feat(plex): search cast & crew names + clickable person cards

Backend (migration 0046_plex_people_search): a new people_text column (cast +
director names, de-duped) is folded into the generated search_vector at weight B,
so the top search box finds titles by an actor/director name and ranks them above
summary-only mentions. New /api/plex/people endpoint returns the cast/crew matching
the term (name prefix or word-start) with a film count + a headshot (pulled from one
representative film's live metadata; image bytes disk-cached).

Frontend: PlexBrowse shows matching people as virtual cards above the grid; clicking
one adds them to the actor/director filter (multi-value, from the C1 work) and clears
the search box so you land on exactly that person's films. Answers the 'why does
"drew" match Rambo?' confusion — it was matching the word in the synopsis; now names
are searchable. i18n en/hu/de.

⚠️ Prod needs a Plex re-sync after deploy to populate people_text (search_vector
regenerates automatically once the column has data).

backend/alembic/versions/0046_plex_people_search.py

@@ -0,0 +1,52 @@
+"""Plex: make cast/crew names searchable
+
+Revision ID: 0046_plex_people_search
+Revises: 0045_plex_filter_meta
+Create Date: 2026-07-06
+
+Adds `people_text` (the item's cast + director names, space-joined) and folds it into the generated
+`search_vector` at weight B — between the title (A) and the summary (C). So the top search box now
+finds titles by an actor/director name and ranks those above films that merely mention the name in
+their synopsis. A generated column's expression can't be altered in place, so we drop + recreate it
+(and its GIN index). `people_text` is populated on the next Plex sync.
+"""
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = "0046_plex_people_search"
+down_revision: Union[str, None] = "0045_plex_filter_meta"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+_CFG = "public.unaccent_simple"
+_NEW_VECTOR = (
+    f"setweight(to_tsvector('{_CFG}', coalesce(title, '')), 'A') || "
+    f"setweight(to_tsvector('{_CFG}', coalesce(people_text, '')), 'B') || "
+    f"setweight(to_tsvector('{_CFG}', left(coalesce(summary, ''), 1000)), 'C')"
+)
+_OLD_VECTOR = (
+    f"setweight(to_tsvector('{_CFG}', coalesce(title, '')), 'A') || "
+    f"setweight(to_tsvector('{_CFG}', left(coalesce(summary, ''), 1000)), 'C')"
+)
+
+
+def _rebuild_vector(expr: str) -> None:
+    op.execute("DROP INDEX IF EXISTS ix_plex_items_search_vector")
+    op.execute("ALTER TABLE plex_items DROP COLUMN search_vector")
+    op.execute(
+        f"ALTER TABLE plex_items ADD COLUMN search_vector tsvector "
+        f"GENERATED ALWAYS AS ({expr}) STORED"
+    )
+    op.execute("CREATE INDEX ix_plex_items_search_vector ON plex_items USING gin (search_vector)")
+
+
+def upgrade() -> None:
+    op.add_column("plex_items", sa.Column("people_text", sa.Text(), nullable=True))
+    _rebuild_vector(_NEW_VECTOR)
+
+
+def downgrade() -> None:
+    _rebuild_vector(_OLD_VECTOR)
+    op.drop_column("plex_items", "people_text")

backend/app/models.py

@@ -1044,10 +1044,14 @@ class PlexItem(Base, TimestampMixin, UpdatedAtMixin):
     genres: Mapped[list | None] = mapped_column(JSONB)
     directors: Mapped[list | None] = mapped_column(JSONB)
     cast_names: Mapped[list | None] = mapped_column(JSONB)
+    # Cast + director names, space-joined — folded into search_vector (weight B) so the search box
+    # finds titles by a person's name (and ranks them above summary-only mentions).
+    people_text: Mapped[str | None] = mapped_column(Text)
     search_vector: Mapped[object | None] = mapped_column(
         TSVECTOR,
         Computed(
             "setweight(to_tsvector('public.unaccent_simple', coalesce(title, '')), 'A') || "
+            "setweight(to_tsvector('public.unaccent_simple', coalesce(people_text, '')), 'B') || "
             "setweight(to_tsvector('public.unaccent_simple', left(coalesce(summary, ''), 1000)), 'C')",
             persisted=True,
         ),

backend/app/plex/sync.py

@@ -183,6 +183,9 @@ def _apply_item(row: PlexItem, lib_id: int, kind: str, meta: dict) -> None:
     row.genres = _tags(meta, "Genre")
     row.directors = _tags(meta, "Director")
     row.cast_names = _tags(meta, "Role", limit=20)
+    # Searchable people blob (cast + directors, de-duped) → folded into search_vector at sync time.
+    people = list(dict.fromkeys((row.directors or []) + (row.cast_names or [])))
+    row.people_text = " ".join(people) or None
     mf = _media_facts(meta)
     row.file_path = mf.get("file_path")
     row.part_key = mf.get("part_key")

backend/app/routes/plex.py

@@ -353,6 +353,73 @@ def facets(
     }
 
 
+def _like_escape(s: str) -> str:
+    return s.replace("\\", "\\\\").replace("%", "\\%").replace("_", "\\_")
+
+
+@router.get("/people")
+def people(
+    q: str,
+    library: str,
+    user: User = Depends(current_user),
+    db: Session = Depends(get_db),
+) -> dict:
+    """Cast/crew whose name matches the search term — drives the virtual person cards shown above the
+    grid. Returns up to a few matches (most films first) with a count + a photo."""
+    term = (q or "").strip()
+    lib = db.query(PlexLibrary).filter_by(plex_key=str(library)).first()
+    if lib is None or lib.kind != "movie" or len(term) < 2:
+        return {"people": []}
+    esc = _like_escape(term)
+    pfx, word = f"{esc}%", f"% {esc}%"  # name starts with the term, or any of its words does
+    rows = db.execute(
+        text(
+            "SELECT name, kind, count(*) c FROM ("
+            "  SELECT jsonb_array_elements_text(cast_names) AS name, 'actor' AS kind FROM plex_items"
+            "    WHERE library_id = :lib AND kind = 'movie'"
+            "  UNION ALL"
+            "  SELECT jsonb_array_elements_text(directors) AS name, 'director' AS kind FROM plex_items"
+            "    WHERE library_id = :lib AND kind = 'movie'"
+            ") x WHERE name ILIKE :pfx OR name ILIKE :word "
+            "GROUP BY name, kind ORDER BY c DESC, name LIMIT 5"
+        ),
+        {"lib": lib.id, "pfx": pfx, "word": word},
+    ).all()
+    return {
+        "people": [
+            {"name": name, "kind": kind, "count": c, "photo": _person_photo(db, lib.id, name, kind)}
+            for name, kind, c in rows
+        ]
+    }
+
+
+def _person_photo(db: Session, lib_id: int, name: str, kind: str) -> str | None:
+    """A person's headshot from one representative film's live metadata (image bytes are disk-cached)."""
+    col = PlexItem.directors if kind == "director" else PlexItem.cast_names
+    rk = (
+        db.query(PlexItem.rating_key)
+        .filter(PlexItem.library_id == lib_id, col.contains([name]))
+        .limit(1)
+        .scalar()
+    )
+    if not rk:
+        return None
+    try:
+        with PlexClient(db) as plex:
+            meta = plex.metadata(rk) or {}
+    except (PlexError, PlexNotConfigured):
+        return None
+    for r in meta.get("Director" if kind == "director" else "Role") or []:
+        if r.get("tag") == name:
+            thumb = str(r.get("thumb") or "")
+            return (
+                f"/api/plex/person-image?u={quote(thumb, safe='')}"
+                if thumb.startswith(_PERSON_IMG_HOST)
+                else None
+            )
+    return None
+
+
 @router.get("/show/{rating_key}")
 def show_detail(
     rating_key: str,