#!/usr/bin/env python3
"""
Tests de non-régression sur des layouts d'identifiants vus en documents réels.
"""
from anonymizer_core_refactored_onnx import (
    RE_SCAN_FILENAME_ARTIFACT,
    anonymise_document_regex,
    load_dictionaries,
)


def test_bacterio_multiline_venue_number_before_ipp_is_masked():
    cfg = load_dictionaries(None)
    text = (
        "Diffusé le :\n"
        "à\n"
        "N° venue :\n"
        "31/07/1973\n"
        "VAN DE GRAAF\n"
        "23176885\n"
        "IPP :\n"
        "2300201230\n"
    )

    anon = anonymise_document_regex([text], [[]], cfg)

    assert "23176885" not in anon.text_out
    assert "[NDA]" in anon.text_out
    assert any(h.kind == "NDA" and h.original == "23176885" for h in anon.audit)


def test_scan_filename_artifact_suffix_is_masked():
    cfg = load_dictionaries(None)
    text = (
        "IPP:\n"
        "16014215\n"
        "Document scanné non\n"
        "éditable pour patient (dont\ngénétique)\n"
        "EXT2-16014215-2300249096.TIF\n"
    )

    anon = anonymise_document_regex([text], [[]], cfg)

    assert RE_SCAN_FILENAME_ARTIFACT.search("EXT2-[IPP]-2300249096.TIF") is not None
    assert "2300249096" not in anon.text_out
    assert "EXT2-[IPP]-[DOSSIER].TIF" in anon.text_out