Maayan Zhitomirsky-Geffet

"""Unit tests for decompiler utility modules:
- bninja/utils/hashes.py
- bninja/utils/json_encoder.py
- bninja/analysis/low_level_normalization.py
"""
import hashlib
import json
import pytest


# ============================================================================
# 1a. hashes.py
# ============================================================================

from redb.extractors.decompiler.bninja.utils.hashes import (
    calculate_md5,
    calculate_sha256,
    calculate_tlsh,
)


class TestCalculateMD5:
    def test_calculate_md5_known_value(self):
        expected = hashlib.md5(b"test").hexdigest()
        assert calculate_md5("test") == expected

    def test_calculate_md5_empty(self):
        expected = hashlib.md5(b"").hexdigest()
        assert calculate_md5("") == expected


class TestCalculateSHA256:
    def test_calculate_sha256_known_value(self):
        expected = hashlib.sha256(b"hello world").hexdigest()
        assert calculate_sha256("hello world") == expected

    def test_calculate_sha256_empty_string(self):
        result = calculate_sha256("")
        assert len(result) == 64
        assert all(c in "0123456789abcdef" for c in result)


class TestCalculateTLSH:
    def test_calculate_tlsh_long_data(self):
        # TLSH requires >= 50 bytes
        data = "A" * 100
        result = calculate_tlsh(data)
        assert result is not None
        assert isinstance(result, str)

    def test_calculate_tlsh_short_data(self):
        data = "A" * 10
        result = calculate_tlsh(data)
        assert result is None

    def test_calculate_tlsh_deterministic(self):
        data = "x" * 200
        assert calculate_tlsh(data) == calculate_tlsh(data)



# ============================================================================
# 1b. json_encoder.py
# ============================================================================

from redb.extractors.decompiler.bninja.utils.json_encoder import BinaryNinjaEncoder


class TestBinaryNinjaEncoder:
    def test_encode_value_confidence_object(self):
        obj = type("VC", (), {"value": 42, "confidence": 255})()
        result = json.dumps(obj, cls=BinaryNinjaEncoder)
        assert json.loads(result) == 42

    def test_encode_str_fallback(self):
        obj = type("Obj", (), {"__str__": lambda self: "custom_repr"})()
        result = json.dumps(obj, cls=BinaryNinjaEncoder)
        assert json.loads(result) == "custom_repr"

    def test_encode_normal_types(self):
        data = {"a": 1, "b": [2, 3], "c": "hello"}
        result = json.dumps(data, cls=BinaryNinjaEncoder)
        assert json.loads(result) == data

    def test_encode_set_via_str(self):
        # Python sets have __str__, so BinaryNinjaEncoder converts them
        # to their string repr instead of raising TypeError.
        result = json.dumps(set([1, 2, 3]), cls=BinaryNinjaEncoder)
        parsed = json.loads(result)
        assert isinstance(parsed, str)
        assert "1" in parsed


# ============================================================================
# 1c. low_level_normalization.py
# ============================================================================

from redb.extractors.decompiler.bninja.analysis.low_level_normalization import (
    LowLevelNormalization,
)


class MockIL:
    """Mock IL node for normalization tests."""
    def __init__(self, operation, operands=None):
        self.operation = operation
        self.operands = operands or []


class TestLowLevelNormalization:
    def setup_method(self):
        self.normalizer = LowLevelNormalization()

    def test_normalize_single_instruction(self):
        node = MockIL(operation=5)
        result = self.normalizer.normalize_instruction_all_levels(node)
        assert result == [5]

    def test_normalize_nested_operands(self):
        child1 = MockIL(operation=10)
        child2 = MockIL(operation=20)
        root = MockIL(operation=1, operands=[child1, child2])
        result = self.normalizer.normalize_instruction_all_levels(root)
        assert result == [1, 10, 20]

    def test_normalize_empty_operands(self):
        node = MockIL(operation=42, operands=[])
        result = self.normalizer.normalize_instruction_all_levels(node)
        assert result == [42]

    def test_normalize_list_operands(self):
        # Simulates phi-node style list operands
        inner = MockIL(operation=99)
        node = MockIL(operation=7, operands=[[inner]])
        result = self.normalizer.normalize_instruction_all_levels(node)
        assert result == [7, 99]

    def test_normalize_none_input(self):
        result = self.normalizer.normalize_instruction_all_levels(None)
        assert result == []