补上完整数据版 K4：年运新字段放在原键后面，Yogini 缺字段不再整段失败。

pytest tests/test_annual_tajika_pack.py tests/test_dasha_master_children.py tests/test_report_longform_gaps2.py：30 项，退出码 0。

Co-Authored-By: Grok (agent loop)

CHANGELOG.md

@@ -1,5 +1,11 @@
 # 印度占星 Skill 更新日志
 
+## 2026-10-23 — 阅读页的 Yogini 子运缺字段时不再整段消失
+
+- 补子运时只在原来的大运行上加子运，不改、不删原来的字段。缺完整年数时仍显示主运，不再整段报错。
+- 年运资料包新加的整宫和 Tripataki 改放在原有字段后面，已有字段的顺序不变。
+- Skill 版本不变（BUG-1425、BUG-1426）。
+
 ## 2026-10-23 — 完整数据版补上行星解释、后两年年运和 72 年年返
 
 - 每颗星和每个宫主的解释加长，仍用「可能」的说法，并写上功能吉凶。土星和 KP 的结果段这一版仍先不做。

docs/BUG_HISTORY.md

@@ -19318,3 +19318,35 @@
 - 复发自：无。
 - 修复版本：分支 `codex/full-data-parity-1-20261021`。
 
+## BUG-1425 | 年运包把新字段插进原有键序
+
+- 状态：resolved
+- 首次发现：2026-10-23
+- 最近更新：2026-10-23
+- 影响面：年运资料包 `build_annual_tajika_pack` 的键序。阅读页按字段名取数，键序本身不改句子，但合同测试和按原序读取的下游会错位。
+- 用户现象：完整数据版全量测试里，年运包合同失败。`annual_houses` 插到了 `mudda_dasha` 前面。
+- 触发条件：生成年运包。B 段把整宫和 Tripataki 写进包时插在既有键中间。
+- 根因：新键写在 `sahams` 和 `mudda_dasha` 之间，改了 2026-10-08 定下的原有顺序。
+- 修复：`annual_houses`、`tripataki_chakra` 改放在原有键（到 `audit` 为止）之后。原有键的顺序和值不变。
+- 验证：`tests/test_annual_tajika_pack.py`（键序断言写了原值 / 新值 / 原因）。与 `tests/test_dasha_master_children.py`、`tests/test_report_longform_gaps2.py` 一起 30 项，退出码 0。
+- 防复发：`test_builds_annual_tajika_pack_contract` 锁住完整键序。再把新键插回中间，这项会失败。
+- 相关记录：issue #91 K4。BUG-1426。
+- 复发自：无
+- 修复版本：分支 `codex/full-data-parity-1-20261021`。
+
+## BUG-1426 | 补 Yogini 子运时缺完整年数，阅读页整段失败
+
+- 状态：resolved
+- 首次发现：2026-10-23
+- 最近更新：2026-10-23
+- 影响面：阅读页「Yogini Dasha 周期」；标准大运表里的 Yogini 行。
+- 用户现象：Yogini 子运整段出不来。全量测试报 `ValueError: yogini antardasha failed for Mangala: 'full_years'`。
+- 触发条件：大运行上没有 `full_years`（或没有未裁切的起点）时，报告渲染去补子运。
+- 根因：`enrich_master_family` 调用 `yogini_antardashas` 时，缺字段直接抛错，渲染在写出标题前就停了。补子运也不该改掉或丢掉这一行原来的字段。
+- 修复：只在原行上追加 `antardasha`。算子运要用的字段若只在嵌套的原始行里，只给计算用，不改原行。缺字段或计算失败时记一条原因日志，原行照旧返回，标题和主运表仍然出现。Narayana、Kalachakra 同样只追加子运，缺日期时不再把整份报告打崩。
+- 验证：`tests/test_report_longform_gaps2.py::test_auxiliary_dasha_headings_render_from_native_modules`。`tests/test_dasha_master_children.py` 新增：每个体系每一行，补充前后原有键和值逐项相同；缺 `full_years` 时不抛错、不改原行。上述三文件 30 项，退出码 0。
+- 防复发：上面两条测试。再把缺字段改成抛错，或补子运时改掉原字段，测试会失败。
+- 相关记录：issue #91 K4。BUG-1425。
+- 复发自：无
+- 修复版本：分支 `codex/full-data-parity-1-20261021`。
+

docs/tasks/PROGRESS-full-data-parity-1-b-20261023.md

@@ -85,3 +85,21 @@ Yogi 记录在本分支曾写成 BUG-1420，后改为 BUG-1421。staging 上 BUG
 | 200 张英文扫描 | pytest tests/test_report_english_sweep.py，屏蔽 PyJHora | 退出码 0，连同三张盘计时一共约 27 分钟 |
 | 前端类型和 lint | 本轮没改前端 | 没跑 |
 | 快速门、全量 pytest | 不跑 | 交给验收 |
+
+## 修复轮（K4）
+
+产品选 A 之后，K1–K3 已在 `cfd0ab68`。本轮只补 K4，没有重做那三项。开工先合并 `origin/staging`（含修复单里的 K4），没有 rebase，没有强推。
+
+| 项 | 结果 |
+| --- | --- |
+| 年运包键序 | `annual_houses`、`tripataki_chakra` 改到原有键（止于 `audit`）之后。`mudda_dasha` 仍紧跟 `sahams` |
+| Yogini 子运 | 只追加子运。缺 `full_years` 时记原因并保留原行，阅读页标题和主运表还在。原有键和值不改 |
+| 断言 | 年运键序写了原值 / 新值 / 原因。新增：每个体系每一行，补充前后原有键和值相同 |
+
+| 检查 | 命令 | 结果 |
+| --- | --- | --- |
+| K4 定向测试 | `.venv\Scripts\python -m pytest tests/test_annual_tajika_pack.py tests/test_dasha_master_children.py tests/test_report_longform_gaps2.py -q` | 30 项，退出码 0 |
+| 前端类型和 lint | 本轮没改前端 | 没跑 |
+| 快速门、全量 pytest、200 张扫描 | 不跑 | 交给验收 |
+
+编号：BUG-1425（年运键序）、BUG-1426（Yogini 子运整段失败）。当前最大号在合并后仍是 BUG-1424，这两条接在后面。

docs/tasks/README.md

@@ -522,7 +522,7 @@
 | `TASK-chart-ascendant-slash-20261021.md` | `PROGRESS-chart-ascendant-slash-20261021.md` | #87 上升状态格「不适用」改「/」+ 悬停 / 点按说明；报告上升行同改并在表下加说明（推翻 10-15 口径） | 已合入 staging | `codex/chart-ascendant-slash-20261021` |
 | `TASK-report-english-han-leak-20261021.md` | `PROGRESS-report-english-han-leak-20261021.md` | #89 个人报告英文版约 4 成整份丢弃：漏译汉字源头补英文孪生字段、修吃空格；200 张固定种子回归 + 快速门 smoke（复发自 BUG-1272）；中文逐字节不变 | 已合入 staging | `codex/report-english-han-leak-20261021` |
 | `TASK-chart-dasha-yogini-20261021.md` | — | #88 星盘页大运右栏切换 Chara / Yogini，Yogini 补小运（按比例，内部核对，页面注明未经外部软件核对）；Ashtottari 另立单（起运表与 PyJHora 8/12 宿不一致） | 已合入 staging | `codex/chart-dasha-yogini-20261021` |
-| `TASK-full-data-parity-1-20261021.md` | `PROGRESS-full-data-parity-1-b-20261023.md` | #91 完整数据版补齐（一）：A 段渲染已在 staging；B 段补 Yogi、大运子运、当年和后两年的月返照与 Tithi、36 个月星历。72 年年返和解释补长未做满 | B 段本地完成，待验收 | `codex/full-data-parity-1-20261021` |
+| `TASK-full-data-parity-1-20261021.md` | `PROGRESS-full-data-parity-1-b-20261023.md` | #91 完整数据版补齐（一）：K1–K3 已做；K4 把年运新字段放到原键之后，Yogini 缺完整年数时不再整段失败（BUG-1425、BUG-1426） | 待验收 | `codex/full-data-parity-1-20261021` |
 | `TASK-full-data-parity-2-20261021.md` | — | #92 完整数据版补齐（二）：研究型推运附录（Tithi Ashtottari、Yoga / Nakshatra Pravesha、Karana 等，口径换成我方真恒星）、四种星座大运第三层、Kalachakra 第三层（简化模型）、Niryaana Shoola 子运只在有书面出处时做、D81 / D108 / D144 研究型分盘；全部标研究口径，不进对话卡与校正；不设大小上限 | 排队（等 #91 合入） | `codex/full-data-parity-2-20261021` |
 | `TASK-goel-jaimini-strength-20261022.md` | `PROGRESS-goel-jaimini-strength-20261022.md` | #94 搬上游公开的 Goel Jaimini 强度与 Brahma / Maheshwara / Rudra（4 个上游原样文件，按 #81 决策 A 先例根目录预算 448→452；来源串改书目引用）；只做计算、不改报告；冻结身份不变 | 已合入 staging | `codex/goel-jaimini-strength-20261022` |
 | `TASK-chart-subject-table-fix-20261022.md` | `PROGRESS-chart-subject-table-fix-20261022.md` | **#86 第 1 次打回**：报告所有表被强制单行（打印跑出纸面）；保存他人复制了档案页代码且已分叉；点编辑框内边距会丢修改 | 已合入 staging | `codex/chart-subject-table-20261021` |
@@ -530,5 +530,5 @@
 | `TASK-synastry-staging-fix-fix-20261022.md` | `PROGRESS-synastry-staging-fix-fix-20261022.md` | **#85 第 1 次打回**（T7～T11 主体通过）：新地址没登记进路由清单致快速门红；CHANGELOG 与 T11 矛盾；侧栏点亮测试是空断言 | 已合入 staging | `codex/synastry-staging-fix-20261021` |
 | `TASK-report-english-han-leak-fix-20261022.md` | `PROGRESS-report-english-han-leak-fix-20261022.md` | **#89 第 1 次打回**：改了冻结的校正打分文件 `jaimini.py`；测试的 PyJHora 屏蔽没传到子进程（测的不是生产那份报告，也是慢的原因）；缺英文时留空藏住漏洞 | 已合入 staging | `codex/report-english-han-leak-20261021` |
 | `TASK-full-data-parity-1-fix-20261022.md` | `PROGRESS-full-data-parity-1-fix-20261022.md` | **#91 A 段第 1 次打回**：新六章 69 节里 16 节在三张虚构盘上整节「未判定」，其中 12 节数据在资料包里、是读错了字段；测试抓不到整节未判定；新六章绕过了报告口径替换；中文附录悄悄丢列和瑜伽条目要量出来 | 本地完成，待验收 | `codex/full-data-parity-1-20261021` |
-| `TASK-full-data-parity-1-b-fix-20261023.md` | `PROGRESS-full-data-parity-1-b-20261023.md` | **#91 B 段打回（第 3 次，产品 10-23 选 A 交回 Grok，并新增 K3 72 年返照表）**：T9 解释补长不在让步清单却没做（行星一章 4,574 字符，目标约 8,970）；D8 后两年缺整宫、Tajika、Tripataki、Saham 轨迹、年主依据 | 待执行 | `codex/full-data-parity-1-20261021` |
+| `TASK-full-data-parity-1-b-fix-20261023.md` | `PROGRESS-full-data-parity-1-b-20261023.md` | **#91 B 段修复**：K1–K3 已交；K4 年运键序与 Yogini 子运字段已补（BUG-1425、BUG-1426） | 待验收 | `codex/full-data-parity-1-20261021` |
 | — | `PROGRESS-full-data-parity-1-fix2-20261022.md` | **#91 A 段第 2 次打回**（Claude 口头修复单 F1～F4）：11 节只印「未判定」和键名改读真实记录；测试只认非第一列真实值、禁止键名格；中文瑜伽组合句由中文源给出（BUG-1418）；六章降为 `##`、节 `###`（BUG-1419） | 本地完成，待验收 | `codex/full-data-parity-1-fix2-20261022` |

scripts/annual_native/dasha_children.py

@@ -7,10 +7,13 @@ and they hang on the majors this repository already calculated.
 
 from __future__ import annotations
 
+import logging
 from copy import deepcopy
 from datetime import datetime, timedelta
 from typing import Any
 
+logger = logging.getLogger(__name__)
+
 
 def _rotate_cycle_rows(cycle_rows: list[dict], start_rashi: str | None) -> list[dict]:
     if not cycle_rows:
@@ -118,10 +121,7 @@ def _copy_rows(rows: list[dict]) -> list[dict]:
     return deepcopy([row for row in rows if isinstance(row, dict)])
 
 
-def _narayana_periods(module: dict) -> list[dict]:
-    sequence = module.get("mahadasha_sequence")
-    if not isinstance(sequence, list):
-        return []
+def _narayana_children(module: dict) -> dict[str, list[dict]]:
     children_by_parent: dict[str, list[dict]] = {}
     boundaries = module.get("dense_boundaries")
     if isinstance(boundaries, list):
@@ -140,6 +140,25 @@ def _narayana_periods(module: dict) -> list[dict]:
                 "parent_lord": row.get("parent_md"),
                 "source": "narayana_dense_boundaries",
             })
+    return children_by_parent
+
+
+def _narayana_periods(module: dict, rows: list[dict] | None = None) -> list[dict]:
+    children_by_parent = _narayana_children(module)
+    if rows:
+        # Keep every field already on the row. Only attach children.
+        periods = []
+        for row in rows:
+            copied = deepcopy(row)
+            if not (isinstance(copied.get("antardasha"), list) and copied.get("antardasha")):
+                sign = copied.get("sign") or copied.get("rashi")
+                parent_key = str(copied.get("period_key") or sign or "")
+                copied["antardasha"] = children_by_parent.get(parent_key, [])
+            periods.append(copied)
+        return periods
+    sequence = module.get("mahadasha_sequence")
+    if not isinstance(sequence, list):
+        return []
     periods = []
     for row in sequence:
         if not isinstance(row, dict):
@@ -157,6 +176,16 @@ def _narayana_periods(module: dict) -> list[dict]:
     return periods
 
 
+def _yogini_call_row(row: dict) -> dict:
+    """Fields the antardasha formula reads. Nested raw_period fills gaps only."""
+    source = dict(row)
+    raw = row.get("raw_period") if isinstance(row.get("raw_period"), dict) else {}
+    for name in ("yogini", "full_years", "unclipped_start_date", "end_date", "start_date"):
+        if source.get(name) is None and raw.get(name) is not None:
+            source[name] = raw[name]
+    return source
+
+
 def _yogini_periods(module: dict, rows: list[dict]) -> list[dict]:
     if any(isinstance(row.get("antardasha"), list) and row.get("antardasha") for row in rows):
         return rows
@@ -168,18 +197,39 @@ def _yogini_periods(module: dict, rows: list[dict]) -> list[dict]:
     birth = None
     if rows:
         birth = rows[0].get("start_date")
-    if not birth:
-        return rows
+        if birth is None:
+            raw = rows[0].get("raw_period") if isinstance(rows[0].get("raw_period"), dict) else {}
+            birth = raw.get("start_date")
     attached = []
     for row in rows:
-        copied = dict(row)
+        copied = deepcopy(row)
+        source = _yogini_call_row(copied)
+        missing = [
+            name for name in ("yogini", "full_years", "unclipped_start_date", "end_date")
+            if source.get(name) is None
+        ]
+        if not birth or missing:
+            logger.warning(
+                "yogini antardasha left off for %s: missing %s",
+                copied.get("yogini"),
+                ",".join(missing) or "start_date",
+            )
+            attached.append(copied)
+            continue
         try:
-            copied["antardasha"] = yogini_antardashas(row, birth_dt=birth, year_days=float(year_days))
+            children = yogini_antardashas(source, birth_dt=birth, year_days=float(year_days))
         except (KeyError, TypeError, ValueError) as exc:
-            raise ValueError(f"yogini antardasha failed for {row.get('yogini')}: {exc}") from exc
-        for child in copied["antardasha"]:
-            child["parent_lord"] = row.get("planet") or row.get("lord")
-            child["main"] = row.get("yogini")
+            logger.warning(
+                "yogini antardasha left off for %s: %s",
+                copied.get("yogini"),
+                exc,
+            )
+            attached.append(copied)
+            continue
+        for child in children:
+            child["parent_lord"] = copied.get("planet") or copied.get("lord")
+            child["main"] = copied.get("yogini")
+        copied["antardasha"] = children
         attached.append(copied)
     return attached
 
@@ -195,15 +245,19 @@ def enrich_master_family(key: str, module: dict | None, periods: list | None) ->
     if key == "vimshottari":
         return rows
     if key == "narayana":
-        return _narayana_periods(module)
+        return _narayana_periods(module, rows)
     if key == "yogini":
         return _yogini_periods(module, rows)
     if key == "kala_chakra":
         if not rows:
             return rows
-        enriched, _ad, _pd = attach_kalachakra_simplified_children(
-            rows,
-            module.get("total_cycle"),
-        )
+        try:
+            enriched, _ad, _pd = attach_kalachakra_simplified_children(
+                rows,
+                module.get("total_cycle"),
+            )
+        except (KeyError, TypeError, ValueError) as exc:
+            logger.warning("kalachakra children left off: %s", exc)
+            return rows
         return enriched
     return rows

scripts/annual_tajika_pack.py

@@ -65,8 +65,6 @@ def build_annual_tajika_pack(
         "tajika_named_yogas": _tajika_named_yogas_field(solar_report),
         "tajika_strength": _tajika_strength_field(solar_report if isinstance(solar_report, dict) else {}),
         "sahams": sahams,
-        "annual_houses": solar_report.get("annual_houses") if isinstance(solar_report, dict) else None,
-        "tripataki_chakra": solar_report.get("tripataki_chakra") if isinstance(solar_report, dict) else None,
         "mudda_dasha": mudda_dasha,
         "patyayini_dasha": patyayini_dasha,
         "monthly_windows": _monthly_windows(mudda_dasha),
@@ -80,6 +78,9 @@ def build_annual_tajika_pack(
         ),
         "exports": {},
         "audit": _audit(profile, solar_report, tajika_report, pyjhora_replay, vedastro_reference),
+        # New full-data fields stay after the original contract keys (K4).
+        "annual_houses": solar_report.get("annual_houses") if isinstance(solar_report, dict) else None,
+        "tripataki_chakra": solar_report.get("tripataki_chakra") if isinstance(solar_report, dict) else None,
     }
     pack["exports"] = _exports(pack)
     return pack

tests/test_annual_tajika_pack.py

@@ -52,6 +52,11 @@ def test_builds_annual_tajika_pack_contract() -> None:
         "report_sections",
         "exports",
         "audit",
+        # 原值：键序止于 audit，sahams 之后直接是 mudda_dasha。
+        # 新值：上面的原序不变；annual_houses、tripataki_chakra 加在 audit 之后。
+        # 原因：完整数据版的年运新字段曾插在 mudda_dasha 前面，打乱既有键序（K4）。
+        "annual_houses",
+        "tripataki_chakra",
     ]
     assert pack["schema"] == "jyotish.annual_tajika_pack.v1"
     assert pack["profile"]["profile_id"]

tests/test_dasha_master_children.py

@@ -2,6 +2,7 @@
 
 from __future__ import annotations
 
+from copy import deepcopy
 from datetime import datetime
 
 from scripts.annual_native.dasha_children import (
@@ -142,3 +143,79 @@ def test_enrich_does_not_invent_children_for_vimshottari() -> None:
     rows = enrich_master_family("vimshottari", {"timeline": []}, [{"lord": "Sun", "years": 6}])
     assert rows == [{"lord": "Sun", "years": 6}]
     assert "antardasha" not in rows[0]
+
+
+def test_enrich_keeps_existing_keys_and_values_on_every_system() -> None:
+    """Child fields may be added. Keys and values that were already there stay."""
+    yogini_row = {
+        "yogini": "Mangala",
+        "planet": "Moon",
+        "lord": "Moon",
+        "full_years": 1,
+        "years": 1.0,
+        "start_date": "1990-01-01T00:00:00",
+        "end_date": "1991-01-01T00:00:00",
+        "unclipped_start_date": "1990-01-01T00:00:00",
+        "marker": "keep-yogini",
+    }
+    narayana_module = {
+        "mahadasha_sequence": [{
+            "sign": "Aries", "lord": "Mars", "years": 7,
+            "start_age": 0, "end_age": 7, "period_key": "md-0",
+        }],
+        "dense_boundaries": [{
+            "level": "AD", "sign": "Taurus", "lord": "Venus",
+            "duration_years": 1, "start_age": 0, "end_age": 1,
+            "parent_md": "Aries", "parent_period_key": "md-0",
+        }],
+    }
+    narayana_row = {
+        "sign": "Aries", "lord": "Mars", "years": 7,
+        "start_age": 0, "end_age": 7, "period_key": "md-0",
+        "marker": "keep-narayana",
+    }
+    kala_row = {
+        "lord": "Sun", "rashi": "Aries", "years": 7, "full_years": 7,
+        "start_date": "1990-01-01T00:00:00", "end_date": "1997-01-01T00:00:00",
+        "marker": "keep-kala",
+    }
+    plain = {"lord": "Sun", "years": 6, "full_years": 6, "marker": "keep-plain"}
+    samples = {
+        "vimshottari": ({}, [{**plain, "marker": "keep-vim"}]),
+        "yogini": ({"dasha_year_days": 365.25}, [yogini_row]),
+        "narayana": (narayana_module, [narayana_row]),
+        "kala_chakra": ({"total_cycle": 7}, [kala_row]),
+        "ashtottari": ({}, [{**plain, "marker": "keep-ashtottari"}]),
+        "tribhagi": ({}, [{**plain, "marker": "keep-tribhagi"}]),
+        "tribhagi_40": ({}, [{**plain, "marker": "keep-tribhagi-40"}]),
+        "shodashottari": ({}, [{**plain, "marker": "keep-shodashottari"}]),
+        "dwadashottari": ({}, [{**plain, "marker": "keep-dwadashottari"}]),
+        "dwisaptatisama": ({}, [{**plain, "marker": "keep-dwisaptatisama"}]),
+        "shattrimshatsama": ({}, [{**plain, "marker": "keep-shattrimshatsama"}]),
+        "panchottari": ({}, [{**plain, "marker": "keep-panchottari"}]),
+        "satabdika": ({}, [{**plain, "marker": "keep-satabdika"}]),
+        "chaturshitisama": ({}, [{**plain, "marker": "keep-chaturshitisama"}]),
+        "shastihayani": ({}, [{**plain, "marker": "keep-shastihayani"}]),
+    }
+    for key, (module, rows) in samples.items():
+        before = deepcopy(rows)
+        enriched = enrich_master_family(key, module, rows)
+        assert rows == before
+        assert len(enriched) == len(before)
+        for old, new in zip(before, enriched):
+            for field, value in old.items():
+                assert field in new
+                assert new[field] == value
+
+
+def test_yogini_enrich_keeps_the_row_when_full_years_is_missing() -> None:
+    row = {
+        "yogini": "Mangala",
+        "planet": "Moon",
+        "years": 1,
+        "start_date": "1990-01-01",
+        "end_date": "1991-01-01",
+    }
+    enriched = enrich_master_family("yogini", {}, [row])
+    assert enriched[0] == row
+    assert "antardasha" not in enriched[0]