langchain/libs/experimental/langchain_experimental/agents/__init__.py

"""**Agent** is a class that uses an LLM to choose 
a sequence of actions to take.

In Chains, a sequence of actions is hardcoded. In Agents,
a language model is used as a reasoning engine to determine which actions
to take and in which order.

Agents select and use **Tools** and **Toolkits** for actions.
"""  # noqa: E501
from langchain_experimental.agents.agent_toolkits import (
    create_csv_agent,
    create_pandas_dataframe_agent,
    create_spark_dataframe_agent,
    create_xorbits_agent,
)

__all__ = [
    "create_csv_agent",
    "create_pandas_dataframe_agent",
    "create_spark_dataframe_agent",
    "create_xorbits_agent",
]