LANGCHAIN Archives - 26 중 4 번째 페이지

[PYTHON/LANGCHAIN] HumanMessage 클래스 : 모델에 멀티모달 데이터 전달하기 1

■ HumanMessage 클래스를 사용해 모델에 멀티모달 데이터를 전달하는 방법을 보여준다. ※ OPENAI_API_KEY 환경 변수 값은 .env 파일에 정의한다. ▶ main.py


import httpx
import base64

from dotenv                  import load_dotenv
from langchain_core.messages import HumanMessage
from langchain_openai        import ChatOpenAI

load_dotenv()

imageURL = "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"

responseAIMessage = httpx.get(imageURL)

contentBytes = responseAIMessage.content

base64Bytes = base64.b64encode(contentBytes)

imageDataString = base64Bytes.decode("utf-8")

hummanMessage = HumanMessage(
    content = [
        {
            "type" : "text",
            "text" : "describe the weather in this image"
        },
        {
            "type"      : "image_url",
            "image_url" : {"url" : f"data:image/jpeg;base64,{imageDataString}" }
        }
    ]
)

chatOpenAI = ChatOpenAI(model = "gpt-4o")

responseAIMessage = chatOpenAI.invoke([hummanMessage])

print(responseAIMessage.content)

"""
The weather in the image appears to be clear and sunny. The sky is mostly blue with scattered clouds, suggesting a pleasant day. The bright sunlight is casting clear shadows, indicating good visibility and a likely warm temperature. The overall scene conveys a calm and serene atmosphere.
"""

import httpx

import base64

from dotenv import load_dotenv

from langchain_core.messages import HumanMessage

from langchain_openai import ChatOpenAI

load_dotenv()

imageURL = "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"

responseAIMessage = httpx.get(imageURL)

contentBytes = responseAIMessage.content

base64Bytes = base64.b64encode(contentBytes)

imageDataString = base64Bytes.decode("utf-8")

hummanMessage = HumanMessage(

content = [

{

"type" : "text",

"text" : "describe the weather in this image"

{

"type" : "image_url",

"image_url" : {"url" : f"data:image/jpeg;base64,{imageDataString}" }

}

]

)

chatOpenAI = ChatOpenAI(model = "gpt-4o")

responseAIMessage = chatOpenAI.invoke([hummanMessage])

print(responseAIMessage.content)

"""

The weather in the image appears to be clear and sunny. The sky is mostly blue with scattered clouds, suggesting a pleasant day. The bright sunlight is casting clear shadows, indicating good visibility and a likely warm temperature. The overall scene conveys a calm and serene atmosphere.

"""

▶

[PYTHON/LANGCHAIN] RunnableConfig 클래스 : 런타임 비밀 값 전달하기

■ RunnableConfig 클래스를 사용해 런타임 비밀 값을 전달하는 방법을 보여준다. ※ RunnableConfig 객체를 사용하여 런타임에 비밀 값을 실행 파일에 전달할 수 있다.

[PYTHON/LANGCHAIN] 모델에서 임시 도구 호출 기능 추가하기 4

■ 모델에서 임시 도구 호출 기능을 추가하는 방법을 보여준다. ※ 도구 출력뿐만 아니라 도구 입력도 반환하는 것이 도움이 될 수 있다. ※

[PYTHON/LANGCHAIN] 모델에서 임시 도구 호출 기능 추가하기 3

■ 모델에서 임시 도구 호출 기능을 추가하는 방법을 보여준다. ▶ main.py


from langchain_core.tools          import tool
from langchain_core.tools          import render_text_description
from langchain_core.prompts        import ChatPromptTemplate
from langchain_community.llms      import Ollama
from langchain_core.output_parsers import JsonOutputParser
from typing                        import TypedDict
from typing                        import Dict
from typing                        import Any
from typing                        import Optional
from langchain_core.runnables      import RunnableConfig

@tool
def multiply(x : float, y : float) -> float:
    """Multiply two numbers together."""
    return x * y

@tool
def add(x : int, y : int) -> int:
    "Add two numbers."
    return x + y

toolList = [multiply, add]

toolListTextDescription = render_text_description(toolList)

systemPromptTemplateString = f"""\
You are an assistant that has access to the following set of tools.
Here are the names and descriptions for each tool:

{toolListTextDescription}

Given the user input, return the name and input of the tool to use.
Return your response as a JSON blob with 'name' and 'argumentDictionary' keys.

The `argumentDictionary` should be a dictionary, with keys corresponding
to the argument names and the values corresponding to the requested values.
"""

chatPromptTemplate = ChatPromptTemplate.from_messages(
    [
        ("system", systemPromptTemplateString),
        ("user"  , "{input}")
    ]
)

ollama = Ollama(model = "phi3:latest")

jsonOutputParser = JsonOutputParser()

class ToolCallRequest(TypedDict):
    """A typed dict that shows the inputs into the invoke_tool function."""

    name               : str
    argumentDictionary : Dict[str, Any]

def invokeTool(toolCallRequest : ToolCallRequest, runnableConfig : Optional[RunnableConfig] = None):
    """A function that we can use the perform a tool invocation.

    Args :
        toolCallRequest : a dict that contains the keys name and arguments.
            The name must match the name of a tool that exists.
            The argumentDictionary are the arguments to that tool.
        runnableConfig : This is configuration information that LangChain uses that contains
            things like callbacks, metadata, etc.See LCEL documentation about RunnableConfig.

    Returns :
        output from the requested tool
    """
    toolNameDictionary = {tool.name : tool for tool in toolList}
    name = toolCallRequest["name"]
    tool = toolNameDictionary[name]
    return tool.invoke(toolCallRequest["argumentDictionary"], config = runnableConfig)

runnableSequence = chatPromptTemplate | ollama | jsonOutputParser | invokeTool

responseValue = runnableSequence.invoke({"input" : "what's thirteen times 4.14137281"})

print(responseValue)

"""
53.83784653
"""

from langchain_core.tools import tool

from langchain_core.tools import render_text_description

from langchain_core.prompts import ChatPromptTemplate

from langchain_community.llms import Ollama

from langchain_core.output_parsers import JsonOutputParser

from typing import TypedDict

from typing import Dict

from typing import Any

from typing import Optional

from langchain_core.runnables import RunnableConfig

@tool

def multiply(x : float, y : float) -> float:

"""Multiply two numbers together."""

return x * y

@tool

def add(x : int, y : int) -> int:

"Add two numbers."

return x + y

toolList = [multiply, add]

toolListTextDescription = render_text_description(toolList)

systemPromptTemplateString = f"""\

You are an assistant that has access to the following set of tools.

Here are the names and descriptions for each tool:

{toolListTextDescription}

Given the user input, return the name and input of the tool to use.

Return your response as a JSON blob with 'name' and 'argumentDictionary' keys.

The `argumentDictionary` should be a dictionary, with keys corresponding

to the argument names and the values corresponding to the requested values.

"""

chatPromptTemplate = ChatPromptTemplate.from_messages(

[

("system", systemPromptTemplateString),

("user" , "{input}")

]

)

ollama = Ollama(model = "phi3:latest")

jsonOutputParser = JsonOutputParser()

class ToolCallRequest(TypedDict):

"""A typed dict that shows the inputs into the invoke_tool function."""

name : str

argumentDictionary : Dict[str, Any]

def invokeTool(toolCallRequest : ToolCallRequest, runnableConfig : Optional[RunnableConfig] = None):

"""A function that we can use the perform a tool invocation.

Args :

toolCallRequest : a dict that contains the keys name and arguments.

The name must match the name of a tool that exists.

The argumentDictionary are the arguments to that tool.

runnableConfig : This is configuration information that LangChain uses that contains

things like callbacks, metadata, etc.See LCEL documentation about RunnableConfig.

Returns :

output from the requested tool

"""

toolNameDictionary = {tool.name : tool for tool in toolList}

name = toolCallRequest["name"]

tool = toolNameDictionary[name]

return tool.invoke(toolCallRequest["argumentDictionary"], config = runnableConfig)

runnableSequence = chatPromptTemplate | ollama | jsonOutputParser | invokeTool

responseValue = runnableSequence.invoke({"input" : "what's thirteen times 4.14137281"})

print(responseValue)

"""

53.83784653

"""

▶ requirements.txt


aiohappyeyeballs==2.4.3
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
dataclasses-json==0.6.7
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.6
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-community==0.3.1
langchain-core==0.3.8
langchain-text-splitters==0.3.0
langsmith==0.1.130
marshmallow==3.22.0
multidict==6.1.0
mypy-extensions==1.0.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic-settings==2.5.2
pydantic_core==2.23.4
python-dotenv==1.0.1
PyYAML==6.0.2
requests==2.32.3
requests-toolbelt==1.0.0
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing-inspect==0.9.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.3

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

dataclasses-json==0.6.7

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.6

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-community==0.3.1

langchain-core==0.3.8

langchain-text-splitters==0.3.0

langsmith==0.1.130

marshmallow==3.22.0

multidict==6.1.0

mypy-extensions==1.0.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic-settings==2.5.2

pydantic_core==2.23.4

python-dotenv==1.0.1

PyYAML==6.0.2

requests==2.32.3

requests-toolbelt==1.0.0

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing-inspect==0.9.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install langchain langchain-community 명령을 실행했다.

[PYTHON/LANGCHAIN] 모델에서 임시 도구 호출 기능 추가하기 2

■ 모델에서 임시 도구 호출 기능을 추가하는 방법을 보여준다. ▶ main.py


from langchain_core.tools          import tool
from langchain_core.tools          import render_text_description
from langchain_core.prompts        import ChatPromptTemplate
from langchain_community.llms      import Ollama
from langchain_core.output_parsers import JsonOutputParser

@tool
def multiply(x : float, y : float) -> float:
    """Multiply two numbers together."""
    return x * y

@tool
def add(x : int, y : int) -> int:
    "Add two numbers."
    return x + y

toolList = [multiply, add]

toolListTextDescription = render_text_description(toolList)

systemPromptTemplateString = f"""\
You are an assistant that has access to the following set of tools.
Here are the names and descriptions for each tool:

{toolListTextDescription}

Given the user input, return the name and input of the tool to use.
Return your response as a JSON blob with 'name' and 'arguments' keys.

The `arguments` should be a dictionary, with keys corresponding
to the argument names and the values corresponding to the requested values.
"""

chatPromptTemplate = ChatPromptTemplate.from_messages(
    [
        ("system", systemPromptTemplateString),
        ("user"  , "{input}")
    ]
)

ollama = Ollama(model = "phi3:latest")

jsonOutputParser = JsonOutputParser()

runnableSequence = chatPromptTemplate | ollama | jsonOutputParser

responseDictionary = runnableSequence.invoke({"input" : "what's 3 plus 1132"})

print(responseDictionary)

"""
{'name': 'add', 'arguments': {'x': 3, 'y': 1132}}
"""

from langchain_core.tools import tool

from langchain_core.tools import render_text_description

from langchain_core.prompts import ChatPromptTemplate

from langchain_community.llms import Ollama

from langchain_core.output_parsers import JsonOutputParser

@tool

def multiply(x : float, y : float) -> float:

"""Multiply two numbers together."""

return x * y

@tool

def add(x : int, y : int) -> int:

"Add two numbers."

return x + y

toolList = [multiply, add]

toolListTextDescription = render_text_description(toolList)

systemPromptTemplateString = f"""\

You are an assistant that has access to the following set of tools.

Here are the names and descriptions for each tool:

{toolListTextDescription}

Given the user input, return the name and input of the tool to use.

Return your response as a JSON blob with 'name' and 'arguments' keys.

The `arguments` should be a dictionary, with keys corresponding

to the argument names and the values corresponding to the requested values.

"""

chatPromptTemplate = ChatPromptTemplate.from_messages(

[

("system", systemPromptTemplateString),

("user" , "{input}")

]

)

ollama = Ollama(model = "phi3:latest")

jsonOutputParser = JsonOutputParser()

runnableSequence = chatPromptTemplate | ollama | jsonOutputParser

responseDictionary = runnableSequence.invoke({"input" : "what's 3 plus 1132"})

print(responseDictionary)

"""

{'name': 'add', 'arguments': {'x': 3, 'y': 1132}}

"""

▶ requirements.txt


aiohappyeyeballs==2.4.3
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
dataclasses-json==0.6.7
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.6
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-community==0.3.1
langchain-core==0.3.8
langchain-text-splitters==0.3.0
langsmith==0.1.130
marshmallow==3.22.0
multidict==6.1.0
mypy-extensions==1.0.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic-settings==2.5.2
pydantic_core==2.23.4
python-dotenv==1.0.1
PyYAML==6.0.2
requests==2.32.3
requests-toolbelt==1.0.0
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing-inspect==0.9.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.3

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

dataclasses-json==0.6.7

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.6

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-community==0.3.1

langchain-core==0.3.8

langchain-text-splitters==0.3.0

langsmith==0.1.130

marshmallow==3.22.0

multidict==6.1.0

mypy-extensions==1.0.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic-settings==2.5.2

pydantic_core==2.23.4

python-dotenv==1.0.1

PyYAML==6.0.2

requests==2.32.3

requests-toolbelt==1.0.0

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing-inspect==0.9.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install langchain langchain-community 명령을 실행했다.

[PYTHON/LANGCHAIN] 모델에서 임시 도구 호출 기능 추가하기 1

■ 모델에서 임시 도구 호출 기능을 추가하는 방법을 보여준다. ▶ main.py


from langchain_core.tools     import tool
from langchain_core.tools     import render_text_description
from langchain_core.prompts   import ChatPromptTemplate
from langchain_community.llms import Ollama

@tool
def multiply(x : float, y : float) -> float:
    """Multiply two numbers together."""
    return x * y

@tool
def add(x : int, y : int) -> int:
    "Add two numbers."
    return x + y

toolList = [multiply, add]

toolListTextDescription = render_text_description(toolList)

systemPromptTemplateString = f"""\
You are an assistant that has access to the following set of tools.
Here are the names and descriptions for each tool:

{toolListTextDescription}

Given the user input, return the name and input of the tool to use.
Return your response as a JSON blob with 'name' and 'arguments' keys.

The `arguments` should be a dictionary, with keys corresponding
to the argument names and the values corresponding to the requested values.
"""

chatPromptTemplate = ChatPromptTemplate.from_messages(
    [
        ("system", systemPromptTemplateString),
        ("user"  , "{input}")
    ]
)

ollama = Ollama(model = "phi3:latest")

runnableSequence = chatPromptTemplate | ollama

response = runnableSequence.invoke({"input" : "what's 3 plus 1132"})

if isinstance(response, str):
    print(response)
else:
    print(response.content)

"""
```json
{
  "name": "add",
  "arguments": {
    "x": 3,
    "y": 1132
  }
}
```

How can I assist you further? (If the user'sin a different context or needs help with something else)
"""

from langchain_core.tools import tool

from langchain_core.tools import render_text_description

from langchain_core.prompts import ChatPromptTemplate

from langchain_community.llms import Ollama

@tool

def multiply(x : float, y : float) -> float:

"""Multiply two numbers together."""

return x * y

@tool

def add(x : int, y : int) -> int:

"Add two numbers."

return x + y

toolList = [multiply, add]

toolListTextDescription = render_text_description(toolList)

systemPromptTemplateString = f"""\

You are an assistant that has access to the following set of tools.

Here are the names and descriptions for each tool:

{toolListTextDescription}

Given the user input, return the name and input of the tool to use.

Return your response as a JSON blob with 'name' and 'arguments' keys.

The `arguments` should be a dictionary, with keys corresponding

to the argument names and the values corresponding to the requested values.

"""

chatPromptTemplate = ChatPromptTemplate.from_messages(

[

("system", systemPromptTemplateString),

("user" , "{input}")

]

)

ollama = Ollama(model = "phi3:latest")

runnableSequence = chatPromptTemplate | ollama

response = runnableSequence.invoke({"input" : "what's 3 plus 1132"})

if isinstance(response, str):

print(response)

else:

print(response.content)

"""

```json

{

"name": "add",

"arguments": {

"x": 3,

"y": 1132

}

```

How can I assist you further? (If the user'sin a different context or needs help with something else)

"""

▶ requirements.txt


aiohappyeyeballs==2.4.3
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
dataclasses-json==0.6.7
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.6
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-community==0.3.1
langchain-core==0.3.8
langchain-text-splitters==0.3.0
langsmith==0.1.130
marshmallow==3.22.0
multidict==6.1.0
mypy-extensions==1.0.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic-settings==2.5.2
pydantic_core==2.23.4
python-dotenv==1.0.1
PyYAML==6.0.2
requests==2.32.3
requests-toolbelt==1.0.0
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing-inspect==0.9.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.3

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

dataclasses-json==0.6.7

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.6

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-community==0.3.1

langchain-core==0.3.8

langchain-text-splitters==0.3.0

langsmith==0.1.130

marshmallow==3.22.0

multidict==6.1.0

mypy-extensions==1.0.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic-settings==2.5.2

pydantic_core==2.23.4

python-dotenv==1.0.1

PyYAML==6.0.2

requests==2.32.3

requests-toolbelt==1.0.0

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing-inspect==0.9.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install langchain langchain-community 명령을 실행했다.

[PYTHON/LANGCHAIN] render_text_description 함수 : 도구 리스트의 텍스트 설명 만들기

■ render_text_description 함수를 사용해 도구 리스트의 텍스트 설명을 만드는 방법을 보여준다. ▶ 예제 코드 (PY)


from langchain_core.tools import tool
from langchain_core.tools import render_text_description

@tool
def multiply(x : float, y : float) -> float:
    """Multiply two numbers together."""
    return x * y

@tool
def add(x : int, y : int) -> int:
    "Add two numbers."
    return x + y

toolList = [multiply, add]

toolListTextDescription = render_text_description(toolList)

"""
multiply(x: float, y: float) -> float - Multiply two numbers together.
add(x: int, y: int) -> int - Add two numbers.
"""

from langchain_core.tools import tool

from langchain_core.tools import render_text_description

@tool

def multiply(x : float, y : float) -> float:

"""Multiply two numbers together."""

return x * y

@tool

def add(x : int, y : int) -> int:

"Add two numbers."

return x + y

toolList = [multiply, add]

toolListTextDescription = render_text_description(toolList)

"""

multiply(x: float, y: float) -> float - Multiply two numbers together.

add(x: int, y: int) -> int - Add two numbers.

"""

※ pip install langchain 명령을 실행했다.

[PYTHON/LANGCHAIN] create_react_agent 함수 : RunnableSequence 클래스의 as_tool 메소드를 사용해 만든 StructuredTool 객체 설정하기

■ create_react_agent 함수에서 RunnableSequence 클래스의 as_tool 메소드를 사용해 만든 StructuredTool 객체를 설정하는 방법을 보여준다. ※ OPENAI_API_KEY 환경 변수 값은 .env 파일에 정의한다.

[PYTHON/LANGCHAIN] create_react_agent 함수 : compiledStateGraph 객체 만들기

■ create_react_agent 함수를 사용해 compiledStateGraph 객체를 만드는 방법을 보여준다. ※ OPENAI_API_KEY 환경 변수 값은 .env 파일에 정의한다. ▶ main.py


from dotenv                      import load_dotenv
from langchain_core.documents    import Document
from langchain_openai            import OpenAIEmbeddings
from langchain_core.vectorstores import InMemoryVectorStore
from langchain_openai            import ChatOpenAI
from langgraph.prebuilt          import create_react_agent

load_dotenv()

documentList = [
    Document(page_content = "Dogs are great companions, known for their loyalty and friendliness."),
    Document(page_content = "Cats are independent pets that often enjoy their own space."         )
]

openAIEmbeddings = OpenAIEmbeddings()

inMemoryVectorStore = InMemoryVectorStore.from_documents(
    documentList,
    embedding = openAIEmbeddings
)

vectorStoreRetriever = inMemoryVectorStore.as_retriever(
    search_type   = "similarity",
    search_kwargs = {"k" : 1}
)

chatOpenAI = ChatOpenAI(model = "gpt-4o-mini")

tool = vectorStoreRetriever.as_tool(
    name        = "pet_info_retriever",
    description = "Get information about pets.",
)

toolList = [tool]

compiledStateGraph = create_react_agent(chatOpenAI, toolList)

for addableUpdatesDict in compiledStateGraph.stream({"messages" : [("human", "What are dogs known for?")]}):
    print(addableUpdatesDict)
    print("-" * 100)

"""
{
    'agent' : {
        'messages' : [
            AIMessage(
                content           = 'Dogs are known for several characteristics and traits, including:\n\n1. **Companionship**: Dogs are often referred to as "man\'s best friend" due to their loyalty and companionship.\n\n2. **Intelligence**: Many dog breeds are highly intelligent and capable of learning a variety of commands and tricks.\n\n3. **Variety of Breeds**: There are hundreds of dog breeds, each with its own unique traits, sizes, and temperaments.\n\n4. **Working Abilities**: Dogs are used in various roles, such as service animals, search and rescue, therapy dogs, and police or military dogs.\n\n5. **Strong Sense of Smell**: Dogs have an exceptional sense of smell, which makes them excellent for tracking and detection purposes.\n\n6. **Social Behavior**: Dogs are social animals and often thrive in the company of humans and other pets.\n\n7. **Playfulness**: Many dogs enjoy playing and being active, which makes them great companions for outdoor activities.\n\n8. **Emotional Support**: Dogs are known to provide emotional support and comfort to their owners, often sensing when someone is feeling down.\n\n9. **Protectiveness**: Many dogs have a natural instinct to protect their home and family, making them good guard animals.\n\n10. **Communication**: Dogs communicate through a combination of vocalizations, body language, and facial expressions. \n\nOverall, dogs are appreciated for their loyalty, intelligence, and the deep bond they can form with humans.',
                additional_kwargs = {'refusal' : None},
                response_metadata = {
                    'token_usage' : {
                        'completion_tokens'         : 299,
                        'prompt_tokens'             : 58,
                        'total_tokens'              : 357,
                        'completion_tokens_details' : {'reasoning_tokens' : 0}
                    },
                    'model_name'         : 'gpt-4o-mini-2024-07-18',
                    'system_fingerprint' : 'fp_f85bea6784',
                    'finish_reason'      : 'stop',
                    'logprobs'           : None
                },
                id                = 'run-b2d78792-6c54-422e-8739-07662d2eb56b-0',
                usage_metadata    = {'input_tokens' : 58, 'output_tokens' : 299, 'total_tokens' : 357}
            )
        ]
    }
}

from dotenv import load_dotenv

from langchain_core.documents import Document

from langchain_openai import OpenAIEmbeddings

from langchain_core.vectorstores import InMemoryVectorStore

from langchain_openai import ChatOpenAI

from langgraph.prebuilt import create_react_agent

load_dotenv()

documentList = [

Document(page_content = "Dogs are great companions, known for their loyalty and friendliness."),

Document(page_content = "Cats are independent pets that often enjoy their own space." )

]

openAIEmbeddings = OpenAIEmbeddings()

inMemoryVectorStore = InMemoryVectorStore.from_documents(

documentList,

embedding = openAIEmbeddings

)

vectorStoreRetriever = inMemoryVectorStore.as_retriever(

search_type = "similarity",

search_kwargs = {"k" : 1}

)

chatOpenAI = ChatOpenAI(model = "gpt-4o-mini")

tool = vectorStoreRetriever.as_tool(

name = "pet_info_retriever",

description = "Get information about pets.",

)

toolList = [tool]

compiledStateGraph = create_react_agent(chatOpenAI, toolList)

for addableUpdatesDict in compiledStateGraph.stream({"messages" : [("human", "What are dogs known for?")]}):

print(addableUpdatesDict)

print("-" * 100)

"""

{

'agent' : {

'messages' : [

AIMessage(

content = 'Dogs are known for several characteristics and traits, including:\n\n1. **Companionship**: Dogs are often referred to as "man\'s best friend" due to their loyalty and companionship.\n\n2. **Intelligence**: Many dog breeds are highly intelligent and capable of learning a variety of commands and tricks.\n\n3. **Variety of Breeds**: There are hundreds of dog breeds, each with its own unique traits, sizes, and temperaments.\n\n4. **Working Abilities**: Dogs are used in various roles, such as service animals, search and rescue, therapy dogs, and police or military dogs.\n\n5. **Strong Sense of Smell**: Dogs have an exceptional sense of smell, which makes them excellent for tracking and detection purposes.\n\n6. **Social Behavior**: Dogs are social animals and often thrive in the company of humans and other pets.\n\n7. **Playfulness**: Many dogs enjoy playing and being active, which makes them great companions for outdoor activities.\n\n8. **Emotional Support**: Dogs are known to provide emotional support and comfort to their owners, often sensing when someone is feeling down.\n\n9. **Protectiveness**: Many dogs have a natural instinct to protect their home and family, making them good guard animals.\n\n10. **Communication**: Dogs communicate through a combination of vocalizations, body language, and facial expressions. \n\nOverall, dogs are appreciated for their loyalty, intelligence, and the deep bond they can form with humans.',

additional_kwargs = {'refusal' : None},

response_metadata = {

'token_usage' : {

'completion_tokens' : 299,

'prompt_tokens' : 58,

'total_tokens' : 357,

'completion_tokens_details' : {'reasoning_tokens' : 0}

'model_name' : 'gpt-4o-mini-2024-07-18',

'system_fingerprint' : 'fp_f85bea6784',

'finish_reason' : 'stop',

'logprobs' : None

id = 'run-b2d78792-6c54-422e-8739-07662d2eb56b-0',

usage_metadata = {'input_tokens' : 58, 'output_tokens' : 299, 'total_tokens' : 357}

)

]

}

""" —————————————————————————————————-

[PYTHON/LANGCHAIN] VectorStoreRetriever 클래스 : as_tool 메소드를 사용해 Tool 객체 만들기

■ VectorStoreRetriever 클래스의 as_tool 메소드를 사용해 Tool 객체를 만드는 방법을 보여준다. ▶ main.py


from langchain_core.documents    import Document
from langchain_openai            import OpenAIEmbeddings
from langchain_core.vectorstores import InMemoryVectorStore

documentList = [
    Document(page_content = "Dogs are great companions, known for their loyalty and friendliness."),
    Document(page_content = "Cats are independent pets that often enjoy their own space."         )
]

openAIEmbeddings = OpenAIEmbeddings()

inMemoryVectorStore = InMemoryVectorStore.from_documents(
    documentList,
    embedding = openAIEmbeddings
)

vectorStoreRetriever = inMemoryVectorStore.as_retriever(
    search_type   = "similarity",
    search_kwargs = {"k" : 1}
)

tool = vectorStoreRetriever.as_tool(
    name        = "pet_info_retriever",
    description = "Get information about pets.",
)

from langchain_core.documents import Document

from langchain_openai import OpenAIEmbeddings

from langchain_core.vectorstores import InMemoryVectorStore

documentList = [

Document(page_content = "Dogs are great companions, known for their loyalty and friendliness."),

Document(page_content = "Cats are independent pets that often enjoy their own space." )

]

openAIEmbeddings = OpenAIEmbeddings()

inMemoryVectorStore = InMemoryVectorStore.from_documents(

documentList,

embedding = openAIEmbeddings

)

vectorStoreRetriever = inMemoryVectorStore.as_retriever(

search_type = "similarity",

search_kwargs = {"k" : 1}

)

tool = vectorStoreRetriever.as_tool(

name = "pet_info_retriever",

description = "Get information about pets.",

)

※ pip install langchain-openai 명령을 실행했다.

[PYTHON/LANGCHAIN] InMemoryVectorStore 클래스 : as_retriever 메소드를 사용해 VectorStoreRetriever 객체 만들기

■ InMemoryVectorStore 클래스의 as_retriever 메소드를 사용해 VectorStoreRetriever 객체를 만드는 방법을 보여준다. ▶ main.py


from langchain_core.documents    import Document
from langchain_openai            import OpenAIEmbeddings
from langchain_core.vectorstores import InMemoryVectorStore

documentList = [
    Document(page_content = "Dogs are great companions, known for their loyalty and friendliness."),
    Document(page_content = "Cats are independent pets that often enjoy their own space."         )
]

openAIEmbeddings = OpenAIEmbeddings()

inMemoryVectorStore = InMemoryVectorStore.from_documents(
    documentList,
    embedding = openAIEmbeddings
)

vectorStoreRetriever = inMemoryVectorStore.as_retriever(
    search_type   = "similarity",
    search_kwargs = {"k" : 1}
)

from langchain_core.documents import Document

from langchain_openai import OpenAIEmbeddings

from langchain_core.vectorstores import InMemoryVectorStore

documentList = [

Document(page_content = "Dogs are great companions, known for their loyalty and friendliness."),

Document(page_content = "Cats are independent pets that often enjoy their own space." )

]

openAIEmbeddings = OpenAIEmbeddings()

inMemoryVectorStore = InMemoryVectorStore.from_documents(

documentList,

embedding = openAIEmbeddings

)

vectorStoreRetriever = inMemoryVectorStore.as_retriever(

search_type = "similarity",

search_kwargs = {"k" : 1}

)

※ pip install langchain-openai 명령을 실행했다.

[PYTHON/LANGCHAIN] InMemoryVectorStore 클래스 : from_documents 정적 메소드를 사용해 InMemoryVectorStore 객체 만들기

■ InMemoryVectorStore 클래스의 from_documents 정적 메소드를 사용해 InMemoryVectorStore 객체를 만드는 방법을 보여준다. ▶ main.py


from langchain_core.documents    import Document
from langchain_openai            import OpenAIEmbeddings
from langchain_core.vectorstores import InMemoryVectorStore

documentList = [
    Document(page_content = "Dogs are great companions, known for their loyalty and friendliness."),
    Document(page_content = "Cats are independent pets that often enjoy their own space."         )
]

openAIEmbeddings = OpenAIEmbeddings()

inMemoryVectorStore = InMemoryVectorStore.from_documents(
    documentList,
    embedding = openAIEmbeddings
)

from langchain_core.documents import Document

from langchain_openai import OpenAIEmbeddings

from langchain_core.vectorstores import InMemoryVectorStore

documentList = [

Document(page_content = "Dogs are great companions, known for their loyalty and friendliness."),

Document(page_content = "Cats are independent pets that often enjoy their own space." )

]

openAIEmbeddings = OpenAIEmbeddings()

inMemoryVectorStore = InMemoryVectorStore.from_documents(

documentList,

embedding = openAIEmbeddings

)

※ pip install langchain-openai 명령을 실행했다.

[PYTHON/LANGCHAIN] Document 클래스 : page_content 속성을 사용해 Document 객체 만들기

■ Document 클래스의 page_content 속성을 사용해 Document 객체를 만드는 방법을 보여준다. ▶ main.py


from langchain_core.documents import Document

documentList = [
    Document(page_content = "Dogs are great companions, known for their loyalty and friendliness."),
    Document(page_content = "Cats are independent pets that often enjoy their own space."         )
]

from langchain_core.documents import Document

documentList = [

Document(page_content = "Dogs are great companions, known for their loyalty and friendliness."),

Document(page_content = "Cats are independent pets that often enjoy their own space." )

]

[PYTHON/LANGCHAIN] RunnableSequence 클래스 : as_tool 메소드를 사용해 StructuredTool 객체 만들기

■ RunnableSequence 클래스의 as_tool 메소드를 사용해 StructuredTool 객체를 만드는 방법을 보여준다. ▶ main.py


from langchain_core.runnables import RunnableLambda

def test1(x : str) -> str:
    return x + "a"

def test2(x : str) -> str:
    return x + "z"

runnableSequence = RunnableLambda(test1) | test2

structuredTool = runnableSequence.as_tool()

responseString = structuredTool.invoke("b")

print(responseString)

"""
baz
"""

from langchain_core.runnables import RunnableLambda

def test1(x : str) -> str:

return x + "a"

def test2(x : str) -> str:

return x + "z"

runnableSequence = RunnableLambda(test1) | test2

structuredTool = runnableSequence.as_tool()

responseString = structuredTool.invoke("b")

print(responseString)

"""

baz

"""

▶ requirements.txt


aiohappyeyeballs==2.4.3
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
async-timeout==4.0.3
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
exceptiongroup==1.2.2
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.5
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-core==0.3.7
langchain-text-splitters==0.3.0
langsmith==0.1.129
multidict==6.1.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic_core==2.23.4
PyYAML==6.0.2
requests==2.32.3
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.3

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

async-timeout==4.0.3

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

exceptiongroup==1.2.2

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.5

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-core==0.3.7

langchain-text-splitters==0.3.0

langsmith==0.1.129

multidict==6.1.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic_core==2.23.4

PyYAML==6.0.2

requests==2.32.3

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install langcahin 명령을

[PYTHON/LANGCHAIN] RunnableLambda 클래스 : as_tool 메소드에서 BaseModel 객체를 설정해 StructuredTool 객체 만들기

■ RunnableLambda 클래스의 as_tool 메소드에서 BaseModel 객체를 설정해 StructuredTool 객체를 만드는 방법을 보여준다. ▶ main.py


from typing                   import Dict
from typing                   import Any
from pydantic                 import BaseModel
from pydantic                 import Field
from typing                   import List
from langchain_core.runnables import RunnableLambda

def test(x : Dict[str, Any]) -> str:
    return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

class TestArgument(BaseModel):
    """Apply a function to an integer and list of integers."""
    a : int       = Field(..., description = "Integer"     )
    b : List[int] = Field(..., description = "List of ints")

structuredTool = runnableLambda.as_tool(TestArgument)

responseString = structuredTool.invoke({"a" : 3, "b" : [1, 2]})

print(responseString)

"""
6
"""

from typing import Dict

from typing import Any

from pydantic import BaseModel

from pydantic import Field

from typing import List

from langchain_core.runnables import RunnableLambda

def test(x : Dict[str, Any]) -> str:

return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

class TestArgument(BaseModel):

"""Apply a function to an integer and list of integers."""

a : int = Field(..., description = "Integer" )

b : List[int] = Field(..., description = "List of ints")

structuredTool = runnableLambda.as_tool(TestArgument)

responseString = structuredTool.invoke({"a" : 3, "b" : [1, 2]})

print(responseString)

"""

▶ requirements.txt


aiohappyeyeballs==2.4.3
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
async-timeout==4.0.3
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
exceptiongroup==1.2.2
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.5
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-core==0.3.7
langchain-text-splitters==0.3.0
langsmith==0.1.129
multidict==6.1.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic_core==2.23.4
PyYAML==6.0.2
requests==2.32.3
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.3

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

async-timeout==4.0.3

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

exceptiongroup==1.2.2

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.5

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-core==0.3.7

langchain-text-splitters==0.3.0

langsmith==0.1.129

multidict==6.1.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic_core==2.23.4

PyYAML==6.0.2

requests==2.32.3

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install

[PYTHON/LANGCHAIN] RunnableLambda 클래스 : as_tool 메소드에서 arg_types 인자를 사용해 도구 인자 타입 설정하기

■ RunnableLambda 클래스의 as_tool 메소드에서 arg_types 인자를 사용해 도구 인자 타입을 설정하는 방법을 보여준다. ▶ main.py


from typing                   import Dict
from typing                   import Any
from typing                   import List
from langchain_core.runnables import RunnableLambda

def test(x : Dict[str, Any]) -> str:
    return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

structuredTool = runnableLambda.as_tool(
    name        = "My tool",
    description = "Explanation of when to use tool.",
    arg_types   = {"a" : int, "b" : List[int]}
)

responseString = structuredTool.invoke({"a" : 3, "b" : [1, 2]})

print(responseString)

"""
6
"""

from typing import Dict

from typing import Any

from typing import List

from langchain_core.runnables import RunnableLambda

def test(x : Dict[str, Any]) -> str:

return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

structuredTool = runnableLambda.as_tool(

name = "My tool",

description = "Explanation of when to use tool.",

arg_types = {"a" : int, "b" : List[int]}

)

responseString = structuredTool.invoke({"a" : 3, "b" : [1, 2]})

print(responseString)

"""

▶ requirements.txt


aiohappyeyeballs==2.4.3
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
async-timeout==4.0.3
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
exceptiongroup==1.2.2
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.5
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-core==0.3.7
langchain-text-splitters==0.3.0
langsmith==0.1.129
multidict==6.1.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic_core==2.23.4
PyYAML==6.0.2
requests==2.32.3
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.3

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

async-timeout==4.0.3

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

exceptiongroup==1.2.2

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.5

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-core==0.3.7

langchain-text-splitters==0.3.0

langsmith==0.1.129

multidict==6.1.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic_core==2.23.4

PyYAML==6.0.2

requests==2.32.3

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip

[PYTHON/LANGCHAIN] StructuredTool 클래스 : invoke 메소드를 사용해 도구 함수 호출하기

■ StructuredTool 클래스의 invoke 메소드를 사용해 도구 함수를 호출하는 방법을 보여준다. ▶ main.py


from typing_extensions        import TypedDict
from typing                   import List
from langchain_core.runnables import RunnableLambda

class Argument(TypedDict):
    a : int
    b : List[int]

def test(x : Argument) -> str:
    return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

structuredTool = runnableLambda.as_tool(
    name        = "My tool",
    description = "Explanation of when to use tool.",
)

responseString = structuredTool.invoke({"a" : 3, "b" : [1, 2]})

print(responseString)

"""
6
"""

from typing_extensions import TypedDict

from typing import List

from langchain_core.runnables import RunnableLambda

class Argument(TypedDict):

a : int

b : List[int]

def test(x : Argument) -> str:

return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

structuredTool = runnableLambda.as_tool(

name = "My tool",

description = "Explanation of when to use tool.",

)

responseString = structuredTool.invoke({"a" : 3, "b" : [1, 2]})

print(responseString)

"""

▶ requirements.txt


aiohappyeyeballs==2.4.3
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
async-timeout==4.0.3
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
exceptiongroup==1.2.2
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.5
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-core==0.3.7
langchain-text-splitters==0.3.0
langsmith==0.1.129
multidict==6.1.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic_core==2.23.4
PyYAML==6.0.2
requests==2.32.3
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.3

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

async-timeout==4.0.3

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

exceptiongroup==1.2.2

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.5

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-core==0.3.7

langchain-text-splitters==0.3.0

langsmith==0.1.129

multidict==6.1.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic_core==2.23.4

PyYAML==6.0.2

requests==2.32.3

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install langchain 명령을

[PYTHON/LANGCHAIN] StructuredTool 클래스 : description/args_schema 속성 사용하기

■ StructuredTool 클래스의 description/args_schema 속성을 사용하는 방법을 보여준다. ▶ main.py


from typing_extensions        import TypedDict
from typing                   import List
from langchain_core.runnables import RunnableLambda

class Argument(TypedDict):
    a : int
    b : List[int]

def test(x : Argument) -> str:
    return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

structuredTool = runnableLambda.as_tool(
    name        = "My tool",
    description = "Explanation of when to use tool.",
)

print(structuredTool.description)
print(structuredTool.args_schema.schema())

"""
Explanation of when to use tool.

{
    'properties' : {
        'a' : {
            'title' : 'A',
            'type'  : 'integer'
        },
        'b' : {
            'items' : {'type' : 'integer'},
            'title' : 'B',
            'type'  : 'array'
        }
    },
    'required' : ['a', 'b'],
    'title'    : 'My tool',
    'type'     : 'object'
}
"""

from typing_extensions import TypedDict

from typing import List

from langchain_core.runnables import RunnableLambda

class Argument(TypedDict):

a : int

b : List[int]

def test(x : Argument) -> str:

return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

structuredTool = runnableLambda.as_tool(

name = "My tool",

description = "Explanation of when to use tool.",

)

print(structuredTool.description)

print(structuredTool.args_schema.schema())

"""

Explanation of when to use tool.

{

'properties' : {

'a' : {

'title' : 'A',

'type' : 'integer'

'b' : {

'items' : {'type' : 'integer'},

'title' : 'B',

'type' : 'array'

}

'required' : ['a', 'b'],

'title' : 'My tool',

'type' : 'object'

}

"""

▶ requirements.txt


aiohappyeyeballs==2.4.3
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
async-timeout==4.0.3
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
exceptiongroup==1.2.2
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.5
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-core==0.3.7
langchain-text-splitters==0.3.0
langsmith==0.1.129
multidict==6.1.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic_core==2.23.4
PyYAML==6.0.2
requests==2.32.3
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.3

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

async-timeout==4.0.3

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

exceptiongroup==1.2.2

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.5

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-core==0.3.7

langchain-text-splitters==0.3.0

langsmith==0.1.129

multidict==6.1.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic_core==2.23.4

PyYAML==6.0.2

requests==2.32.3

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install langchain 명령을 실행했다.

[PYTHON/LANGCHAIN] RunnableLambda 클래스 : as_tool 메소드를 사용해 StructuredTool 객체 만들기

■ RunnableLambda 클래스의 as_tool 메소드를 사용해 StructuredTool 객체를 만드는 방법을 보여준다. ▶ main.py


from typing_extensions        import TypedDict
from typing                   import List
from langchain_core.runnables import RunnableLambda

class Argument(TypedDict):
    a : int
    b : List[int]

def test(x : Argument) -> str:
    return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

structuredTool = runnableLambda.as_tool(
    name        = "My tool",
    description = "Explanation of when to use tool.",
)

from typing_extensions import TypedDict

from typing import List

from langchain_core.runnables import RunnableLambda

class Argument(TypedDict):

a : int

b : List[int]

def test(x : Argument) -> str:

return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

structuredTool = runnableLambda.as_tool(

name = "My tool",

description = "Explanation of when to use tool.",

)

▶ requirements.txt


aiohappyeyeballs==2.4.3
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
async-timeout==4.0.3
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
exceptiongroup==1.2.2
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.5
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-core==0.3.7
langchain-text-splitters==0.3.0
langsmith==0.1.129
multidict==6.1.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic_core==2.23.4
PyYAML==6.0.2
requests==2.32.3
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.3

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

async-timeout==4.0.3

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

exceptiongroup==1.2.2

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.5

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-core==0.3.7

langchain-text-splitters==0.3.0

langsmith==0.1.129

multidict==6.1.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic_core==2.23.4

PyYAML==6.0.2

requests==2.32.3

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install langchain 명령을

[PYTHON/LANGCHAIN] RunnableLambda 클래스 : 생성자를 사용해 RunnableLambda 객체 만들기

■ RunnableLambda 클래스의 생성자를 사용해 RunnableLambda 객체를 만드는 방법을 보여준다. ▶ main.py


from typing_extensions        import TypedDict
from typing                   import List
from langchain_core.runnables import RunnableLambda

class Argument(TypedDict):
    a : int
    b : List[int]

def test(x : Argument) -> str:
    return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

from typing_extensions import TypedDict

from typing import List

from langchain_core.runnables import RunnableLambda

class Argument(TypedDict):

a : int

b : List[int]

def test(x : Argument) -> str:

return str(x["a"] * max(x["b"]))

runnableLambda = RunnableLambda(test)

▶ requirements.txt


aiohappyeyeballs==2.4.3
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
async-timeout==4.0.3
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
exceptiongroup==1.2.2
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.5
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-core==0.3.7
langchain-text-splitters==0.3.0
langsmith==0.1.129
multidict==6.1.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic_core==2.23.4
PyYAML==6.0.2
requests==2.32.3
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.3

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

async-timeout==4.0.3

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

exceptiongroup==1.2.2

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.5

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-core==0.3.7

langchain-text-splitters==0.3.0

langsmith==0.1.129

multidict==6.1.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic_core==2.23.4

PyYAML==6.0.2

requests==2.32.3

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install langchain 명령을 실행했다.

[PYTHON/LANGCHAIN] BaseTool 클래스 : inovoke 메소드를 사용해 컨텐트/아티펙트 구하기

■ BaseTool 클래스의 inovoke 메소드를 사용해 컨텐트/아티펙트를 구하는 방법을 보여준다. ▶ main.py


import random

from langchain_core.tools import BaseTool
from typing               import Tuple
from typing               import List

class GenerateRandomFloatValueTool(BaseTool):
    name            : str = "GenerateRandomFloatValueTool"
    description     : str = "Generate size random floats in the range [minimum, maximum]."
    response_format : str = "content_and_artifact"

    digitCount : int = 2

    def _run(self, minimum : float, maximum : float, count : int) -> Tuple[str, List[float]]:
        range_ = maximum - minimum
        array = [
            round(minimum + (range_ * random.random()), ndigits = self.digitCount)
            for _ in range(count)
        ]
        content = f"Generated {count} floats in [{minimum}, {maximum}], rounded to {self.digitCount} decimals."
        return content, array

generateRandomFloatValueTool = GenerateRandomFloatValueTool(digitCount = 4)

responseToolMessage = generateRandomFloatValueTool.invoke(
    {
        "name" : "generateRandomFloatValueTool",
        "args" : {"minimum" : 0.1, "maximum" : 3.3333, "count" : 3},
        "id"   : "123",      # required
        "type" : "tool_call" # required
    }
)

print(responseToolMessage)

"""
content='Generated 3 floats in [0.1, 3.3333], rounded to 4 decimals.' name='GenerateRandomFloatValueTool' tool_call_id='123' artifact=[1.5382, 2.6612, 2.3324]
"""

import random

from langchain_core.tools import BaseTool

from typing import Tuple

from typing import List

class GenerateRandomFloatValueTool(BaseTool):

name : str = "GenerateRandomFloatValueTool"

description : str = "Generate size random floats in the range [minimum, maximum]."

response_format : str = "content_and_artifact"

digitCount : int = 2

def _run(self, minimum : float, maximum : float, count : int) -> Tuple[str, List[float]]:

range_ = maximum - minimum

array = [

round(minimum + (range_ * random.random()), ndigits = self.digitCount)

for _ in range(count)

]

content = f"Generated {count} floats in [{minimum}, {maximum}], rounded to {self.digitCount} decimals."

return content, array

generateRandomFloatValueTool = GenerateRandomFloatValueTool(digitCount = 4)

responseToolMessage = generateRandomFloatValueTool.invoke(

{

"name" : "generateRandomFloatValueTool",

"args" : {"minimum" : 0.1, "maximum" : 3.3333, "count" : 3},

"id" : "123", # required

"type" : "tool_call" # required

}

)

print(responseToolMessage)

"""

content='Generated 3 floats in [0.1, 3.3333], rounded to 4 decimals.' name='GenerateRandomFloatValueTool' tool_call_id='123' artifact=[1.5382, 2.6612, 2.3324]

"""

▶ requirements.txt


aiohappyeyeballs==2.4.2
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
async-timeout==4.0.3
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
exceptiongroup==1.2.2
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.5
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-core==0.3.6
langchain-text-splitters==0.3.0
langsmith==0.1.129
multidict==6.1.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic_core==2.23.4
PyYAML==6.0.2
requests==2.32.3
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.2

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

async-timeout==4.0.3

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

exceptiongroup==1.2.2

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.5

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-core==0.3.6

langchain-text-splitters==0.3.0

langsmith==0.1.129

multidict==6.1.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic_core==2.23.4

PyYAML==6.0.2

requests==2.32.3

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install langchain 명령을 실행했다.

[PYTHON/LANGCHAIN] BaseTool 클래스 : inovoke 메소드를 사용해 컨텐트 구하기

■ BaseTool 클래스의 inovoke 메소드를 사용해 컨텐트를 구하는 방법을 보여준다. ▶ main.py


import random

from langchain_core.tools import BaseTool
from typing               import Tuple
from typing               import List

class GenerateRandomFloatValueTool(BaseTool):
    name            : str = "GenerateRandomFloatValueTool"
    description     : str = "Generate size random floats in the range [minimum, maximum]."
    response_format : str = "content_and_artifact"

    digitCount : int = 2

    def _run(self, minimum : float, maximum : float, count : int) -> Tuple[str, List[float]]:
        range_ = maximum - minimum
        array = [
            round(minimum + (range_ * random.random()), ndigits = self.digitCount)
            for _ in range(count)
        ]
        content = f"Generated {count} floats in [{minimum}, {maximum}], rounded to {self.digitCount} decimals."
        return content, array

generateRandomFloatValueTool = GenerateRandomFloatValueTool(digitCount = 4)

responseString = generateRandomFloatValueTool.invoke({"minimum" : 0.1, "maximum" : 3.3333, "count" : 3})

print(responseString)

"""
Generated 3 floats in [0.1, 3.3333], rounded to 4 decimals.
"""

import random

from langchain_core.tools import BaseTool

from typing import Tuple

from typing import List

class GenerateRandomFloatValueTool(BaseTool):

name : str = "GenerateRandomFloatValueTool"

description : str = "Generate size random floats in the range [minimum, maximum]."

response_format : str = "content_and_artifact"

digitCount : int = 2

def _run(self, minimum : float, maximum : float, count : int) -> Tuple[str, List[float]]:

range_ = maximum - minimum

array = [

round(minimum + (range_ * random.random()), ndigits = self.digitCount)

for _ in range(count)

]

content = f"Generated {count} floats in [{minimum}, {maximum}], rounded to {self.digitCount} decimals."

return content, array

generateRandomFloatValueTool = GenerateRandomFloatValueTool(digitCount = 4)

responseString = generateRandomFloatValueTool.invoke({"minimum" : 0.1, "maximum" : 3.3333, "count" : 3})

print(responseString)

"""

Generated 3 floats in [0.1, 3.3333], rounded to 4 decimals.

"""

▶ requirements.txt


aiohappyeyeballs==2.4.2
aiohttp==3.10.8
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.6.0
async-timeout==4.0.3
attrs==24.2.0
certifi==2024.8.30
charset-normalizer==3.3.2
exceptiongroup==1.2.2
frozenlist==1.4.1
greenlet==3.1.1
h11==0.14.0
httpcore==1.0.5
httpx==0.27.2
idna==3.10
jsonpatch==1.33
jsonpointer==3.0.0
langchain==0.3.1
langchain-core==0.3.6
langchain-text-splitters==0.3.0
langsmith==0.1.129
multidict==6.1.0
numpy==1.26.4
orjson==3.10.7
packaging==24.1
pydantic==2.9.2
pydantic_core==2.23.4
PyYAML==6.0.2
requests==2.32.3
sniffio==1.3.1
SQLAlchemy==2.0.35
tenacity==8.5.0
typing_extensions==4.12.2
urllib3==2.2.3
yarl==1.13.1

aiohappyeyeballs==2.4.2

aiohttp==3.10.8

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.6.0

async-timeout==4.0.3

attrs==24.2.0

certifi==2024.8.30

charset-normalizer==3.3.2

exceptiongroup==1.2.2

frozenlist==1.4.1

greenlet==3.1.1

h11==0.14.0

httpcore==1.0.5

httpx==0.27.2

idna==3.10

jsonpatch==1.33

jsonpointer==3.0.0

langchain==0.3.1

langchain-core==0.3.6

langchain-text-splitters==0.3.0

langsmith==0.1.129

multidict==6.1.0

numpy==1.26.4

orjson==3.10.7

packaging==24.1

pydantic==2.9.2

pydantic_core==2.23.4

PyYAML==6.0.2

requests==2.32.3

sniffio==1.3.1

SQLAlchemy==2.0.35

tenacity==8.5.0

typing_extensions==4.12.2

urllib3==2.2.3

yarl==1.13.1

※ pip install langchain 명령을 실행했다.

[PYTHON/LANGCHAIN] StructuredTool 클래스 : invoke 메소드에서 AIMessage 객체의 tool_calls 속성 값을 사용해 컨텐트 구하기

■ StructuredTool 클래스의 invoke 메소드에서 AIMessage 객체의 tool_calls 속성 값을 사용해 컨텐트를 구하는 방법을 보여준다. ※ OPENAI_API_KEY 환경 변수 값은 .env 파일에

[PYTHON/LANGCHAIN] StructuredTool 클래스 : invoke 메소드에서 AIMessage 객체의 tool_calls 속성값을 사용해 ToolMessage 객체 구하기

■ StructuredTool 클래스의 invoke 메소드에서 AIMessage 객체의 tool_calls 속성 값을 사용해 ToolMessage 객체를 구하는 방법을 보여준다. ※ OPENAI_API_KEY 환경 변수 값은 .env

[PYTHON/LANGCHAIN] RunnableBindable 클래스 : invoke 메소드를 사용해 모델에서 도구 호출하기

■ RunnableBindable 클래스의 invoke 메소드를 사용해 모델에서 도구를 호출하는 방법을 보여준다. ※ OPENAI_API_KEY 환경 변수 값은 .env 파일에 정의한다. ▶ main.py


import random

from dotenv               import load_dotenv
from langchain_core.tools import tool
from typing               import Tuple
from typing               import List
from langchain_openai     import ChatOpenAI

load_dotenv()

@tool(response_format = "content_and_artifact")
def generateRandomIntegerValues(minimumValue : int, maximumValue : int, count : int) -> Tuple[str, List[int]]:
    """Generate size random ints in the range [minimumValue, maximumValue]."""
    array = [random.randint(minimumValue, maximumValue) for _ in range(count)]
    content = f"Successfully generated array of {count} random ints in [{minimumValue}, {maximumValue}]."
    return content, array

chatOpenAI = ChatOpenAI(model = "gpt-4o-mini")

runnableBindable = chatOpenAI.bind_tools([generateRandomIntegerValues])

responseAIMessage = runnableBindable.invoke("generate 6 positive ints less than 25")

print(responseAIMessage.tool_calls)

"""
[{'name': 'generateRandomIntegerValues', 'args': {'minimumValue': 1, 'maximumValue': 24, 'count': 6}, 'id': 'call_F9ZZFBIsd6HXz30bb0yZTWgl', 'type': 'tool_call'}]
"""

import random

from dotenv import load_dotenv

from langchain_core.tools import tool

from typing import Tuple

from typing import List

from langchain_openai import ChatOpenAI

load_dotenv()

@tool(response_format = "content_and_artifact")

def generateRandomIntegerValues(minimumValue : int, maximumValue : int, count : int) -> Tuple[str, List[int]]:

"""Generate size random ints in the range [minimumValue, maximumValue]."""

array = [random.randint(minimumValue, maximumValue) for _ in range(count)]

content = f"Successfully generated array of {count} random ints in [{minimumValue}, {maximumValue}]."

return content, array

chatOpenAI = ChatOpenAI(model = "gpt-4o-mini")

runnableBindable = chatOpenAI.bind_tools([generateRandomIntegerValues])

responseAIMessage = runnableBindable.invoke("generate 6 positive ints less than 25")

print(responseAIMessage.tool_calls)

"""

[{'name': 'generateRandomIntegerValues', 'args': {'minimumValue': 1, 'maximumValue': 24, 'count': 6}, 'id': 'call_F9ZZFBIsd6HXz30bb0yZTWgl', 'type': 'tool_call'}]

"""