test_md_header_text_splitter_fenced_code_block_interleaved() — langchain Function Reference

Source Code

libs/text-splitters/tests/unit_tests/test_text_splitters.py lines 1491–1524

def test_md_header_text_splitter_fenced_code_block_interleaved(
    fence: str, other_fence: str
) -> None:
    """Test markdown splitter by header: Interleaved fenced code block."""
    markdown_document = (
        "# This is a Header\n\n"
        f"{fence}\n"
        "foo\n"
        "# Not a header\n"
        f"{other_fence}\n"
        "# Not a header\n"
        f"{fence}"
    )

    headers_to_split_on = [
        ("#", "Header 1"),
        ("##", "Header 2"),
    ]

    markdown_splitter = MarkdownHeaderTextSplitter(
        headers_to_split_on=headers_to_split_on,
    )
    output = markdown_splitter.split_text(markdown_document)

    expected_output = [
        Document(
            page_content=(
                f"{fence}\nfoo\n# Not a header\n{other_fence}\n# Not a header\n{fence}"
            ),
            metadata={"Header 1": "This is a Header"},
        ),
    ]

    assert output == expected_output