"""Test text splitters that require an integration."""

import pytest

from langchain.text_splitter import CharacterTextSplitter


def test_huggingface_type_check() -> None:
    """Test that type checks are done properly on input."""
    with pytest.raises(ValueError):
        CharacterTextSplitter.from_huggingface_tokenizer("foo")


def test_huggingface_tokenizer() -> None:
    """Test text splitter that uses a HuggingFace tokenizer."""
    from transformers import GPT2TokenizerFast

    tokenizer = GPT2TokenizerFast.from_pretrained("gpt2")
    text_splitter = CharacterTextSplitter.from_huggingface_tokenizer(
        tokenizer, separator=" ", chunk_size=1, chunk_overlap=0
    )
    output = text_splitter.split_text("foo bar")
    assert output == ["foo", "bar"]