如何配置运行时链内部结构

先决条件

本指南假定您熟悉以下概念

• Runnable 接口
• 链接 runnables
• 绑定运行时参数

有时您可能想要在链中试验，甚至向最终用户公开多种不同的方法来做事。这可能包括调整温度等参数，甚至更换一个模型。为了使这种体验尽可能简单，我们定义了两种方法。

• configurable_fields 方法。 这允许您配置 runnable 的特定字段。
  • 这与 runnables 上的 .bind 方法相关，但允许您在运行时为链中的给定步骤指定参数，而不是预先指定它们。
• configurable_alternatives 方法。 使用此方法，您可以列出在运行时可以设置的任何特定 runnable 的替代方案，并将它们交换为那些指定的替代方案。

可配置字段

让我们通过一个示例，该示例在运行时配置聊天模型字段（如温度）

%pip install --upgrade --quiet langchain langchain-openai

import os
from getpass import getpass

if "OPENAI_API_KEY" not in os.environ:
    os.environ["OPENAI_API_KEY"] = getpass()

配置聊天模型上的字段

如果使用 init_chat_model 创建聊天模型，则可以在构造函数中指定可配置字段

from langchain.chat_models import init_chat_model

llm = init_chat_model(
    "openai:gpt-4o-mini",
    configurable_fields=("temperature",),
)

API 参考:init_chat_model

然后，您可以使用 .with_config 在运行时设置参数

response = llm.with_config({"temperature": 0}).invoke("Hello")
print(response.content)

Hello! How can I assist you today?

提示

除了像温度这样的调用参数外，以这种方式配置字段还可以扩展到客户端和其他属性。

与工具一起使用

当绑定工具时，此方法也适用

from langchain_core.tools import tool


@tool
def get_weather(location: str):
    """Get the weather."""
    return "It's sunny."


llm_with_tools = llm.bind_tools([get_weather])
response = llm_with_tools.with_config({"temperature": 0}).invoke(
    "What's the weather in SF?"
)
response.tool_calls

API 参考:tool

[{'name': 'get_weather',
  'args': {'location': 'San Francisco'},
  'id': 'call_B93EttzlGyYUhzbIIiMcl3bE',
  'type': 'tool_call'}]

除了 .with_config 之外，我们现在可以在直接传递配置时包含参数。请参见下面的示例，在该示例中，我们允许在 langgraph agent 内部配置底层模型温度

! pip install --upgrade langgraph

from langgraph.prebuilt import create_react_agent

agent = create_react_agent(llm, [get_weather])

response = agent.invoke(
    {"messages": "What's the weather in Boston?"},
    {"configurable": {"temperature": 0}},
)

API 参考:create_react_agent

在任意 Runnables 上配置字段

您还可以在任意 Runnables 上使用 .configurable_fields 方法，如下所示

from langchain_core.prompts import PromptTemplate
from langchain_core.runnables import ConfigurableField
from langchain_openai import ChatOpenAI

model = ChatOpenAI(temperature=0).configurable_fields(
    temperature=ConfigurableField(
        id="llm_temperature",
        name="LLM Temperature",
        description="The temperature of the LLM",
    )
)

model.invoke("pick a random number")

API 参考:PromptTemplate | ConfigurableField | ChatOpenAI

AIMessage(content='17', response_metadata={'token_usage': {'completion_tokens': 1, 'prompt_tokens': 11, 'total_tokens': 12}, 'model_name': 'gpt-3.5-turbo', 'system_fingerprint': 'fp_c2295e73ad', 'finish_reason': 'stop', 'logprobs': None}, id='run-ba26a0da-0a69-4533-ab7f-21178a73d303-0')

上面，我们将 temperature 定义为 ConfigurableField，我们可以在运行时设置它。为此，我们使用 with_config 方法，如下所示

model.with_config(configurable={"llm_temperature": 0.9}).invoke("pick a random number")

AIMessage(content='12', response_metadata={'token_usage': {'completion_tokens': 1, 'prompt_tokens': 11, 'total_tokens': 12}, 'model_name': 'gpt-3.5-turbo', 'system_fingerprint': 'fp_c2295e73ad', 'finish_reason': 'stop', 'logprobs': None}, id='run-ba8422ad-be77-4cb1-ac45-ad0aae74e3d9-0')

请注意，字典中传递的 llm_temperature 条目的键与 ConfigurableField 的 id 相同。

我们也可以这样做，仅影响链中的一个步骤

prompt = PromptTemplate.from_template("Pick a random number above {x}")
chain = prompt | model

chain.invoke({"x": 0})

AIMessage(content='27', response_metadata={'token_usage': {'completion_tokens': 1, 'prompt_tokens': 14, 'total_tokens': 15}, 'model_name': 'gpt-3.5-turbo', 'system_fingerprint': 'fp_c2295e73ad', 'finish_reason': 'stop', 'logprobs': None}, id='run-ecd4cadd-1b72-4f92-b9a0-15e08091f537-0')

chain.with_config(configurable={"llm_temperature": 0.9}).invoke({"x": 0})

AIMessage(content='35', response_metadata={'token_usage': {'completion_tokens': 1, 'prompt_tokens': 14, 'total_tokens': 15}, 'model_name': 'gpt-3.5-turbo', 'system_fingerprint': 'fp_c2295e73ad', 'finish_reason': 'stop', 'logprobs': None}, id='run-a916602b-3460-46d3-a4a8-7c926ec747c0-0')

使用 HubRunnables

这对于允许切换提示非常有用

from langchain.runnables.hub import HubRunnable

prompt = HubRunnable("rlm/rag-prompt").configurable_fields(
    owner_repo_commit=ConfigurableField(
        id="hub_commit",
        name="Hub Commit",
        description="The Hub commit to pull from",
    )
)

prompt.invoke({"question": "foo", "context": "bar"})

API 参考:HubRunnable

ChatPromptValue(messages=[HumanMessage(content="You are an assistant for question-answering tasks. Use the following pieces of retrieved context to answer the question. If you don't know the answer, just say that you don't know. Use three sentences maximum and keep the answer concise.\nQuestion: foo \nContext: bar \nAnswer:")])

prompt.with_config(configurable={"hub_commit": "rlm/rag-prompt-llama"}).invoke(
    {"question": "foo", "context": "bar"}
)

ChatPromptValue(messages=[HumanMessage(content="[INST]<<SYS>> You are an assistant for question-answering tasks. Use the following pieces of retrieved context to answer the question. If you don't know the answer, just say that you don't know. Use three sentences maximum and keep the answer concise.<</SYS>> \nQuestion: foo \nContext: bar \nAnswer: [/INST]")])

可配置替代方案

configurable_alternatives() 方法允许我们用替代方案替换链中的步骤。 在下面，我们用另一个聊天模型替换一个聊天模型

%pip install --upgrade --quiet langchain-anthropic

import os
from getpass import getpass

if "ANTHROPIC_API_KEY" not in os.environ:
    os.environ["ANTHROPIC_API_KEY"] = getpass()

WARNING: You are using pip version 22.0.4; however, version 24.0 is available.
You should consider upgrading via the '/Users/jacoblee/.pyenv/versions/3.10.5/bin/python -m pip install --upgrade pip' command.
Note: you may need to restart the kernel to use updated packages.

from langchain_anthropic import ChatAnthropic
from langchain_core.prompts import PromptTemplate
from langchain_core.runnables import ConfigurableField
from langchain_openai import ChatOpenAI

llm = ChatAnthropic(
    model="claude-3-haiku-20240307", temperature=0
).configurable_alternatives(
    # This gives this field an id
    # When configuring the end runnable, we can then use this id to configure this field
    ConfigurableField(id="llm"),
    # This sets a default_key.
    # If we specify this key, the default LLM (ChatAnthropic initialized above) will be used
    default_key="anthropic",
    # This adds a new option, with name `openai` that is equal to `ChatOpenAI()`
    openai=ChatOpenAI(),
    # This adds a new option, with name `gpt4` that is equal to `ChatOpenAI(model="gpt-4")`
    gpt4=ChatOpenAI(model="gpt-4"),
    # You can add more configuration options here
)
prompt = PromptTemplate.from_template("Tell me a joke about {topic}")
chain = prompt | llm

# By default it will call Anthropic
chain.invoke({"topic": "bears"})

API 参考:ChatAnthropic | PromptTemplate | ConfigurableField | ChatOpenAI

AIMessage(content="Here's a bear joke for you:\n\nWhy don't bears wear socks? \nBecause they have bear feet!\n\nHow's that? I tried to come up with a simple, silly pun-based joke about bears. Puns and wordplay are a common way to create humorous bear jokes. Let me know if you'd like to hear another one!", response_metadata={'id': 'msg_018edUHh5fUbWdiimhrC3dZD', 'model': 'claude-3-haiku-20240307', 'stop_reason': 'end_turn', 'stop_sequence': None, 'usage': {'input_tokens': 13, 'output_tokens': 80}}, id='run-775bc58c-28d7-4e6b-a268-48fa6661f02f-0')

# We can use `.with_config(configurable={"llm": "openai"})` to specify an llm to use
chain.with_config(configurable={"llm": "openai"}).invoke({"topic": "bears"})

AIMessage(content="Why don't bears like fast food?\n\nBecause they can't catch it!", response_metadata={'token_usage': {'completion_tokens': 15, 'prompt_tokens': 13, 'total_tokens': 28}, 'model_name': 'gpt-3.5-turbo', 'system_fingerprint': 'fp_c2295e73ad', 'finish_reason': 'stop', 'logprobs': None}, id='run-7bdaa992-19c9-4f0d-9a0c-1f326bc992d4-0')

# If we use the `default_key` then it uses the default
chain.with_config(configurable={"llm": "anthropic"}).invoke({"topic": "bears"})

AIMessage(content="Here's a bear joke for you:\n\nWhy don't bears wear socks? \nBecause they have bear feet!\n\nHow's that? I tried to come up with a simple, silly pun-based joke about bears. Puns and wordplay are a common way to create humorous bear jokes. Let me know if you'd like to hear another one!", response_metadata={'id': 'msg_01BZvbmnEPGBtcxRWETCHkct', 'model': 'claude-3-haiku-20240307', 'stop_reason': 'end_turn', 'stop_sequence': None, 'usage': {'input_tokens': 13, 'output_tokens': 80}}, id='run-59b6ee44-a1cd-41b8-a026-28ee67cdd718-0')

使用提示

我们可以做类似的事情，但在提示之间交替

llm = ChatAnthropic(model="claude-3-haiku-20240307", temperature=0)
prompt = PromptTemplate.from_template(
    "Tell me a joke about {topic}"
).configurable_alternatives(
    # This gives this field an id
    # When configuring the end runnable, we can then use this id to configure this field
    ConfigurableField(id="prompt"),
    # This sets a default_key.
    # If we specify this key, the default prompt (asking for a joke, as initialized above) will be used
    default_key="joke",
    # This adds a new option, with name `poem`
    poem=PromptTemplate.from_template("Write a short poem about {topic}"),
    # You can add more configuration options here
)
chain = prompt | llm

# By default it will write a joke
chain.invoke({"topic": "bears"})

AIMessage(content="Here's a bear joke for you:\n\nWhy don't bears wear socks? \nBecause they have bear feet!", response_metadata={'id': 'msg_01DtM1cssjNFZYgeS3gMZ49H', 'model': 'claude-3-haiku-20240307', 'stop_reason': 'end_turn', 'stop_sequence': None, 'usage': {'input_tokens': 13, 'output_tokens': 28}}, id='run-8199af7d-ea31-443d-b064-483693f2e0a1-0')

# We can configure it write a poem
chain.with_config(configurable={"prompt": "poem"}).invoke({"topic": "bears"})

AIMessage(content="Here is a short poem about bears:\n\nMajestic bears, strong and true,\nRoaming the forests, wild and free.\nPowerful paws, fur soft and brown,\nCommanding respect, nature's crown.\n\nForaging for berries, fishing streams,\nProtecting their young, fierce and keen.\nMighty bears, a sight to behold,\nGuardians of the wilderness, untold.\n\nIn the wild they reign supreme,\nEmbodying nature's grand theme.\nBears, a symbol of strength and grace,\nCaptivating all who see their face.", response_metadata={'id': 'msg_01Wck3qPxrjURtutvtodaJFn', 'model': 'claude-3-haiku-20240307', 'stop_reason': 'end_turn', 'stop_sequence': None, 'usage': {'input_tokens': 13, 'output_tokens': 134}}, id='run-69414a1e-51d7-4bec-a307-b34b7d61025e-0')

使用提示和 LLM

我们还可以配置多个事物！这是一个使用提示和 LLM 完成此操作的示例。

llm = ChatAnthropic(
    model="claude-3-haiku-20240307", temperature=0
).configurable_alternatives(
    # This gives this field an id
    # When configuring the end runnable, we can then use this id to configure this field
    ConfigurableField(id="llm"),
    # This sets a default_key.
    # If we specify this key, the default LLM (ChatAnthropic initialized above) will be used
    default_key="anthropic",
    # This adds a new option, with name `openai` that is equal to `ChatOpenAI()`
    openai=ChatOpenAI(),
    # This adds a new option, with name `gpt4` that is equal to `ChatOpenAI(model="gpt-4")`
    gpt4=ChatOpenAI(model="gpt-4"),
    # You can add more configuration options here
)
prompt = PromptTemplate.from_template(
    "Tell me a joke about {topic}"
).configurable_alternatives(
    # This gives this field an id
    # When configuring the end runnable, we can then use this id to configure this field
    ConfigurableField(id="prompt"),
    # This sets a default_key.
    # If we specify this key, the default prompt (asking for a joke, as initialized above) will be used
    default_key="joke",
    # This adds a new option, with name `poem`
    poem=PromptTemplate.from_template("Write a short poem about {topic}"),
    # You can add more configuration options here
)
chain = prompt | llm

# We can configure it write a poem with OpenAI
chain.with_config(configurable={"prompt": "poem", "llm": "openai"}).invoke(
    {"topic": "bears"}
)

AIMessage(content="In the forest deep and wide,\nBears roam with grace and pride.\nWith fur as dark as night,\nThey rule the land with all their might.\n\nIn winter's chill, they hibernate,\nIn spring they emerge, hungry and great.\nWith claws sharp and eyes so keen,\nThey hunt for food, fierce and lean.\n\nBut beneath their tough exterior,\nLies a gentle heart, warm and superior.\nThey love their cubs with all their might,\nProtecting them through day and night.\n\nSo let us admire these majestic creatures,\nIn awe of their strength and features.\nFor in the wild, they reign supreme,\nThe mighty bears, a timeless dream.", response_metadata={'token_usage': {'completion_tokens': 133, 'prompt_tokens': 13, 'total_tokens': 146}, 'model_name': 'gpt-3.5-turbo', 'system_fingerprint': 'fp_c2295e73ad', 'finish_reason': 'stop', 'logprobs': None}, id='run-5eec0b96-d580-49fd-ac4e-e32a0803b49b-0')

# We can always just configure only one if we want
chain.with_config(configurable={"llm": "openai"}).invoke({"topic": "bears"})

AIMessage(content="Why don't bears wear shoes?\n\nBecause they have bear feet!", response_metadata={'token_usage': {'completion_tokens': 13, 'prompt_tokens': 13, 'total_tokens': 26}, 'model_name': 'gpt-3.5-turbo', 'system_fingerprint': 'fp_c2295e73ad', 'finish_reason': 'stop', 'logprobs': None}, id='run-c1b14c9c-4988-49b8-9363-15bfd479973a-0')

保存配置

我们还可以轻松地将配置的链保存为自己的对象

openai_joke = chain.with_config(configurable={"llm": "openai"})

openai_joke.invoke({"topic": "bears"})

AIMessage(content="Why did the bear break up with his girlfriend? \nBecause he couldn't bear the relationship anymore!", response_metadata={'token_usage': {'completion_tokens': 20, 'prompt_tokens': 13, 'total_tokens': 33}, 'model_name': 'gpt-3.5-turbo', 'system_fingerprint': 'fp_c2295e73ad', 'finish_reason': 'stop', 'logprobs': None}, id='run-391ebd55-9137-458b-9a11-97acaff6a892-0')

下一步

您现在知道如何在运行时配置链的内部步骤。

要了解更多信息，请参阅本节中关于 runnables 的其他操作指南，包括

• 使用 .bind() 作为设置 runnable 的运行时参数的更简单方法