langchain/tests/integration_tests/agent/test_pandas_agent.py
Mike Wang c3044b1bf0
[test] Add integration_test for PandasAgent (#4056)
- confirm creation
- confirm functionality with a simple dimension check.

The test now is calling OpenAI API directly, but learning from
@vowelparrot that we’re caching the requests, so that it’s not that
expensive. I also found we’re calling OpenAI api in other integration
tests. Please lmk if there is any concern of real external API calls. I
can alternatively make a fake LLM for this test. Thanks
2023-05-05 14:49:02 -07:00

31 lines
925 B
Python

import re
import numpy as np
import pytest
from pandas import DataFrame
from langchain.agents import create_pandas_dataframe_agent
from langchain.agents.agent import AgentExecutor
from langchain.llms import OpenAI
@pytest.fixture(scope="module")
def df() -> DataFrame:
random_data = np.random.rand(4, 4)
df = DataFrame(random_data, columns=["name", "age", "food", "sport"])
return df
def test_pandas_agent_creation(df: DataFrame) -> None:
agent = create_pandas_dataframe_agent(OpenAI(temperature=0), df)
assert isinstance(agent, AgentExecutor)
def test_data_reading(df: DataFrame) -> None:
agent = create_pandas_dataframe_agent(OpenAI(temperature=0), df)
assert isinstance(agent, AgentExecutor)
response = agent.run("how many rows in df? Give me a number.")
result = re.search(rf".*({df.shape[0]}).*", response)
assert result is not None
assert result.group(1) is not None