US01-03: Discover and Reconcile Stable Assets (#48)

tests/unit/test_hashing.py

@@ -0,0 +1,48 @@
+"""Hashing: byte identity vs normalized-pixel identity."""
+
+import numpy as np
+from PIL import Image
+
+from photo_pipeline.services import hashing
+
+
+def _write(path, seed, size=(48, 32)):
+    rng = np.random.default_rng(seed)
+    arr = rng.integers(0, 256, (size[1], size[0], 3), dtype=np.uint8)
+    Image.fromarray(arr).save(path, quality=95)
+
+
+def test_sha256_is_deterministic_and_distinct(tmp_path):
+    a, b = tmp_path / "a.jpg", tmp_path / "b.jpg"
+    _write(a, 1)
+    _write(b, 2)
+    assert hashing.sha256_file(a) == hashing.sha256_file(a)
+    assert hashing.sha256_file(a) != hashing.sha256_file(b)
+
+
+def test_trailing_bytes_change_sha_but_not_pixels(tmp_path):
+    """Appending bytes after a JPEG (an EXIF-only-style edit) changes the file
+    hash while decoded pixels — and the pixel hash — stay the same."""
+    img = tmp_path / "a.jpg"
+    _write(img, 1)
+    sha_before = hashing.sha256_file(img)
+    pix_before = hashing.pixel_sha256(img)
+
+    with open(img, "ab") as handle:
+        handle.write(b"\xff\xfe trailing comment bytes")
+
+    assert hashing.sha256_file(img) != sha_before
+    assert hashing.pixel_sha256(img) == pix_before
+
+
+def test_different_images_have_different_pixel_hashes(tmp_path):
+    a, b = tmp_path / "a.jpg", tmp_path / "b.jpg"
+    _write(a, 1)
+    _write(b, 2)
+    assert hashing.pixel_sha256(a) != hashing.pixel_sha256(b)
+
+
+def test_safe_pixel_sha256_returns_none_for_garbage(tmp_path):
+    bad = tmp_path / "bad.jpg"
+    bad.write_bytes(b"not an image")
+    assert hashing.safe_pixel_sha256(bad) is None

tests/unit/test_path_policy.py

@@ -0,0 +1,83 @@
+"""Path policy: supported extensions, exclusion, deterministic discovery, and
+symlink-escape rejection."""
+
+import os
+
+import pytest
+
+from photo_pipeline import path_policy
+
+
+def test_is_supported_is_case_insensitive():
+    assert path_policy.is_supported("a.jpg")
+    assert path_policy.is_supported("a.JPG")
+    assert path_policy.is_supported("a.HEIC")
+    assert not path_policy.is_supported("a.txt")
+    assert not path_policy.is_supported("a.mp4")
+
+
+def test_is_excluded_matches_ignore_and_thumb_anywhere():
+    assert path_policy.is_excluded("lib/_IGNORE/secret.jpg")
+    assert path_policy.is_excluded("lib/album/.@__thumb/t.jpg")
+    assert not path_policy.is_excluded("lib/album/photo.jpg")
+
+
+def test_discovery_finds_supported_and_skips_excluded_and_unsupported(tmp_path):
+    (tmp_path / "album").mkdir()
+    (tmp_path / "_IGNORE").mkdir()
+    (tmp_path / "album" / ".@__thumb").mkdir()
+    keep = [tmp_path / "root.jpg", tmp_path / "album" / "a.PNG"]
+    for p in keep:
+        p.write_bytes(b"x")
+    (tmp_path / "notes.txt").write_bytes(b"x")
+    (tmp_path / "_IGNORE" / "secret.jpg").write_bytes(b"x")
+    (tmp_path / "album" / ".@__thumb" / "thumb.jpg").write_bytes(b"x")
+
+    found = path_policy.discover([tmp_path])
+    assert found == sorted(keep)
+
+
+def test_discovery_is_deterministic(tmp_path):
+    for i in range(5):
+        (tmp_path / f"{i}.jpg").write_bytes(b"x")
+    assert path_policy.discover([tmp_path]) == path_policy.discover([tmp_path])
+
+
+def test_resolve_within_accepts_inside_and_rejects_outside(tmp_path):
+    root = tmp_path / "lib"
+    root.mkdir()
+    inside = root / "a.jpg"
+    inside.write_bytes(b"x")
+    assert path_policy.resolve_within(root, inside) == inside.resolve()
+    with pytest.raises(path_policy.PathPolicyError):
+        path_policy.resolve_within(root, tmp_path / "outside.jpg")
+
+
+@pytest.mark.skipif(not hasattr(os, "symlink"), reason="symlinks unsupported")
+def test_symlink_escaping_root_fails_discovery(tmp_path):
+    root = tmp_path / "lib"
+    root.mkdir()
+    outside = tmp_path / "outside.jpg"
+    outside.write_bytes(b"x")
+    link = root / "link.jpg"
+    try:
+        os.symlink(outside, link)
+    except (OSError, NotImplementedError):
+        pytest.skip("cannot create symlink on this platform")
+    with pytest.raises(path_policy.PathPolicyError):
+        path_policy.discover([root])
+
+
+@pytest.mark.skipif(not hasattr(os, "symlink"), reason="symlinks unsupported")
+def test_symlink_within_root_is_allowed(tmp_path):
+    root = tmp_path / "lib"
+    root.mkdir()
+    target = root / "real.jpg"
+    target.write_bytes(b"x")
+    link = root / "alias.jpg"
+    try:
+        os.symlink(target, link)
+    except (OSError, NotImplementedError):
+        pytest.skip("cannot create symlink on this platform")
+    found = path_policy.discover([root])
+    assert target in found and link in found