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