DATASET · research

Daybreak 置信度空模型闸门

research.stock_leg_attribution_null_gate

研究成品

将真实信号与打乱及规模匹配的空模型比较，记录 lift、deflated Sharpe 和逐日闸门结论。

下载本类 JSON 下载完整 JSON.gz 先获取 Key

字段56

关键字段1

更新口径producer_schedule

时间列eval_date

单页上限5000

分页窄筛选

PIT不支持

规模<1K

WHAT IT ANSWERS

这张表适合回答什么

判断置信度是否超过随机基线
检查 deflated Sharpe
审计每日 gate verdict

GRAIN & TIME

一行代表什么

业务粒度: 每行代表“Daybreak 置信度空模型闸门”的一条记录；逻辑粒度由 eval_date 共同确定。
时间口径: 以 eval_date 为主要查询时间轴；除下方限制特别说明外，时间戳按 UTC、日期按来源市场日历解释。
更新节奏: 由对应生产任务或事件触发更新，不承诺固定日历频率；查询前先检查最新时间。

KEY FIELDS

先看这些字段

eval_date: eval date 对应的日历日期。

RECOMMENDED JOINS

怎样和其他表关联

该表主要按自身业务键使用，目前没有推荐的跨表主键联接。

HOW TO QUERY SAFELY

查询方法与已知限制

先用 catalog 确认当前契约；可按 catalog 中列出的 filters 进一步缩小范围；单页最多 5000 行，该表不提供稳定 cursor，结果过大时必须缩小标的、日期或其他筛选。

返回范围仍受 Friends 权限、单页上限和 catalog 中的筛选契约约束。

MACHINE CONTRACT

筛选、排序与分页

UNVERSIONED

默认排序: eval_date
唯一键: —
读取模式: canonical_final_active
最大时间窗: 未单独限制

可用 filters

gate_passed→ gate_passed · integer

oos→ oos · integer

POINT IN TIME

可复现读取

NOT AVAILABLE

该资产没有 PIT 契约。普通 rows 端点仍按当前公开读模式返回数据；不要把它表述成历史时点快照。

COPY-READY CALLS

怎么调用

同一数据集可用 HTTP、Python SDK、CLI 或 MCP；Key 始终只通过环境变量/Bearer Header 传递。

READ ONLY

HTTP

curl --fail-with-body -H "Authorization: Bearer $FRIENDS_DATA_API_KEY" "https://api.alphahubs.uk/api/v1/data/datasets/research/stock_leg_attribution_null_gate/rows?limit=100&start=2026-08-01&end=2026-08-01"

PYTHON SDK

from friends_data_api import FriendsDataClient, Settings

async with FriendsDataClient(Settings.from_env()) as client:
    page = await client.get_dataset_rows('research', 'stock_leg_attribution_null_gate', limit=100, start='2026-08-01', end='2026-08-01')
    print(page.data)
    if page.has_more:
        print("next cursor:", page.next_cursor)

CLI

friends-data get-dataset-rows research stock_leg_attribution_null_gate --limit 100 --start 2026-08-01 --end 2026-08-01

MCP TOOL CALL

{
  "arguments": {
    "dataset": "stock_leg_attribution_null_gate",
    "end": "2026-08-01",
    "filters": {},
    "limit": 100,
    "namespace": "research",
    "start": "2026-08-01"
  },
  "tool": "get_dataset_rows"
}

FIELD DICTIONARY

56 个返回字段

descriptionSource 可在 JSON 中区分仓库注释、标准术语表和名称推导说明。

筛选字段

#	字段	ClickHouse 类型	JSON	可空	角色	说明
1	`eval_date`date	`Date`	`string / date`	no	time, default_order, warehouse_primary_key	Calendar date for eval date (YYYY-MM-DD).名称推导
2	`side`	`LowCardinality(String)`	`string`	no	warehouse_primary_key	Source-normalized field: side.名称推导
3	`n_dates`	`UInt32`	`integer`	no	value	Count of dates.名称推导
4	`n_rows`	`UInt32`	`integer`	no	value	Count of rows.名称推导
5	`n_explosive`	`UInt32`	`integer`	no	value	Count of explosive.名称推导
6	`n_trials`	`UInt32`	`integer`	no	value	Count of trials.名称推导
7	`conviction_lift`	`Float64`	`number`	no	value	Source-normalized field: conviction lift.名称推导
8	`shuffled_p`	`Float64`	`number`	no	value	Source-normalized field: shuffled p.名称推导
9	`size_matched_p`	`Float64`	`number`	no	value	Source-normalized field: size matched p.名称推导
10	`beats_shuffled`	`UInt8`	`integer`	no	value	Source-normalized field: beats shuffled.名称推导
11	`beats_size_matched`	`UInt8`	`integer`	no	value	Source-normalized field: beats size matched.名称推导
12	`p`	`Float64`	`number`	no	value	Source-normalized field: p.名称推导
13	`deflated_sharpe`	`Float64`	`number`	no	value	Source-normalized field: deflated sharpe.名称推导
14	`oos`	`UInt8`	`integer`	no	filter	Source-normalized field: oos.名称推导
15	`gate_passed`	`UInt8`	`integer`	no	filter	Source-normalized field: gate passed.名称推导
16	`fail_reason`	`String`	`string`	no	value	Source-normalized field: fail reason.名称推导
17	`n_effective`	`Int32`	`integer`	no	value	Count of effective.名称推导
18	`streak_required`	`Int32`	`integer`	no	value	Source-normalized field: streak required.名称推导
19	`streak_clean`	`Int32`	`integer`	no	value	Source-normalized field: streak clean.名称推导
20	`family`	`LowCardinality(String)`	`string`	no	value	Source-normalized field: family.名称推导
21	`n_trials_source`	`LowCardinality(String)`	`string`	no	value	Count of trials source.名称推导
22	`search_face_hash`	`String`	`string`	no	value	Source-normalized field: search face hash.名称推导
23	`residualizer_primary_basis`	`LowCardinality(String)`	`string`	no	value	Source-normalized field: residualizer primary basis.名称推导
24	`residualizer_shadow_basis`	`LowCardinality(String)`	`string`	no	value	Source-normalized field: residualizer shadow basis.名称推导
25	`residualizer_rollout_mode`	`LowCardinality(String)`	`string`	no	value	Source-normalized field: residualizer rollout mode.名称推导
26	`residualizer_shadow_ready`	`UInt8`	`integer`	no	value	Source-normalized field: residualizer shadow ready.名称推导
27	`residualizer_provenance_json`	`String`	`string`	no	value	Source-normalized field: residualizer provenance json.名称推导
28	`residualizer_dsr_family`	`LowCardinality(String)`	`string`	no	value	Source-normalized field: residualizer dsr family.名称推导
29	`residualizer_shadow_conviction_lift`	`Float64`	`number`	no	value	Source-normalized field: residualizer shadow conviction lift.名称推导
30	`residualizer_shadow_shuffled_p`	`Float64`	`number`	no	value	Source-normalized field: residualizer shadow shuffled p.名称推导
31	`residualizer_shadow_size_matched_p`	`Float64`	`number`	no	value	Source-normalized field: residualizer shadow size matched p.名称推导
32	`residualizer_shadow_beats_shuffled`	`UInt8`	`integer`	no	value	Source-normalized field: residualizer shadow beats shuffled.名称推导
33	`residualizer_shadow_beats_size_matched`	`UInt8`	`integer`	no	value	Source-normalized field: residualizer shadow beats size matched.名称推导
34	`residualizer_shadow_p`	`Float64`	`number`	no	value	Source-normalized field: residualizer shadow p.名称推导
35	`residualizer_shadow_deflated_sharpe`	`Float64`	`number`	no	value	Source-normalized field: residualizer shadow deflated sharpe.名称推导
36	`residualizer_shadow_oos`	`UInt8`	`integer`	no	value	Source-normalized field: residualizer shadow oos.名称推导
37	`residualizer_shadow_gate_passed`	`UInt8`	`integer`	no	value	Source-normalized field: residualizer shadow gate passed.名称推导
38	`residualizer_shadow_n_effective`	`Int32`	`integer`	no	value	Source-normalized field: residualizer shadow n effective.名称推导
39	`factor_alpha_basis`	`LowCardinality(String)`	`string`	no	value	Source-normalized field: factor alpha basis.名称推导
40	`factor_alpha_ready`	`UInt8`	`integer`	no	value	Source-normalized field: factor alpha ready.名称推导
41	`factor_alpha_coverage_frac`	`Float64`	`number`	no	value	Source-normalized field: factor alpha coverage frac.名称推导
42	`factor_alpha_betas_json`	`String`	`string`	no	value	Source-normalized field: factor alpha betas json.名称推导
43	`factor_alpha_conviction_lift`	`Float64`	`number`	no	value	Source-normalized field: factor alpha conviction lift.名称推导
44	`factor_alpha_shuffled_p`	`Float64`	`number`	no	value	Source-normalized field: factor alpha shuffled p.名称推导
45	`factor_alpha_size_matched_p`	`Float64`	`number`	no	value	Source-normalized field: factor alpha size matched p.名称推导
46	`factor_alpha_beats_shuffled`	`UInt8`	`integer`	no	value	Source-normalized field: factor alpha beats shuffled.名称推导
47	`factor_alpha_beats_size_matched`	`UInt8`	`integer`	no	value	Source-normalized field: factor alpha beats size matched.名称推导
48	`factor_alpha_p`	`Float64`	`number`	no	value	Source-normalized field: factor alpha p.名称推导
49	`factor_alpha_deflated_sharpe`	`Float64`	`number`	no	value	Source-normalized field: factor alpha deflated sharpe.名称推导
50	`factor_alpha_raw_deflated_sharpe`	`Float64`	`number`	no	value	Source-normalized field: factor alpha raw deflated sharpe.名称推导
51	`factor_alpha_oos`	`UInt8`	`integer`	no	value	Source-normalized field: factor alpha oos.名称推导
52	`factor_alpha_gate_passed`	`UInt8`	`integer`	no	value	Source-normalized field: factor alpha gate passed.名称推导
53	`factor_alpha_n_effective`	`Int32`	`integer`	no	value	Source-normalized field: factor alpha n effective.名称推导
54	`computed_at`UTC_datetime	`DateTime64(3, 'UTC')`	`string / date-time`	no	value	UTC timestamp for computed at.名称推导
55	`version`	`UInt64`	`integer`	no	provenance	Monotonic row version used for correction-aware reads.标准术语表
56	`is_deleted`	`UInt8`	`integer`	no	value	Soft-delete marker; public serving excludes current tombstones.标准术语表

字典口径

结构来自 Friends 只读身份可见的真实 serving relation；权限、过滤、排序、PIT 与分页来自可执行 DatasetSpec。Schema observed at 2026-10-06T18:08:52Z。