perf(feed): collapse count + facets into one /feed/summary pass (E2 S4, PERF-13)

The FE fired /feed/count (1 pass) + /facets (3 passes) on every filter change =
4 full scans of the subscribed set. New /api/feed/summary does ONE grouped query
(the base filter with tags stripped, GROUP BY channel_id -> per-channel video
counts) and derives the total count + every tag's distinct-channel facet count in
Python on the ~few-hundred-row channel map, faithfully mirroring filtered_query's
channel-based tag logic + the /facets drill-down (own category dropped, unless a
conjunctive topic-AND). Returns {count, facets} — the same numbers the two old
endpoints returned.

FE: a single shared query (Feed reads .count, Sidebar reads .facets); same key +
filters, so React Query dedupes to ONE request. Invalidations repointed to
qk.feedSummary. The old /feed/count + /facets endpoints (and their now-unused FE
api/qk) are kept as the equivalence-test oracle + safety net; a later cleanup can
drop them.

New equivalence test (test_feed_views, 8 filter combos incl. cross-category,
disjunctive/conjunctive topic modes, search): /feed/summary count == /feed/count
and facets == /facets, byte-identical. Structural win: 2 requests / 4-6 queries
-> 1 request / 2-3 queries (the review measured 530 -> 144 ms). Gate green:
344 pytest, frontend tsc/eslint/knip/prettier/222 vitest.

frontend/src/lib/queryKeys.test.ts

@@ -16,6 +16,8 @@ describe("qk — bare vs parameterized branch", () => {
     expect(qk.feedCount(FILTERS)).toEqual(["feed-count", FILTERS]);
     expect(qk.facets()).toEqual(["facets"]);
     expect(qk.facets(FILTERS)).toEqual(["facets", FILTERS]);
+    expect(qk.feedSummary()).toEqual(["feed-summary"]);
+    expect(qk.feedSummary(FILTERS)).toEqual(["feed-summary", FILTERS]);
     expect(qk.playlist()).toEqual(["playlist"]);
     expect(qk.plexShow()).toEqual(["plex-show"]);
     expect(qk.plexShow("s1")).toEqual(["plex-show", "s1"]);