tests/test_reextract_archives.py

@@ -6,7 +6,7 @@ import io
 import zipfile
 
 import pytest
-from PIL import Image
+from PIL import Image, ImageDraw
 from sqlalchemy import select
 
 from backend.app.models import (
@@ -28,6 +28,18 @@ def _jpeg(color, size=256):
     return buf.getvalue()
 
 
+def _patterned_jpeg(idx, size=256):
+    """Structurally distinct per idx so the members don't phash-collapse — solid
+    colors collide to distance 0 and the second would dedupe away (the two-axis
+    dedup gotcha). A rectangle at an idx-derived position keeps phashes apart."""
+    img = Image.new("RGB", (size, size), "white")
+    x = 20 + idx * 60
+    ImageDraw.Draw(img).rectangle([x, x, x + 70, x + 70], fill="black")
+    buf = io.BytesIO()
+    img.save(buf, "JPEG")
+    return buf.getvalue()
+
+
 def test_reextract_links_archive_members_to_post(db_sync, tmp_path, monkeypatch):
     from backend.app.tasks import ml as ml_mod
     from backend.app.tasks import thumbnail as thumb_mod
@@ -122,10 +134,10 @@ def test_reextract_timebox_resumes_from_cursor(db_sync, tmp_path, monkeypatch):
     store_dir = images_root / "attachments" / "two"
     store_dir.mkdir(parents=True)
     att_ids = []
-    for n, color in enumerate(("red", "blue")):
-        arc = store_dir / f"{n}_archive_{color}"
+    for n in range(2):
+        arc = store_dir / f"{n}_archive"
         with zipfile.ZipFile(arc, "w") as zf:
-            zf.writestr(f"{color}.jpg", _jpeg(color))
+            zf.writestr(f"member{n}.jpg", _patterned_jpeg(n))
         sha = hashlib.sha256(arc.read_bytes()).hexdigest()
         att = PostAttachment(
             post_id=post.id, artist_id=artist.id, sha256=sha, path=str(arc),