feat(feed): per-user search finds in Mine scope + full-text relevance search

Two related search improvements:

1) Your own live-search results now belong to your Mine feed. A new per-user
   search_finds table (migration 0030) records each video you surface via your
   YouTube search (the route inserts them idempotently). The Mine feed becomes
   'your non-hidden subscriptions OR your search finds', and the Source filter
   now applies in Mine too: organic = subscriptions, search = your search finds,
   all = both (default stays organic, so the main feed is unchanged). The shared
   Library keeps using the global via_search flag.

2) Feed search ranks by relevance instead of a whole-phrase substring. A custom
   unaccent_simple text-search config + GIN index (migration 0031) back a
   YouTube-like fuzzy match: word-order-independent, multi-word AND, prefix on
   the word being typed, accent-insensitive. A new 'relevance' sort orders by
   ts_rank; the channel name still matches as a substring. The rank is scaled to
   an integer so the keyset cursor pages it exactly (a raw float4 breaks paging).
   _filtered_query returns the rank expr so only the feed list uses it.

backend/alembic/versions/0031_title_fts.py

@@ -0,0 +1,49 @@
+"""full-text relevance search on video titles
+
+Revision ID: 0031_title_fts
+Revises: 0030_search_finds
+Create Date: 2026-06-30
+
+Adds a PostgreSQL full-text search index over video titles so the feed search box can rank by
+relevance (a YouTube-like "fuzzy" search: word-order-independent, multi-word AND, prefix on the
+word being typed) instead of a whole-phrase substring match.
+
+A custom text-search configuration `unaccent_simple` = the `simple` config (no stemming, no
+stopwords — right for a multilingual HU/EN/DE catalog) plus the `unaccent` dictionary, so the
+search stays accent-insensitive ("tiesto" matches "Tiësto"). A GIN expression index on
+`to_tsvector('public.unaccent_simple', coalesce(title,''))` makes the @@ match fast; the feed
+query uses the exact same expression so the planner picks the index. `unaccent` is already
+enabled (migration 0029). The config-based 2-arg to_tsvector is IMMUTABLE, so it's index-safe.
+"""
+from typing import Sequence, Union
+
+from alembic import op
+
+revision: str = "0031_title_fts"
+down_revision: Union[str, None] = "0030_search_finds"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.execute(
+        "CREATE TEXT SEARCH CONFIGURATION public.unaccent_simple (COPY = pg_catalog.simple)"
+    )
+    op.execute(
+        """
+        ALTER TEXT SEARCH CONFIGURATION public.unaccent_simple
+        ALTER MAPPING FOR asciiword, asciihword, hword_asciipart, word, hword, hword_part
+        WITH unaccent, simple
+        """
+    )
+    op.execute(
+        """
+        CREATE INDEX ix_videos_title_fts ON videos
+        USING gin (to_tsvector('public.unaccent_simple', coalesce(title, '')))
+        """
+    )
+
+
+def downgrade() -> None:
+    op.execute("DROP INDEX IF EXISTS ix_videos_title_fts")
+    op.execute("DROP TEXT SEARCH CONFIGURATION IF EXISTS public.unaccent_simple")