[AZ-175] Media table integration with XxHash64 content hashing and status lifecycle

Made-with: Cursor

tests/test_media_hash.py

@@ -0,0 +1,46 @@
+import xxhash
+
+
+def test_compute_media_content_hash_small_file():
+    # Arrange
+    from media_hash import compute_media_content_hash
+
+    data = b"x" * 100
+    expected = xxhash.xxh64(len(data).to_bytes(8, "little") + data).hexdigest()
+    # Act
+    out = compute_media_content_hash(data)
+    # Assert
+    assert out == expected
+
+
+def test_compute_media_content_hash_large_file():
+    # Arrange
+    from media_hash import compute_media_content_hash
+
+    data = (bytes(range(256)) * 20)[:5000]
+    n = len(data)
+    mid = (n - 1024) // 2
+    blob = (
+        n.to_bytes(8, "little")
+        + data[0:1024]
+        + data[mid : mid + 1024]
+        + data[-1024:]
+    )
+    expected = xxhash.xxh64(blob).hexdigest()
+    # Act
+    out = compute_media_content_hash(data)
+    # Assert
+    assert out == expected
+
+
+def test_compute_media_content_hash_virtual_prefix():
+    # Arrange
+    from media_hash import compute_media_content_hash
+
+    # Act
+    v = compute_media_content_hash(b"abc", virtual=True)
+    n = compute_media_content_hash(b"abc", virtual=False)
+    # Assert
+    assert v.startswith("V")
+    assert not n.startswith("V")
+    assert v == "V" + n