From ed0b08967e2ee27f49f77fb1ff246f1641cfd7af Mon Sep 17 00:00:00 2001
From: andrew <andrew@kretzer.club>
Date: Thu, 9 Jul 2026 11:23:10 -0600
Subject: [PATCH] Fall back to rm when a fuzzy-dedup delete target isn't
 tracked in beets

A handful of confirmed candidates reference files that are no longer in
the beets DB (removed by some other pass, or never re-imported after a
move) but are still sitting on disk -- beet remove can't match them by
id or path either way, so they failed forever. If the file still exists
after a "No matching items found" error, remove it directly instead of
counting it as a failure. Applies to both --apply and --apply-pairs.
---
 pipeline/lib/find-fuzzy-dupes.py | 33 +++++++++++++++++++++++++++-----
 1 file changed, 28 insertions(+), 5 deletions(-)

diff --git a/pipeline/lib/find-fuzzy-dupes.py b/pipeline/lib/find-fuzzy-dupes.py
index 256559a..f4a2fed 100755
--- a/pipeline/lib/find-fuzzy-dupes.py
+++ b/pipeline/lib/find-fuzzy-dupes.py
@@ -180,9 +180,22 @@ def _apply_pairs(pairs_file: str, emit_json: bool) -> int:
             capture_output=True, text=True,
         )
         if result.returncode != 0:
-            print(f"  FAILED: {delete_path} — {result.stderr.strip()}", file=sys.stderr)
-            failed += 1
-            continue
+            if "No matching items found" in result.stderr and os.path.exists(delete_path):
+                # Not tracked in beets at all (e.g. it was already removed
+                # from the DB by some other pass, or never re-imported after
+                # a move) but the duplicate file itself is still sitting on
+                # disk -- `beet remove` can never match it by id or path
+                # either way, so just remove the file directly.
+                try:
+                    os.remove(delete_path)
+                except OSError as exc:
+                    print(f"  FAILED: {delete_path} — not in beets, then rm failed: {exc}", file=sys.stderr)
+                    failed += 1
+                    continue
+            else:
+                print(f"  FAILED: {delete_path} — {result.stderr.strip()}", file=sys.stderr)
+                failed += 1
+                continue
         print(f"  DELETE {delete_path}")
         deleted += 1
         if emit_json:
@@ -372,8 +385,18 @@ def main() -> int:
             capture_output=True, text=True,
         )
         if result.returncode != 0:
-            print(f"  FAILED: {path} — {result.stderr.strip()}", file=sys.stderr)
-            failed += 1
+            if "No matching items found" in result.stderr and os.path.exists(path):
+                # Item was already gone from beets by the time we got here
+                # (e.g. removed by an earlier pass) but the file is still on
+                # disk -- remove it directly instead of counting it failed.
+                try:
+                    os.remove(path)
+                except OSError as exc:
+                    print(f"  FAILED: {path} — not in beets, then rm failed: {exc}", file=sys.stderr)
+                    failed += 1
+            else:
+                print(f"  FAILED: {path} — {result.stderr.strip()}", file=sys.stderr)
+                failed += 1
     deleted = len(delete_targets) - failed - skipped_unconfirmed
     print(f"[fuzzy-dupes] done. {deleted} deleted, {failed} failed, {skipped_unconfirmed} skipped (unconfirmed).")
     return 1 if failed else 0