macro_split/tests/test_detection_and_numbering.py

import pytest
from rdkit import Chem

from macro_lactone_toolkit import (
    AmbiguousMacrolactoneError,
    MacroLactoneAnalyzer,
    MacrolactoneDetectionError,
    MacrolactoneFragmenter,
)

from .helpers import (
    build_ambiguous_smiles,
    build_macrolactone,
    build_non_standard_ring_atom_macrolactone,
    build_overlapping_candidate_macrolactone,
)


@pytest.mark.parametrize("ring_size", [12, 14, 16, 20])
def test_analyzer_detects_supported_ring_sizes(ring_size: int):
    built = build_macrolactone(ring_size)
    analyzer = MacroLactoneAnalyzer()

    assert analyzer.get_valid_ring_sizes(built.smiles) == [ring_size]


def test_analyzer_rejects_non_lactone_macrocycle():
    analyzer = MacroLactoneAnalyzer()

    assert analyzer.get_valid_ring_sizes("C1CCCCCCCCCCC1") == []


@pytest.mark.parametrize("ring_size", [12, 14, 16, 20])
def test_analyzer_classifies_supported_ring_sizes(ring_size: int):
    built = build_macrolactone(ring_size)
    analyzer = MacroLactoneAnalyzer()

    result = analyzer.classify_macrocycle(built.smiles)

    assert result.classification == "standard_macrolactone"
    assert result.ring_size == ring_size
    assert result.primary_reason_code is None
    assert result.primary_reason_message is None
    assert result.all_reason_codes == []
    assert result.all_reason_messages == []
    assert result.candidate_ring_sizes == [ring_size]


def test_analyzer_classifies_ring_heteroatom_as_non_standard():
    built = build_non_standard_ring_atom_macrolactone()
    analyzer = MacroLactoneAnalyzer()

    result = analyzer.classify_macrocycle(built.smiles)

    assert result.classification == "non_standard_macrocycle"
    assert result.ring_size == 16
    assert result.primary_reason_code == "contains_non_carbon_ring_atoms_outside_positions_1_2"
    assert result.primary_reason_message == "Ring positions 3..N contain non-carbon atoms."
    assert result.all_reason_codes == ["contains_non_carbon_ring_atoms_outside_positions_1_2"]
    assert result.candidate_ring_sizes == [16]


def test_analyzer_classifies_overlapping_candidates_as_non_standard():
    built = build_overlapping_candidate_macrolactone()
    analyzer = MacroLactoneAnalyzer()

    result = analyzer.classify_macrocycle(built.smiles)

    assert result.classification == "non_standard_macrocycle"
    assert result.ring_size == 12
    assert result.primary_reason_code == "multiple_overlapping_macrocycle_candidates"
    assert result.primary_reason_message == "Overlapping macrolactone candidate rings were detected."
    assert result.all_reason_codes == ["multiple_overlapping_macrocycle_candidates"]
    assert result.candidate_ring_sizes == [12]


def test_analyzer_classifies_non_lactone_macrocycle():
    analyzer = MacroLactoneAnalyzer()

    result = analyzer.classify_macrocycle("C1CCCCCCCCCCC1")

    assert result.classification == "not_macrolactone"
    assert result.ring_size is None
    assert result.primary_reason_code == "no_lactone_ring_in_12_to_20_range"
    assert result.primary_reason_message == "No 12-20 membered lactone ring was detected."
    assert result.all_reason_codes == ["no_lactone_ring_in_12_to_20_range"]
    assert result.candidate_ring_sizes == []


def test_analyzer_explicit_ring_size_miss_returns_requested_ring_not_found():
    built = build_macrolactone(12)
    analyzer = MacroLactoneAnalyzer()

    result = analyzer.classify_macrocycle(built.smiles, ring_size=16)

    assert result.classification == "not_macrolactone"
    assert result.ring_size is None
    assert result.primary_reason_code == "requested_ring_size_not_found"
    assert result.primary_reason_message == "The requested ring size was not detected as a lactone ring."
    assert result.all_reason_codes == ["requested_ring_size_not_found"]
    assert result.candidate_ring_sizes == []


def test_fragmenter_auto_numbers_ring_with_expected_positions():
    built = build_macrolactone(16, {5: "methyl"})
    result = MacrolactoneFragmenter().number_molecule(built.mol)

    assert result.ring_size == 16
    assert result.position_to_atom == built.position_to_atom
    assert set(result.position_to_atom) == set(range(1, 17))
    assert result.atom_to_position == {
        atom_idx: position for position, atom_idx in built.position_to_atom.items()
    }

    carbonyl_atom = built.mol.GetAtomWithIdx(result.position_to_atom[1])
    assert carbonyl_atom.GetSymbol() == "C"
    assert any(
        bond.GetBondType() == Chem.BondType.DOUBLE and bond.GetOtherAtom(carbonyl_atom).GetSymbol() == "O"
        for bond in carbonyl_atom.GetBonds()
    )

    ester_oxygen = built.mol.GetAtomWithIdx(result.position_to_atom[2])
    assert ester_oxygen.GetSymbol() == "O"


def test_fragmenter_requires_explicit_ring_size_for_ambiguous_molecule():
    ambiguous_smiles = build_ambiguous_smiles()

    with pytest.raises(AmbiguousMacrolactoneError):
        MacrolactoneFragmenter().number_molecule(ambiguous_smiles)


def test_fragmenter_raises_for_missing_macrolactone():
    with pytest.raises(
        MacrolactoneDetectionError,
        match="classification=not_macrolactone primary_reason_code=no_lactone_ring_in_12_to_20_range",
    ):
        MacrolactoneFragmenter().number_molecule("CCO")


def test_fragmenter_rejects_non_standard_macrocycle_with_reason_code():
    built = build_non_standard_ring_atom_macrolactone()

    with pytest.raises(
        MacrolactoneDetectionError,
        match="classification=non_standard_macrocycle "
        "primary_reason_code=contains_non_carbon_ring_atoms_outside_positions_1_2",
    ):
        MacrolactoneFragmenter().number_molecule(built.smiles)


def test_fragmenter_rejects_non_standard_macrocycle_during_fragmentation():
    built = build_overlapping_candidate_macrolactone()

    with pytest.raises(
        MacrolactoneDetectionError,
        match="classification=non_standard_macrocycle "
        "primary_reason_code=multiple_overlapping_macrocycle_candidates",
    ):
        MacrolactoneFragmenter().fragment_molecule(built.smiles)


def test_explicit_ring_size_selects_requested_ring():
    built = build_macrolactone(14)
    result = MacrolactoneFragmenter(ring_size=14).number_molecule(built.smiles)

    assert result.ring_size == 14


def test_explicit_ring_size_rejects_wrong_ring():
    built = build_macrolactone(12)

    with pytest.raises(MacrolactoneDetectionError):
        MacrolactoneFragmenter(ring_size=16).number_molecule(built.smiles)