import pytest

from app.engine.source_extraction import chunk_with_offsets


def test_chunk_offsets_and_overlap_are_exact():
    text = "가" * 1200
    chunks = list(chunk_with_offsets(text, size=500, stride=250))
    assert chunks[0] == (0, 500, text[0:500])
    assert chunks[1] == (250, 750, text[250:750])
    assert chunks[-1][1] == len(text)


def test_chunk_parameters_must_be_positive():
    with pytest.raises(ValueError):
        list(chunk_with_offsets("본문", size=0, stride=1))