64 lines
2.1 KiB
Python
64 lines
2.1 KiB
Python
import pytest
|
|
|
|
from app.modules.rag.intent_router_v2.analysis.normalization import QueryNormalizer
|
|
|
|
pytestmark = pytest.mark.intent_router
|
|
|
|
|
|
def test_query_normalizer_collapses_whitespace() -> None:
|
|
normalizer = QueryNormalizer()
|
|
|
|
normalized = normalizer.normalize(" Объясни как работает \n класс X ")
|
|
|
|
assert normalized == "Объясни как работает класс X"
|
|
|
|
|
|
def test_query_normalizer_canonicalizes_quotes() -> None:
|
|
normalizer = QueryNormalizer()
|
|
|
|
normalized = normalizer.normalize('Уточни «текст» и “текст”')
|
|
|
|
assert normalized == 'Уточни "текст" и "текст"'
|
|
|
|
|
|
def test_query_normalizer_preserves_backticks_verbatim() -> None:
|
|
normalizer = QueryNormalizer()
|
|
|
|
normalized = normalizer.normalize("Уточни по коду `def build(x):` ")
|
|
|
|
assert normalized == "Уточни по коду `def build(x):`"
|
|
|
|
|
|
def test_query_normalizer_preserves_latin_and_cyrillic_file_paths() -> None:
|
|
normalizer = QueryNormalizer()
|
|
|
|
normalized = normalizer.normalize("Сверь app/core/config.py и «docs/руководство.md»")
|
|
|
|
assert "app/core/config.py" in normalized
|
|
assert "docs/руководство.md" in normalized
|
|
assert "config. py" not in normalized
|
|
assert "руководство. md" not in normalized
|
|
|
|
|
|
def test_query_normalizer_punctuation_spacing_does_not_break_extensions() -> None:
|
|
normalizer = QueryNormalizer()
|
|
|
|
normalized = normalizer.normalize("Проверь docs/spec.md , затем app/main.py !")
|
|
|
|
assert "docs/spec.md" in normalized
|
|
assert "app/main.py" in normalized
|
|
assert "spec. md" not in normalized
|
|
assert "main. py" not in normalized
|
|
|
|
|
|
def test_query_normalizer_idempotent_and_without_enrichment() -> None:
|
|
normalizer = QueryNormalizer()
|
|
raw = ' Прочитай «README.md» и docs/spec.md '
|
|
|
|
once = normalizer.normalize(raw)
|
|
twice = normalizer.normalize(once)
|
|
|
|
assert twice == once
|
|
assert "documentation" not in once.lower()
|
|
assert "class" not in once.lower()
|