langchain/langchain/agents/agent_toolkits/pandas/base.py

69 lines
2.4 KiB
Python

"""Agent for working with pandas objects."""
from typing import Any, Dict, List, Optional
from langchain.agents.agent import AgentExecutor
from langchain.agents.agent_toolkits.pandas.prompt import PREFIX, SUFFIX
from langchain.agents.mrkl.base import ZeroShotAgent
from langchain.base_language import BaseLanguageModel
from langchain.callbacks.base import BaseCallbackManager
from langchain.chains.llm import LLMChain
from langchain.tools.python.tool import PythonAstREPLTool
def create_pandas_dataframe_agent(
llm: BaseLanguageModel,
df: Any,
callback_manager: Optional[BaseCallbackManager] = None,
prefix: str = PREFIX,
suffix: str = SUFFIX,
input_variables: Optional[List[str]] = None,
verbose: bool = False,
return_intermediate_steps: bool = False,
max_iterations: Optional[int] = 15,
max_execution_time: Optional[float] = None,
early_stopping_method: str = "force",
agent_executor_kwargs: Optional[Dict[str, Any]] = None,
**kwargs: Dict[str, Any],
) -> AgentExecutor:
"""Construct a pandas agent from an LLM and dataframe."""
try:
import pandas as pd
except ImportError:
raise ValueError(
"pandas package not found, please install with `pip install pandas`"
)
if not isinstance(df, pd.DataFrame):
raise ValueError(f"Expected pandas object, got {type(df)}")
if input_variables is None:
input_variables = ["df", "input", "agent_scratchpad"]
tools = [PythonAstREPLTool(locals={"df": df})]
prompt = ZeroShotAgent.create_prompt(
tools, prefix=prefix, suffix=suffix, input_variables=input_variables
)
partial_prompt = prompt.partial(df=str(df.head().to_markdown()))
llm_chain = LLMChain(
llm=llm,
prompt=partial_prompt,
callback_manager=callback_manager,
)
tool_names = [tool.name for tool in tools]
agent = ZeroShotAgent(
llm_chain=llm_chain,
allowed_tools=tool_names,
callback_manager=callback_manager,
**kwargs,
)
return AgentExecutor.from_agent_and_tools(
agent=agent,
tools=tools,
callback_manager=callback_manager,
verbose=verbose,
return_intermediate_steps=return_intermediate_steps,
max_iterations=max_iterations,
max_execution_time=max_execution_time,
early_stopping_method=early_stopping_method,
**(agent_executor_kwargs or {}),
)