chatkit/backend/tests/test_answer_grid.py

"""Unit tests for answer_grid module."""

from __future__ import annotations

import pytest

from app.answer_grid import (
    AnswerGrid,
    QuestionMeta,
    StudentRow,
    diff_student,
    from_dict,
    normalize_letter,
    seed_from_parsed,
    to_dict,
)


# ---------------------------------------------------------------------------
# normalize_letter
# ---------------------------------------------------------------------------


@pytest.mark.unit
@pytest.mark.parametrize(
    "raw, expected",
    [
        ("A", "A"),
        ("a", "A"),
        (" b ", "B"),
        ("C)", "C"),
        ("D) foo", "D"),
        ("=", "="),
        (" = ", "="),
        ("-", "="),
        ("–", "="),
        ("—", "="),
        ("", None),
        (None, None),
        ("   ", None),
        ("N/A", None),
        ("invalid", None),
        ("AB", None),
    ],
)
def test_normalize_letter(raw, expected):
    assert normalize_letter(raw) == expected


# ---------------------------------------------------------------------------
# from_dict validation
# ---------------------------------------------------------------------------


@pytest.mark.unit
def test_from_dict_minimal_valid():
    payload = {
        "total_questions": 2,
        "official_answers": ["A", "B"],
        "students": [{"name": "Alice", "answers": ["A", "C"]}],
        "questions": [
            {"number": 1, "text": "Q1", "options": ["A) x", "B) y"]},
            {"number": 2, "text": "Q2", "options": ["A) x", "B) y"]},
        ],
    }
    grid = from_dict(payload)
    assert grid.total_questions == 2
    assert grid.official_answers == ("A", "B")
    assert grid.students[0].name == "Alice"
    assert grid.students[0].answers == ("A", "C")


@pytest.mark.unit
def test_from_dict_pads_short_student_row():
    payload = {
        "total_questions": 3,
        "official_answers": ["A", "B", "C"],
        "students": [{"name": "Alice", "answers": ["A"]}],
        "questions": [{"text": "Q"}, {"text": "Q"}, {"text": "Q"}],
    }
    grid = from_dict(payload)
    assert grid.students[0].answers == ("A", None, None)


@pytest.mark.unit
def test_from_dict_truncates_long_student_row():
    payload = {
        "total_questions": 2,
        "official_answers": ["A", "B"],
        "students": [{"name": "Alice", "answers": ["A", "B", "C", "D"]}],
        "questions": [],
    }
    grid = from_dict(payload)
    assert grid.students[0].answers == ("A", "B")


@pytest.mark.unit
@pytest.mark.parametrize(
    "bad_field",
    [
        {"total_questions": 0},
        {"total_questions": -1},
        {"total_questions": "three"},
    ],
)
def test_from_dict_rejects_bad_n(bad_field):
    payload = {
        "total_questions": 2,
        "official_answers": ["A", "B"],
        "students": [],
        "questions": [],
    }
    payload.update(bad_field)
    with pytest.raises(ValueError):
        from_dict(payload)


@pytest.mark.unit
def test_from_dict_default_student_name():
    payload = {
        "total_questions": 1,
        "official_answers": ["A"],
        "students": [{"answers": ["A"]}],
        "questions": [],
    }
    grid = from_dict(payload)
    assert grid.students[0].name == "Student 1"


# ---------------------------------------------------------------------------
# to_dict roundtrip
# ---------------------------------------------------------------------------


@pytest.mark.unit
def test_to_dict_roundtrip():
    payload = {
        "total_questions": 2,
        "official_answers": ["A", "B"],
        "students": [{"name": "Alice", "answers": ["A", "C"]}],
        "questions": [
            {"number": 1, "text": "Q1", "options": ["A) x", "B) y"]},
            {"number": 2, "text": "Q2", "options": ["A) x", "B) y"]},
        ],
    }
    grid = from_dict(payload)
    again = from_dict(to_dict(grid))
    assert grid == again


# ---------------------------------------------------------------------------
# seed_from_parsed
# ---------------------------------------------------------------------------


@pytest.mark.unit
def test_seed_from_parsed_basic():
    questions = {
        "questions": [
            {"number": 1, "text": "Q1", "options": ["A) a", "B) b"]},
            {"number": 2, "text": "Q2", "options": ["A) a", "B) b"]},
        ]
    }
    students = {
        "students": [
            {
                "name": "Alice",
                "answers": [
                    {"question_number": 1, "answer": "A"},
                    {"question_number": 2, "answer": "b"},
                ],
            }
        ]
    }
    teachers = {
        "answers": [
            {"question_number": 1, "correct_answer": "A"},
            {"question_number": 2, "correct_answer": "B"},
        ]
    }
    grid = seed_from_parsed(questions, students, teachers)
    assert grid.total_questions == 2
    assert grid.official_answers == ("A", "B")
    assert grid.students[0].answers == ("A", "B")
    assert grid.questions[0].text == "Q1"


@pytest.mark.unit
def test_seed_from_parsed_legacy_dash_becomes_correct_marker():
    """Legacy '-' in a student answer normalizes to the canonical '=' marker."""
    questions = {"questions": [{"number": 1, "text": "Q"}, {"number": 2, "text": "Q"}]}
    students = {
        "students": [
            {
                "name": "Alice",
                "answers": [
                    {"question_number": 1, "answer": "-"},
                    {"question_number": 2, "answer": "D"},
                ],
            }
        ]
    }
    teachers = {
        "answers": [
            {"question_number": 1, "correct_answer": "A"},
            {"question_number": 2, "correct_answer": "B"},
        ]
    }
    grid = seed_from_parsed(questions, students, teachers)
    assert grid.students[0].answers == ("=", "D")


@pytest.mark.unit
def test_seed_from_parsed_preserves_explicit_equals():
    questions = {"questions": [{"number": 1, "text": "Q"}]}
    students = {
        "students": [
            {"name": "Alice", "answers": [{"question_number": 1, "answer": "="}]}
        ]
    }
    teachers = {"answers": [{"question_number": 1, "correct_answer": "A"}]}
    grid = seed_from_parsed(questions, students, teachers)
    assert grid.students[0].answers == ("=",)


@pytest.mark.unit
def test_seed_from_parsed_no_questions_derives_n_from_teacher():
    questions = {}
    students = {
        "students": [{"name": "A", "answers": [{"question_number": 1, "answer": "A"}]}]
    }
    teachers = {
        "answers": [
            {"question_number": 1, "correct_answer": "A"},
            {"question_number": 2, "correct_answer": "B"},
            {"question_number": 3, "correct_answer": "C"},
        ]
    }
    grid = seed_from_parsed(questions, students, teachers)
    assert grid.total_questions == 3
    assert grid.official_answers == ("A", "B", "C")


@pytest.mark.unit
def test_seed_from_parsed_all_empty_gives_n_one():
    grid = seed_from_parsed({}, {}, {})
    assert grid.total_questions == 1
    assert grid.students == ()
    assert grid.official_answers == (None,)


# ---------------------------------------------------------------------------
# diff_student
# ---------------------------------------------------------------------------


@pytest.mark.unit
def test_diff_student_returns_only_wrong():
    grid = AnswerGrid(
        total_questions=3,
        official_answers=("A", "B", "C"),
        students=(StudentRow(name="Alice", answers=("A", "D", "C")),),
        questions=(
            QuestionMeta(number=1, text="Q1", options=()),
            QuestionMeta(number=2, text="Q2", options=()),
            QuestionMeta(number=3, text="Q3", options=()),
        ),
    )
    wrongs = diff_student(grid, 0)
    assert len(wrongs) == 1
    assert wrongs[0].question.number == 2
    assert wrongs[0].student_answer == "D"
    assert wrongs[0].correct_answer == "B"


@pytest.mark.unit
def test_diff_student_blank_student_answer_is_wrong():
    grid = AnswerGrid(
        total_questions=2,
        official_answers=("A", "B"),
        students=(StudentRow(name="Alice", answers=("A", None)),),
        questions=(
            QuestionMeta(number=1, text="Q1", options=()),
            QuestionMeta(number=2, text="Q2", options=()),
        ),
    )
    wrongs = diff_student(grid, 0)
    assert len(wrongs) == 1
    assert wrongs[0].student_answer is None


@pytest.mark.unit
def test_diff_student_equals_marker_is_correct():
    """A '=' cell means 'student answered correctly' and must never appear in wrongs."""
    grid = AnswerGrid(
        total_questions=3,
        official_answers=("A", "B", "C"),
        students=(StudentRow(name="Alice", answers=("=", "X", "=")),),
        questions=tuple(
            QuestionMeta(number=i + 1, text=f"Q{i+1}", options=()) for i in range(3)
        ),
    )
    wrongs = diff_student(grid, 0)
    assert len(wrongs) == 1
    assert wrongs[0].question.number == 2
    assert wrongs[0].student_answer == "X"


@pytest.mark.unit
def test_diff_student_skips_equals_in_official():
    grid = AnswerGrid(
        total_questions=2,
        official_answers=("A", "="),  # Q2 has no authoritative key
        students=(StudentRow(name="Alice", answers=("X", "Y")),),
        questions=(
            QuestionMeta(number=1, text="Q1", options=()),
            QuestionMeta(number=2, text="Q2", options=()),
        ),
    )
    wrongs = diff_student(grid, 0)
    assert len(wrongs) == 1
    assert wrongs[0].question.number == 1


@pytest.mark.unit
def test_diff_student_skips_blank_official_key():
    grid = AnswerGrid(
        total_questions=2,
        official_answers=("A", None),  # Q2 key missing
        students=(StudentRow(name="Alice", answers=("B", "C")),),
        questions=(
            QuestionMeta(number=1, text="Q1", options=()),
            QuestionMeta(number=2, text="Q2", options=()),
        ),
    )
    wrongs = diff_student(grid, 0)
    assert len(wrongs) == 1
    assert wrongs[0].question.number == 1


@pytest.mark.unit
def test_diff_student_all_correct():
    grid = AnswerGrid(
        total_questions=2,
        official_answers=("A", "B"),
        students=(StudentRow(name="Alice", answers=("A", "B")),),
        questions=(
            QuestionMeta(number=1, text="Q1", options=()),
            QuestionMeta(number=2, text="Q2", options=()),
        ),
    )
    assert diff_student(grid, 0) == []


@pytest.mark.unit
def test_diff_student_out_of_range():
    grid = AnswerGrid(
        total_questions=1,
        official_answers=("A",),
        students=(),
        questions=(QuestionMeta(number=1, text="Q", options=()),),
    )
    with pytest.raises(IndexError):
        diff_student(grid, 0)


@pytest.mark.unit
def test_diff_student_preserves_question_order():
    grid = AnswerGrid(
        total_questions=4,
        official_answers=("A", "B", "C", "D"),
        students=(StudentRow(name="Alice", answers=("X", "B", "Y", "Z")),),
        questions=tuple(
            QuestionMeta(number=i + 1, text=f"Q{i+1}", options=()) for i in range(4)
        ),
    )
    wrongs = diff_student(grid, 0)
    assert [w.question.number for w in wrongs] == [1, 3, 4]