langchain.ChatAnthropicDataset

kedro_datasets_experimental.langchain.ChatAnthropicDataset

ChatAnthropicDataset(credentials, kwargs=None)

Bases: AbstractDataset[None, ChatAnthropic]

ChatAnthropicDataset loads a ChatAnthropic langchain model.

Example usage for the YAML API

catalog.yml

claude_instant_1:
    type: langchain.ChatAnthropicDataset
    kwargs:
        model: "claude-instant-1"
        temperature: 0.0
    credentials: anthropic

credentials.yml

anthropic:
    anthropic_api_url: <anthropic-api-base>
    anthropic_api_key: <anthropic-api-key>

Example usage for the Python API

from kedro_datasets_experimental.langchain import ChatAnthropicDataset

llm = ChatAnthropicDataset(
    credentials={
        "anthropic_api_url": "xxx",
        "anthropic_api_key": "xxx",
    },
    kwargs={
        "model": "claude-instant-1",
        "temperature": 0.0,
    },
).load()

# See: https://python.langchain.com/docs/integrations/chat/anthropic
llm.invoke("Hello world!")

Parameters:

• credentials (dict[str, str]) –

  must contain anthropic_api_url and anthropic_api_key.

• kwargs (dict[str, Any], default: None ) –

  keyword arguments passed to the ChatAnthropic constructor.

Source code in kedro-datasets/kedro_datasets_experimental/langchain/_anthropic.py

def __init__(self, credentials: dict[str, str], kwargs: dict[str, Any] = None):
    """Constructor.

    Args:
        credentials: must contain `anthropic_api_url` and `anthropic_api_key`.
        kwargs: keyword arguments passed to the ChatAnthropic constructor.
    """
    self.anthropic_api_url = credentials["anthropic_api_url"]
    self.anthropic_api_key = credentials["anthropic_api_key"]
    self.kwargs = kwargs or {}

anthropic_api_key instance-attribute

anthropic_api_key = credentials['anthropic_api_key']

anthropic_api_url instance-attribute

anthropic_api_url = credentials['anthropic_api_url']

kwargs instance-attribute

kwargs = kwargs or {}

_describe

_describe()

Source code in kedro-datasets/kedro_datasets_experimental/langchain/_anthropic.py

def _describe(self) -> dict[str, Any]:
    return {**self.kwargs}

load

load()

Source code in kedro-datasets/kedro_datasets_experimental/langchain/_anthropic.py

def load(self) -> ChatAnthropic:
    return ChatAnthropic(
        anthropic_api_url=self.anthropic_api_url,
        anthropic_api_key=self.anthropic_api_key,
        **self.kwargs,
    )

save

save(data)

Source code in kedro-datasets/kedro_datasets_experimental/langchain/_anthropic.py

def save(self, data: None) -> NoReturn:
    raise DatasetError(f"{self.__class__.__name__} is a read only dataset type")