tests/test_transcriber_normalization.py

import re

from app.core.config import get_settings
from app.services.transcriber import _normalise_transcript_text


def _get_test_replacements():
    replacements = get_settings().TRANSCRIBE_REPLACEMENTS
    return tuple(
        (re.compile(pattern, re.IGNORECASE), replacement)
        for pattern, replacement in replacements.items()
    )


def test_normalise_walk_experience_to_work_experience() -> None:
    query = "uh what is his walk experience in a professional setting"
    replacements = _get_test_replacements()
    assert _normalise_transcript_text(query, replacements) == "what is his work experience in a professional setting"


def test_normalise_text_stack_to_tech_stack() -> None:
    replacements = _get_test_replacements()
    assert _normalise_transcript_text("what text stack does he use", replacements) == "what tech stack does he use"


def test_keeps_clean_transcript_unchanged() -> None:
    original = "What technologies and skills does he work with?"
    replacements = _get_test_replacements()
    assert _normalise_transcript_text(original, replacements) == original