tests/test_hash_engine.py

import hashlib
from hdk.core.hash_engine import HashEngine
from hdk.models import HDKNode, NodeLevel, EventType


def test_content_hash_deterministic():
    engine = HashEngine()
    h1 = engine.hash_content("hello world")
    h2 = engine.hash_content("hello world")
    assert h1 == h2
    assert len(h1) == 64  # SHA-256 hex


def test_content_hash_different_input():
    engine = HashEngine()
    h1 = engine.hash_content("hello")
    h2 = engine.hash_content("hello!")
    assert h1 != h2


def test_hierarchical_hash_includes_parents():
    engine = HashEngine()
    content = engine.hash_content("my query")
    h_no_parent = engine.hash_with_parents(content, [])
    h_with_parent = engine.hash_with_parents(content, ["aaa111"])
    assert h_no_parent != h_with_parent


def test_changing_ancestor_breaks_hash():
    """Core HDK property: modifying any ancestor invalidates descendants."""
    engine = HashEngine()
    root = engine.hash_content("project_1")
    tab = engine.hash_with_parents(engine.hash_content("tab_1"), [root])
    conv = engine.hash_with_parents(engine.hash_content("conv_1"), [tab, root])
    query = engine.hash_with_parents(engine.hash_content("prompt"), [conv, tab, root])

    # Now change root
    root_tampered = engine.hash_content("project_1_MODIFIED")
    query_with_tampered = engine.hash_with_parents(
        engine.hash_content("prompt"), [conv, tab, root_tampered]
    )
    assert query != query_with_tampered


def _leaf(x: str) -> bytes:
    return hashlib.sha256(b"\x00" + x.encode()).digest()


def _node(left: bytes, right: bytes) -> bytes:
    return hashlib.sha256(b"\x01" + left + right).digest()


def test_merkle_root_single():
    """A single leaf is hashed with the leaf prefix, never returned bare."""
    engine = HashEngine()
    root = engine.compute_merkle_root(["aaa"])
    assert root != "aaa"
    assert root == _leaf("aaa").hex()


def test_merkle_root_pair():
    engine = HashEngine()
    root = engine.compute_merkle_root(["aaa", "bbb"])
    assert root == _node(_leaf("aaa"), _leaf("bbb")).hex()


def test_merkle_root_odd_count():
    """RFC 6962 shape for 3 leaves: H(0x01 || H(0x01||La||Lb) || Lc)."""
    engine = HashEngine()
    a, b, c = "a" * 64, "b" * 64, "c" * 64
    root3 = engine.compute_merkle_root([a, b, c])
    assert root3 == _node(_node(_leaf(a), _leaf(b)), _leaf(c)).hex()
    # no duplication of the odd leaf
    assert root3 != engine.compute_merkle_root([a, b, c, c])


def test_merkle_root_matches_rfc6962_split():
    """Bottom-up result equals RFC 6962's recursive split at the largest power of two."""
    def mth(d):
        if len(d) == 1:
            return _leaf(d[0])
        k = 1
        while k * 2 < len(d):
            k *= 2
        return _node(mth(d[:k]), mth(d[k:]))

    engine = HashEngine()
    for n in range(1, 18):
        data = [f"leaf{i}" for i in range(n)]
        assert engine.compute_merkle_root(data) == mth(data).hex(), n


def _chain(engine):
    root = HDKNode(id="r", content_hash=engine.hash_content("project"), parent_chain=[])
    root.self_hash = engine.hash_with_parents(root.content_hash, [])
    root.node_hash = engine.create_node_hash(root)
    child = HDKNode(id="c", content_hash=engine.hash_content("query"),
                    parent_chain=[root.node_hash])
    child.self_hash = engine.hash_with_parents(child.content_hash, child.parent_chain)
    child.node_hash = engine.create_node_hash(child)
    return root, child


def test_verify_chain_valid():
    """parent_chain holds the parent's node_hash, and verify_chain accepts that."""
    engine = HashEngine()
    root_node, child_node = _chain(engine)
    assert engine.verify_chain(child_node, [root_node]) is True


def test_verify_chain_tampered():
    engine = HashEngine()
    root_node, child_node = _chain(engine)

    # wrong parent hash in the chain
    child_node.parent_chain = ["TAMPERED_HASH"]
    assert engine.verify_chain(child_node, [root_node]) is False


def test_verify_chain_tampered_parent():
    engine = HashEngine()
    root_node, child_node = _chain(engine)
    root_node.model = "forged"  # parent's fields no longer match its node_hash
    assert engine.verify_chain(child_node, [root_node]) is False


def test_verify_chain_missing_parent():
    engine = HashEngine()
    root_node, child_node = _chain(engine)
    assert engine.verify_chain(child_node, []) is False