(</span><span>"</span><span>What is the square root of the year Plato was born?</span><span>"</span><span>)</span>
<span>print</span><span>(</span><span>result</span><span>)</span>
<span>from</span> <span>langchain.agents</span> <span>import</span> <span>load_tools</span><span>,</span> <span>initialize_agent</span><span>,</span> <span>AgentType</span>
<span>from</span> <span>langchain.llms</span> <span>import</span> <span>OpenAI</span>

<span>llm</span> <span>=</span> <span>OpenAI</span><span>(</span><span>temperature</span><span>=</span><span>0</span><span>)</span>
<span>tools</span> <span>=</span> <span>load_tools</span><span>([</span><span>"</span><span>wikipedia</span><span>"</span><span>,</span> <span>"</span><span>llm-math</span><span>"</span><span>],</span> <span>llm</span><span>=</span><span>llm</span><span>)</span>

<span>agent</span> <span>=</span> <span>initialize_agent</span><span>(</span>
    <span>tools</span><span>,</span> 
    <span>llm</span><span>,</span> 
    <span>agent</span><span>=</span><span>AgentType</span><span>.</span><span>ZERO_SHOT_REACT_DESCRIPTION</span><span>,</span>
    <span>verbose</span><span>=</span><span>True</span>
<span>)</span>

<span>result</span> <span>=</span> <span>agent</span><span>.</span><span>run</span><span>(</span><span>"</span><span>What is the square root of the year Plato was born?</span><span>"</span><span>)</span>
<span>print</span><span>(</span><span>result</span><span>)</span>
from langchain.agents import load_tools, initialize_agent, AgentType
from langchain.llms import OpenAI

llm = OpenAI(temperature=0)
tools = load_tools(["wikipedia", "llm-math"], llm=llm)

agent = initialize_agent(
    tools, 
    llm, 
    agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
    verbose=True
)

result = agent.run("What is the square root of the year Plato was born?")
print(result)

Enter fullscreen mode Exit fullscreen mode

This example creates an agent that can use Wikipedia and perform mathematical calculations to answer complex questions.

3.3 Creating Custom Agents

You can create custom agents by defining your own tools and agent classes. This allows for highly specialized agents tailored to specific tasks or domains.

Here’s an example of a custom agent: 

<span>from</span> <span>langchain.agents</span> <span>import</span> <span>Tool</span><span>,</span> <span>AgentExecutor</span><span>,</span> <span>LLMSingleActionAgent</span>
<span>from</span> <span>langchain.prompts</span> <span>import</span> <span>StringPromptTemplate</span>
<span>from</span> <span>langchain</span> <span>import</span> <span>OpenAI</span><span>,</span> <span>SerpAPIWrapper</span><span>,</span> <span>LLMChain</span>
<span>from</span> <span>typing</span> <span>import</span> <span>List</span><span>,</span> <span>Union</span>
<span>from</span> <span>langchain.schema</span> <span>import</span> <span>AgentAction</span><span>,</span> <span>AgentFinish</span>
<span>import</span> <span>re</span>
<span># Define custom tools </span><span>search</span> <span>=</span> <span>SerpAPIWrapper</span><span>()</span>
<span>tools</span> <span>=</span> <span>[</span>
    <span>Tool</span><span>(</span>
        <span>name</span><span>=</span><span>"</span><span>Search</span><span>"</span><span>,</span>
        <span>func</span><span>=</span><span>search</span><span>.</span><span>run</span><span>,</span>
        <span>description</span><span>=</span><span>"</span><span>Useful for answering questions about current events</span><span>"</span>
    <span>)</span>
<span>]</span>
<span># Define a custom prompt template </span><span>template</span> <span>=</span> <span>"""</span><span>Answer the following questions as best you can: {input} Use the following format: Question: the input question you must answer Thought: you should always think about what to do Action: the action to take, should be one of [{tool_names}] Action Input: the input to the action Observation: the result of the action </span><span>...</span> <span>(</span><span>this</span> <span>Thought</span><span>/</span><span>Action</span><span>/</span><span>Action</span> <span>Input</span><span>/</span><span>Observation</span> <span>can</span> <span>repeat</span> <span>N</span> <span>times</span><span>)</span>
<span>Thought</span><span>:</span> <span>I</span> <span>now</span> <span>know</span> <span>the</span> <span>final</span> <span>answer</span>
<span>Final</span> <span>Answer</span><span>:</span> <span>the</span> <span>final</span> <span>answer</span> <span>to</span> <span>the</span> <span>original</span> <span>input</span> <span>question</span>
<span>Begin! Question: {input} Thought: To answer this question, I need to search for current information. {agent_scratchpad}</span><span>"""</span>
<span>class</span> <span>CustomPromptTemplate</span><span>(</span><span>StringPromptTemplate</span><span>):</span>
    <span>template</span><span>:</span> <span>str</span>
    <span>tools</span><span>:</span> <span>List</span><span>[</span><span>Tool</span><span>]</span>
    <span>def</span> <span>format</span><span>(</span><span>self</span><span>,</span> <span>**</span><span>kwargs</span><span>)</span> <span>-></span> <span>str</span><span>:</span>
        <span>intermediate_steps</span> <span>=</span> <span>kwargs</span><span>.</span><span>pop</span><span>(</span><span>"</span><span>intermediate_steps</span><span>"</span><span>)</span>
        <span>thoughts</span> <span>=</span> <span>""</span>
        <span>for</span> <span>action</span><span>,</span> <span>observation</span> <span>in</span> <span>intermediate_steps</span><span>:</span>
            <span>thoughts</span> <span>+=</span> <span>action</span><span>.</span><span>log</span>
            <span>thoughts</span> <span>+=</span> <span>f</span><span>"</span><span>\n</span><span>Observation: </span><span>{</span><span>observation</span><span>}</span><span>\n</span><span>Thought: </span><span>"</span>
        <span>kwargs</span><span>[</span><span>"</span><span>agent_scratchpad</span><span>"</span><span>]</span> <span>=</span> <span>thoughts</span>
        <span>kwargs</span><span>[</span><span>"</span><span>tool_names</span><span>"</span><span>]</span> <span>=</span> <span>"</span><span>, </span><span>"</span><span>.</span><span>join</span><span>([</span><span>tool</span><span>.</span><span>name</span> <span>for</span> <span>tool</span> <span>in</span> <span>self</span><span>.</span><span>tools</span><span>])</span>
        <span>return</span> <span>self</span><span>.</span><span>template</span><span>.</span><span>format</span><span>(</span><span>**</span><span>kwargs</span><span>)</span>
<span>prompt</span> <span>=</span> <span>CustomPromptTemplate</span><span>(</span>
    <span>template</span><span>=</span><span>template</span><span>,</span>
    <span>tools</span><span>=</span><span>tools</span><span>,</span>
    <span>input_variables</span><span>=</span><span>[</span><span>"</span><span>input</span><span>"</span><span>,</span> <span>"</span><span>intermediate_steps</span><span>"</span><span>]</span>
<span>)</span>
<span># Define a custom output parser </span><span>class</span> <span>CustomOutputParser</span><span>:</span>
    <span>def</span> <span>parse</span><span>(</span><span>self</span><span>,</span> <span>llm_output</span><span>:</span> <span>str</span><span>)</span> <span>-></span> <span>Union</span><span>[</span><span>AgentAction</span><span>,</span> <span>AgentFinish</span><span>]:</span>
        <span>if</span> <span>"</span><span>Final Answer:</span><span>"</span> <span>in</span> <span>llm_output</span><span>:</span>
            <span>return</span> <span>AgentFinish</span><span>(</span>
                <span>return_values</span><span>=</span><span>{</span><span>"</span><span>output</span><span>"</span><span>:</span> <span>llm_output</span><span>.</span><span>split</span><span>(</span><span>"</span><span>Final Answer:</span><span>"</span><span>)[</span><span>-</span><span>1</span><span>].</span><span>strip</span><span>()},</span>
                <span>log</span><span>=</span><span>llm_output</span><span>,</span>
            <span>)</span>
        <span>action_match</span> <span>=</span> <span>re</span><span>.</span><span>search</span><span>(</span><span>r</span><span>"</span><span>Action: (\w+)</span><span>"</span><span>,</span> <span>llm_output</span><span>,</span> <span>re</span><span>.</span><span>DOTALL</span><span>)</span>
        <span>action_input_match</span> <span>=</span> <span>re</span><span>.</span><span>search</span><span>(</span><span>r</span><span>"</span><span>Action Input: (.*)</span><span>"</span><span>,</span> <span>llm_output</span><span>,</span> <span>re</span><span>.</span><span>DOTALL</span><span>)</span>
        <span>if</span> <span>not</span> <span>action_match</span> <span>or</span> <span>not</span> <span>action_input_match</span><span>:</span>
            <span>raise</span> <span>ValueError</span><span>(</span><span>f</span><span>"</span><span>Could not parse LLM output: `</span><span>{</span><span>llm_output</span><span>}</span><span>`</span><span>"</span><span>)</span>
        <span>action</span> <span>=</span> <span>action_match</span><span>.</span><span>group</span><span>(</span><span>1</span><span>).</span><span>strip</span><span>()</span>
        <span>action_input</span> <span>=</span> <span>action_input_match</span><span>.</span><span>group</span><span>(</span><span>1</span><span>).</span><span>strip</span><span>(</span><span>"</span><span> </span><span>"</span><span>).</span><span>strip</span><span>(</span><span>'"'</span><span>)</span>
        <span>return</span> <span>AgentAction</span><span>(</span><span>tool</span><span>=</span><span>action</span><span>,</span> <span>tool_input</span><span>=</span><span>action_input</span><span>,</span> <span>log</span><span>=</span><span>llm_output</span><span>)</span>
<span># Create the custom output parser </span><span>output_parser</span> <span>=</span> <span>CustomOutputParser</span><span>()</span>
<span># Define the LLM chain </span><span>llm</span> <span>=</span> <span>OpenAI</span><span>(</span><span>temperature</span><span>=</span><span>0</span><span>)</span>
<span>llm_chain</span> <span>=</span> <span>LLMChain</span><span>(</span><span>llm</span><span>=</span><span>llm</span><span>,</span> <span>prompt</span><span>=</span><span>prompt</span><span>)</span>
<span># Define the custom agent </span><span>agent</span> <span>=</span> <span>LLMSingleActionAgent</span><span>(</span>
    <span>llm_chain</span><span>=</span><span>llm_chain</span><span>,</span>
    <span>output_parser</span><span>=</span><span>output_parser</span><span>,</span>
    <span>stop</span><span>=</span><span>[</span><span>"</span><span>\n</span><span>Observation:</span><span>"</span><span>],</span>
    <span>allowed_tools</span><span>=</span><span>[</span><span>tool</span><span>.</span><span>name</span> <span>for</span> <span>tool</span> <span>in</span> <span>tools</span><span>]</span>
<span>)</span>
<span># Create an agent executor </span><span>agent_executor</span> <span>=</span> <span>AgentExecutor</span><span>.</span><span>from_agent_and_tools</span><span>(</span><span>agent</span><span>=</span><span>agent</span><span>,</span> <span>tools</span><span>=</span><span>tools</span><span>,</span> <span>,</span> <span>verbose</span><span>=</span><span>True</span><span>)</span>
<span># Run the agent </span><span>result</span> <span>=</span> <span>agent_executor</span><span>.</span><span>run</span><span>(</span><span>“</span><span>What</span><span>’</span><span>s</span> <span>the</span> <span>latest</span> <span>news</span> <span>about</span> <span>AI</span><span>?”</span><span>)</span>
<span>print</span><span>(</span><span>result</span><span>)</span>
<span>from</span> <span>langchain.agents</span> <span>import</span> <span>Tool</span><span>,</span> <span>AgentExecutor</span><span>,</span> <span>LLMSingleActionAgent</span>
<span>from</span> <span>langchain.prompts</span> <span>import</span> <span>StringPromptTemplate</span>
<span>from</span> <span>langchain</span> <span>import</span> <span>OpenAI</span><span>,</span> <span>SerpAPIWrapper</span><span>,</span> <span>LLMChain</span>
<span>from</span> <span>typing</span> <span>import</span> <span>List</span><span>,</span> <span>Union</span>
<span>from</span> <span>langchain.schema</span> <span>import</span> <span>AgentAction</span><span>,</span> <span>AgentFinish</span>
<span>import</span> <span>re</span>

<span># Define custom tools </span><span>search</span> <span>=</span> <span>SerpAPIWrapper</span><span>()</span>
<span>tools</span> <span>=</span> <span>[</span>
    <span>Tool</span><span>(</span>
        <span>name</span><span>=</span><span>"</span><span>Search</span><span>"</span><span>,</span>
        <span>func</span><span>=</span><span>search</span><span>.</span><span>run</span><span>,</span>
        <span>description</span><span>=</span><span>"</span><span>Useful for answering questions about current events</span><span>"</span>
    <span>)</span>
<span>]</span>

<span># Define a custom prompt template </span><span>template</span> <span>=</span> <span>"""</span><span>Answer the following questions as best you can: {input} Use the following format: Question: the input question you must answer Thought: you should always think about what to do Action: the action to take, should be one of [{tool_names}] Action Input: the input to the action Observation: the result of the action </span><span>...</span> <span>(</span><span>this</span> <span>Thought</span><span>/</span><span>Action</span><span>/</span><span>Action</span> <span>Input</span><span>/</span><span>Observation</span> <span>can</span> <span>repeat</span> <span>N</span> <span>times</span><span>)</span>
<span>Thought</span><span>:</span> <span>I</span> <span>now</span> <span>know</span> <span>the</span> <span>final</span> <span>answer</span>
<span>Final</span> <span>Answer</span><span>:</span> <span>the</span> <span>final</span> <span>answer</span> <span>to</span> <span>the</span> <span>original</span> <span>input</span> <span>question</span>

<span>Begin! Question: {input} Thought: To answer this question, I need to search for current information. {agent_scratchpad}</span><span>"""</span>

<span>class</span> <span>CustomPromptTemplate</span><span>(</span><span>StringPromptTemplate</span><span>):</span>
    <span>template</span><span>:</span> <span>str</span>
    <span>tools</span><span>:</span> <span>List</span><span>[</span><span>Tool</span><span>]</span>

    <span>def</span> <span>format</span><span>(</span><span>self</span><span>,</span> <span>**</span><span>kwargs</span><span>)</span> <span>-></span> <span>str</span><span>:</span>
        <span>intermediate_steps</span> <span>=</span> <span>kwargs</span><span>.</span><span>pop</span><span>(</span><span>"</span><span>intermediate_steps</span><span>"</span><span>)</span>
        <span>thoughts</span> <span>=</span> <span>""</span>
        <span>for</span> <span>action</span><span>,</span> <span>observation</span> <span>in</span> <span>intermediate_steps</span><span>:</span>
            <span>thoughts</span> <span>+=</span> <span>action</span><span>.</span><span>log</span>
            <span>thoughts</span> <span>+=</span> <span>f</span><span>"</span><span>\n</span><span>Observation: </span><span>{</span><span>observation</span><span>}</span><span>\n</span><span>Thought: </span><span>"</span>
        <span>kwargs</span><span>[</span><span>"</span><span>agent_scratchpad</span><span>"</span><span>]</span> <span>=</span> <span>thoughts</span>
        <span>kwargs</span><span>[</span><span>"</span><span>tool_names</span><span>"</span><span>]</span> <span>=</span> <span>"</span><span>, </span><span>"</span><span>.</span><span>join</span><span>([</span><span>tool</span><span>.</span><span>name</span> <span>for</span> <span>tool</span> <span>in</span> <span>self</span><span>.</span><span>tools</span><span>])</span>
        <span>return</span> <span>self</span><span>.</span><span>template</span><span>.</span><span>format</span><span>(</span><span>**</span><span>kwargs</span><span>)</span>

<span>prompt</span> <span>=</span> <span>CustomPromptTemplate</span><span>(</span>
    <span>template</span><span>=</span><span>template</span><span>,</span>
    <span>tools</span><span>=</span><span>tools</span><span>,</span>
    <span>input_variables</span><span>=</span><span>[</span><span>"</span><span>input</span><span>"</span><span>,</span> <span>"</span><span>intermediate_steps</span><span>"</span><span>]</span>
<span>)</span>

<span># Define a custom output parser </span><span>class</span> <span>CustomOutputParser</span><span>:</span>
    <span>def</span> <span>parse</span><span>(</span><span>self</span><span>,</span> <span>llm_output</span><span>:</span> <span>str</span><span>)</span> <span>-></span> <span>Union</span><span>[</span><span>AgentAction</span><span>,</span> <span>AgentFinish</span><span>]:</span>
        <span>if</span> <span>"</span><span>Final Answer:</span><span>"</span> <span>in</span> <span>llm_output</span><span>:</span>
            <span>return</span> <span>AgentFinish</span><span>(</span>
                <span>return_values</span><span>=</span><span>{</span><span>"</span><span>output</span><span>"</span><span>:</span> <span>llm_output</span><span>.</span><span>split</span><span>(</span><span>"</span><span>Final Answer:</span><span>"</span><span>)[</span><span>-</span><span>1</span><span>].</span><span>strip</span><span>()},</span>
                <span>log</span><span>=</span><span>llm_output</span><span>,</span>
            <span>)</span>

        <span>action_match</span> <span>=</span> <span>re</span><span>.</span><span>search</span><span>(</span><span>r</span><span>"</span><span>Action: (\w+)</span><span>"</span><span>,</span> <span>llm_output</span><span>,</span> <span>re</span><span>.</span><span>DOTALL</span><span>)</span>
        <span>action_input_match</span> <span>=</span> <span>re</span><span>.</span><span>search</span><span>(</span><span>r</span><span>"</span><span>Action Input: (.*)</span><span>"</span><span>,</span> <span>llm_output</span><span>,</span> <span>re</span><span>.</span><span>DOTALL</span><span>)</span>

        <span>if</span> <span>not</span> <span>action_match</span> <span>or</span> <span>not</span> <span>action_input_match</span><span>:</span>
            <span>raise</span> <span>ValueError</span><span>(</span><span>f</span><span>"</span><span>Could not parse LLM output: `</span><span>{</span><span>llm_output</span><span>}</span><span>`</span><span>"</span><span>)</span>

        <span>action</span> <span>=</span> <span>action_match</span><span>.</span><span>group</span><span>(</span><span>1</span><span>).</span><span>strip</span><span>()</span>
        <span>action_input</span> <span>=</span> <span>action_input_match</span><span>.</span><span>group</span><span>(</span><span>1</span><span>).</span><span>strip</span><span>(</span><span>"</span><span> </span><span>"</span><span>).</span><span>strip</span><span>(</span><span>'"'</span><span>)</span>

        <span>return</span> <span>AgentAction</span><span>(</span><span>tool</span><span>=</span><span>action</span><span>,</span> <span>tool_input</span><span>=</span><span>action_input</span><span>,</span> <span>log</span><span>=</span><span>llm_output</span><span>)</span>

<span># Create the custom output parser </span><span>output_parser</span> <span>=</span> <span>CustomOutputParser</span><span>()</span>

<span># Define the LLM chain </span><span>llm</span> <span>=</span> <span>OpenAI</span><span>(</span><span>temperature</span><span>=</span><span>0</span><span>)</span>
<span>llm_chain</span> <span>=</span> <span>LLMChain</span><span>(</span><span>llm</span><span>=</span><span>llm</span><span>,</span> <span>prompt</span><span>=</span><span>prompt</span><span>)</span>

<span># Define the custom agent </span><span>agent</span> <span>=</span> <span>LLMSingleActionAgent</span><span>(</span>
    <span>llm_chain</span><span>=</span><span>llm_chain</span><span>,</span>
    <span>output_parser</span><span>=</span><span>output_parser</span><span>,</span>
    <span>stop</span><span>=</span><span>[</span><span>"</span><span>\n</span><span>Observation:</span><span>"</span><span>],</span>
    <span>allowed_tools</span><span>=</span><span>[</span><span>tool</span><span>.</span><span>name</span> <span>for</span> <span>tool</span> <span>in</span> <span>tools</span><span>]</span>
<span>)</span>

<span># Create an agent executor </span><span>agent_executor</span> <span>=</span> <span>AgentExecutor</span><span>.</span><span>from_agent_and_tools</span><span>(</span><span>agent</span><span>=</span><span>agent</span><span>,</span> <span>tools</span><span>=</span><span>tools</span><span>,</span> <span>,</span> <span>verbose</span><span>=</span><span>True</span><span>)</span>
<span># Run the agent </span><span>result</span> <span>=</span> <span>agent_executor</span><span>.</span><span>run</span><span>(</span><span>“</span><span>What</span><span>’</span><span>s</span> <span>the</span> <span>latest</span> <span>news</span> <span>about</span> <span>AI</span><span>?”</span><span>)</span>

<span>print</span><span>(</span><span>result</span><span>)</span>
from langchain.agents import Tool, AgentExecutor, LLMSingleActionAgent
from langchain.prompts import StringPromptTemplate
from langchain import OpenAI, SerpAPIWrapper, LLMChain
from typing import List, Union
from langchain.schema import AgentAction, AgentFinish
import re

# Define custom tools search = SerpAPIWrapper()
tools = [
    Tool(
        name="Search",
        func=search.run,
        description="Useful for answering questions about current events"
    )
]

# Define a custom prompt template template = """Answer the following questions as best you can: {input} Use the following format: Question: the input question you must answer Thought: you should always think about what to do Action: the action to take, should be one of [{tool_names}] Action Input: the input to the action Observation: the result of the action ... (this Thought/Action/Action Input/Observation can repeat N times)
Thought: I now know the final answer
Final Answer: the final answer to the original input question

Begin! Question: {input} Thought: To answer this question, I need to search for current information. {agent_scratchpad}"""

class CustomPromptTemplate(StringPromptTemplate):
    template: str
    tools: List[Tool]

    def format(self, **kwargs) -> str:
        intermediate_steps = kwargs.pop("intermediate_steps")
        thoughts = ""
        for action, observation in intermediate_steps:
            thoughts += action.log
            thoughts += f"\nObservation: {observation}\nThought: "
        kwargs["agent_scratchpad"] = thoughts
        kwargs["tool_names"] = ", ".join([tool.name for tool in self.tools])
        return self.template.format(**kwargs)

prompt = CustomPromptTemplate(
    template=template,
    tools=tools,
    input_variables=["input", "intermediate_steps"]
)

# Define a custom output parser class CustomOutputParser:
    def parse(self, llm_output: str) -> Union[AgentAction, AgentFinish]:
        if "Final Answer:" in llm_output:
            return AgentFinish(
                return_values={"output": llm_output.split("Final Answer:")[-1].strip()},
                log=llm_output,
            )

        action_match = re.search(r"Action: (\w+)", llm_output, re.DOTALL)
        action_input_match = re.search(r"Action Input: (.*)", llm_output, re.DOTALL)

        if not action_match or not action_input_match:
            raise ValueError(f"Could not parse LLM output: `{llm_output}`")

        action = action_match.group(1).strip()
        action_input = action_input_match.group(1).strip(" ").strip('"')

        return AgentAction(tool=action, tool_input=action_input, log=llm_output)

# Create the custom output parser output_parser = CustomOutputParser()

# Define the LLM chain llm = OpenAI(temperature=0)
llm_chain = LLMChain(llm=llm, prompt=prompt)

# Define the custom agent agent = LLMSingleActionAgent(
    llm_chain=llm_chain,
    output_parser=output_parser,
    stop=["\nObservation:"],
    allowed_tools=[tool.name for tool in tools]
)

# Create an agent executor agent_executor = AgentExecutor.from_agent_and_tools(agent=agent, tools=tools, , verbose=True)
# Run the agent result = agent_executor.run(“What’s the latest news about AI?”)

print(result)

Enter fullscreen mode Exit fullscreen mode

Conclusion

LangChain’s chains and agents offer robust capabilities for constructing sophisticated AI-driven systems. When integrated with large language models (LLMs), they enable the creation of adaptable, smart applications designed to tackle a variety of tasks. As you progress through your LangChain journey, feel free to experiment with diverse chain types, agent setups, and custom modules to fully harness the framework’s potential.

LangChian (4 Part Series)

1 Mastering LangChain: Part 1 – Introduction to LangChain and Its Key Components
2 Part 2: Mastering Prompts and Language Models with LangChain
3 Part 3: Building Powerful Chains and Agents in LangChain
4 LangChain Part 4 – Leveraging Memory and Storage in LangChain: A Comprehensive Guide

原文链接：Part 3: Building Powerful Chains and Agents in LangChain

展开阅读全文
© 版权声明
文章版权声明 1、本网站名称：拾光赋
2、本站永久网址：https://www.blogs.ink
3、本网站的文章部分内容可能来源于网络，仅供大家学习与参考，如有侵权，请联系站长QQ：805375623进行删除处理。
4、本站一切资源不代表本站立场，并不代表本站赞同其观点和对其真实性负责。
5、本站一律禁止以任何方式发布或转载任何违法的相关信息，访客发现请向站长举报
6、本站资源大多存储在云盘，如发现链接失效，请联系我们我们会第一时间更新。

THE END
Python（EN）
# python# llm# openai# langchain
喜欢就支持一下吧
点赞8 分享
Happiness isn't about getting what you want all the time, it's about loving what you have.
幸福并不是一味得到自己想要的，而是珍爱自己拥有的
kity的头像-拾光赋
相关推荐
评论 抢沙发

请登录后发表评论

    暂无评论内容

文章目录
kity的头像-拾光赋
文章目录
今日剩余 59.3%
本周剩余 79.9%
本月剩余 43.8%
本年剩余 79.1%
最近评论
Vito的头像-拾光赋

Vito5个月前2

你就是我心中的那首忐忑，总是让我惊心动魄
Jason.Meng的头像-拾光赋钻石会员

Jason.Meng徽章-人气大使-拾光赋5个月前2

总是会在人生的不同阶段反复爱上大佬的文章~
每日一言
If you don't enjoy your life, sorrow, sadness, fear, shame and guilt will.
如果你不好好享受生活，你的悲伤、难过、害怕、羞愧和内疚会代替你享受