Filter soft-deleted and missing refs from visibility queries

list_references_by_asset_id and list_tags_with_usage were not filtering
out deleted_at/is_missing refs, allowing /view?filename=blake3:... to
serve files through hidden references and inflating tag usage counts.
Add list_all_file_paths_by_asset_id for orphan cleanup which
intentionally needs unfiltered access.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

app/assets/database/queries/__init__.py

@@ -31,6 +31,7 @@ from app.assets.database.queries.asset_reference import (
     get_unenriched_references,
     get_unreferenced_unhashed_asset_ids,
     insert_reference,
+    list_all_file_paths_by_asset_id,
     list_references_by_asset_id,
     list_references_page,
     mark_references_missing_outside_prefixes,
@@ -99,6 +100,7 @@ __all__ = [
     "get_unenriched_references",
     "get_unreferenced_unhashed_asset_ids",
     "insert_reference",
+    "list_all_file_paths_by_asset_id",
     "list_references_by_asset_id",
     "list_references_page",
     "list_tag_counts_for_filtered_assets",

app/assets/database/queries/asset_reference.py

@@ -549,6 +549,8 @@ def list_references_by_asset_id(
         session.execute(
             select(AssetReference)
             .where(AssetReference.asset_id == asset_id)
+            .where(AssetReference.is_missing == False)  # noqa: E712
+            .where(AssetReference.deleted_at.is_(None))
             .order_by(AssetReference.id.asc())
         )
         .scalars()
@@ -556,6 +558,25 @@ def list_references_by_asset_id(
     )
 
 
+def list_all_file_paths_by_asset_id(
+    session: Session,
+    asset_id: str,
+) -> list[str]:
+    """Return every file_path for an asset, including soft-deleted/missing refs.
+
+    Used for orphan cleanup where all on-disk files must be removed.
+    """
+    return list(
+        session.execute(
+            select(AssetReference.file_path)
+            .where(AssetReference.asset_id == asset_id)
+            .where(AssetReference.file_path.isnot(None))
+        )
+        .scalars()
+        .all()
+    )
+
+
 def upsert_reference(
     session: Session,
     asset_id: str,

app/assets/database/queries/tags.py

@@ -275,6 +275,7 @@ def list_tags_with_usage(
         .select_from(AssetReferenceTag)
         .join(AssetReference, AssetReference.id == AssetReferenceTag.asset_reference_id)
         .where(build_visible_owner_clause(owner_id))
+        .where(AssetReference.is_missing == False)  # noqa: E712
         .where(AssetReference.deleted_at.is_(None))
         .group_by(AssetReferenceTag.tag_name)
         .subquery()
@@ -311,6 +312,7 @@ def list_tags_with_usage(
             select(AssetReferenceTag.tag_name)
             .join(AssetReference, AssetReference.id == AssetReferenceTag.asset_reference_id)
             .where(build_visible_owner_clause(owner_id))
+            .where(AssetReference.is_missing == False)  # noqa: E712
             .where(AssetReference.deleted_at.is_(None))
             .group_by(AssetReferenceTag.tag_name)
         )