fix(system): three next-runs, completed-date data_date, running mark

Split /system schedule into fans / snapshot+UP / discovery next-runs.
Drive data_date and snapshot_days from fully completed video snapshots
only, keep prior success date on failure, show task failure records, and
mark in-progress tasks as 运行中 without wall-clock next-run jumps.

tests/test_snapshot.py

@@ -256,13 +256,19 @@ def test_fan_failure_placeholder():
         transport = SnapshotTransport({}, fan={"code": -500, "message": "x", "data": None})
         db, repo, service = make_service(tmp, transport, retries=2)
         try:
+            # 合并跑：视频完成日不受粉丝失败影响
             result = service.run(date=DATE)
             assert result.total == 0
             assert result.fans_failed == 1
-            assert result.completed is False
+            assert result.completed is True
+            assert repo.kv_get("snapshot:last_completed") == DATE
             row = repo.get_up_stat(DATE, 37754047)
             assert row["ok"] == 0
             assert row["follower"] is None and row["following"] is None
+
+            fans_only = service.run_fans(date=DATE, force=True)
+            assert fans_only.fans_failed == 1
+            assert fans_only.completed is False
         finally:
             db.close()