From 068a8e675f07b0bd4bd788b4f1e5be382bb396f9 Mon Sep 17 00:00:00 2001
From: Holden <holden@arch.fyi>
Date: Wed, 17 Jun 2026 01:12:28 +0000
Subject: [PATCH] fix: 4 correctness bugs from full-codebase audit

- executor: restore effective_count when replacement upload fails all retries
  (delete succeeded but slot was never filled, leaving cap undercount)
- diversity: skip zero-norm embeddings before dedup/FPS selection
  (InsightFace zeros pass dedup with similarity 0 and score distance 1.0,
  getting selected first as maximally diverse)
- cli: exclude None from skip_ids in _smaller_duplicate_ids
  (p.get('id') without None guard lets None into the set, silently
  dropping every other id-less person from the processed list)
- embeddings: select face nearest crop centre instead of largest by area
  (25% margin can pull a bigger neighbouring face into the crop;
  largest-face selection then embeds the wrong person)
---
 winnow/cli.py        |  1 +
 winnow/diversity.py  |  3 +++
 winnow/embeddings.py | 11 ++++++++---
 winnow/executor.py   |  6 ++++++
 4 files changed, 18 insertions(+), 3 deletions(-)

diff --git a/winnow/cli.py b/winnow/cli.py
index 7ac5447..d864b01 100644
--- a/winnow/cli.py
+++ b/winnow/cli.py
@@ -84,6 +84,7 @@ def _handle_duplicate_people(people: list[dict]) -> list[dict]:
             p.get("id")
             for ps in groups.values()
             for p in sorted(ps, key=lambda x: x.get("assetCount", 0), reverse=True)[1:]
+            if p.get("id") is not None
         }
 
     skip_ids = _smaller_duplicate_ids(duplicates)
diff --git a/winnow/diversity.py b/winnow/diversity.py
index 2b55c2e..154b4aa 100644
--- a/winnow/diversity.py
+++ b/winnow/diversity.py
@@ -303,6 +303,9 @@ def _select_by_embedding(
 
             emb = get_embedding(embed_img, asset_id=asset["id"])
             if emb is not None:
+                if np.linalg.norm(emb) < 1e-6:
+                    logger.debug("Zero-norm embedding for asset %s, skipping", asset["id"])
+                    continue
                 embeddings.append(emb)
                 valid_candidates.append(asset)
                 confidence_scores.append(confidence)
diff --git a/winnow/embeddings.py b/winnow/embeddings.py
index a715162..7787da7 100644
--- a/winnow/embeddings.py
+++ b/winnow/embeddings.py
@@ -218,9 +218,14 @@ def get_face_embedding(img_pil: Image.Image) -> np.ndarray | None:
         if not faces:
             return None
 
-        # Return embedding of largest face
-        largest = max(faces, key=lambda f: (f.bbox[2] - f.bbox[0]) * (f.bbox[3] - f.bbox[1]))
-        return largest.embedding
+        # Return embedding of the face nearest the crop centre; a large margin can pull
+        # a bigger neighbouring face into frame, and max-by-area would pick the wrong person.
+        cx, cy = img_pil.width / 2, img_pil.height / 2
+        nearest = min(
+            faces,
+            key=lambda f: ((f.bbox[0] + f.bbox[2]) / 2 - cx) ** 2 + ((f.bbox[1] + f.bbox[3]) / 2 - cy) ** 2,
+        )
+        return nearest.embedding
     except Exception as e:
         logger.error("Error getting face embedding: %s", e)
         return None
diff --git a/winnow/executor.py b/winnow/executor.py
index 52b7401..e263f4c 100644
--- a/winnow/executor.py
+++ b/winnow/executor.py
@@ -608,6 +608,12 @@ def upload_to_frigate(jobs: list[dict]) -> None:
                             progress.console.print(
                                 f"    [red]✗ {fname}: {type(e).__name__} - {e} (after {max_retries} attempts)[/red]"
                             )
+                    else:
+                        # All retries exhausted without a successful upload.
+                        # Restore the slot freed by the preceding delete so the next
+                        # candidate still sees at_cap=True and must beat the replacement gate.
+                        if at_cap:
+                            effective_count += 1
 
                     progress.advance(upload_task)