diff --git a/CHANGELOG.md b/CHANGELOG.md index 4a213016..f51751c6 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -2,10 +2,10 @@ ## 2026-11-05 — #166 占星师 11-05 答复：待核口径加标签，严格 Gajakesari 不再把凶性水星算作吉星支持，新增 PVR 双宫主 Arudha 对照 -- Shadbala 结果的 `method_variants` 写明两项尚待占星师核定的口径：七分盘尊位力本垣 45 分暂只在 D1 按度数段计（`saptavargaja_moolatrikona`），水星、木星相照暂按四分之一计（`drik_mercury_jupiter`）；水星自然吉凶口径改名，写明不计交点是本站的实现选择、亏月计入。数值都不变（BUG-1613、BUG-1614）。 -- 分盘里按分盘内度数判本垣（全读 D9 表、咨询 D9 摘要）标为待核；此前文档里的「V.P. Jain 例题」更正为 PyJHora 测试转录的夹具（BUG-1613）。 -- Gajakesari 严格定义：与太阳、火星、土星或亏月同座的水星不再算作吉星支持，与 Shadbala 用同一条规则（BUG-1615）。Matrunasa 199 的效果文字改为它自己的条件，198 写明凶星集合、交点不给相照是保守口径（BUG-1616）。 -- 新增天蝎 / 水瓶 Arudha 的 PVR 双宫主比较，只作参考，不进默认 A10 / AL / UL、星盘页和生时校正（BUG-1617）。 +- Shadbala 结果的 `method_variants` 写明两项尚待占星师核定的口径：七分盘尊位力本垣 45 分暂只在 D1 按度数段计（`saptavargaja_moolatrikona`），水星、木星相照暂按四分之一计（`drik_mercury_jupiter`）；水星自然吉凶口径改名，写明不计交点是本站的实现选择、亏月计入。数值都不变（BUG-1622、BUG-1623）。 +- 分盘里按分盘内度数判本垣（全读 D9 表、咨询 D9 摘要）标为待核；此前文档里的「V.P. Jain 例题」更正为 PyJHora 测试转录的夹具（BUG-1622）。 +- Gajakesari 严格定义：与太阳、火星、土星或亏月同座的水星不再算作吉星支持，与 Shadbala 用同一条规则（BUG-1624）。Matrunasa 199 的效果文字改为它自己的条件，198 写明凶星集合、交点不给相照是保守口径（BUG-1625）。 +- 新增天蝎 / 水瓶 Arudha 的 PVR 双宫主比较，只作参考，不进默认 A10 / AL / UL、星盘页和生时校正（BUG-1626）。 - 生时校正分数不变（仍为 scoring-15）；冻结记录按 ERR-110 更新为 `astrologer_166c_2026_11_05`，密封 20 例与报时 900 次回放逐条相同。Skill 版本不变；没有数据库迁移。 ## 2026-11-04 — #166 第 9 条：Narayana 大运默认改用 Sanjay Rath 版 diff --git a/docs/BUG_HISTORY.md b/docs/BUG_HISTORY.md index 1831123c..0f287fcf 100644 --- a/docs/BUG_HISTORY.md +++ b/docs/BUG_HISTORY.md @@ -22075,7 +22075,7 @@ - 复发自：无 - 修复版本：分支 `feat/global-boot-loader`。 -## BUG-1613 | 七分盘尊位力的本垣 45 分「只在 D1、只按度数段」被当作 BPHS 定案；「V.P. Jain 例题」实为 PyJHora 测试转录 +## BUG-1622 | 七分盘尊位力的本垣 45 分「只在 D1、只按度数段」被当作 BPHS 定案；「V.P. Jain 例题」实为 PyJHora 测试转录 - 状态：resolved（只加标签、注释与测试，数值不变；未部署） - 首次发现 / 最近更新：2026-11-05（#166 占星师答复，Gitea #166 评论 39249 第 1 项，对应 39236 第 1 项、问题清单第 17 / 19 条）/ 2026-11-05 @@ -22085,13 +22085,13 @@ - 根因：Santhanam 版 27.2–4（印刷页 265）确认本垣 45 分，并说同样的值适用于其余六个分盘；「只在 D1」不是这段译文写明的例外，D1 整座还是度数段、其他六盘怎么计（分盘星座还是分盘内度数、D30 如何换算）都没有解决。现行实现是临时选择，却没有标出。V.P. Jain 一项来自 PyJHora `pvr_tests.py::shadbala_VPJainBook_tests`，没有书名、版本和原页（ERR-081 已记过，数据核对文档又写成了「例题」）。 - 修复：`SAPTAVARGAJA_MOOLATRIKONA_CONVENTION = "d1_degree_band_only_provisional"`，写进 `method_variants.saptavargaja_moolatrikona`，注释写明是临时实现、待核，不是 BPHS 定案；`jyotish_engine.VARGA_MOOLATRIKONA_BASIS = "in_varga_degree_band_pending_review"`，D9 扩展表每行加 `moolatrikona_basis`，咨询 D9 摘要加同名字段（数据卡不复制这个字段）。三处数据核对 / 竞品文档把「例题 / 书例」改为「PyJHora 测试转录的夹具」并注明更正。`astrologer_questions.md` 末尾加 2026-11-05 七项落地状态表。没有改任何计算。 - 验证：`tests/test_astrologer_rulings_166c.py`：标签值、`method_variants` 输出、D1 度数段内 45 / 段外 30、其他六盘从不得 45（行为未变）、D9 摘要标签且狮子 25° 太阳仍判自有星座；`test_shadbala_service_api.py` 期望的 `method_variants` 加两项（原值 / 新值 / 原因注释）。 -- 冻结身份（ERR-110，本轮全部 shadbala.py 改动一次重冻）：`shadbala.py` 在冻结身份里（holdout_v3 `frozen_scoring.files` 与扩展身份）。本轮所有 shadbala.py 改动（BUG-1613、BUG-1614 的标签，BUG-1615 共用的水星规则函数）放在同一提交 `5ffef30a`；在本分支全部代码提交之后（`09fd3a42`）按 ERR-110 改 FREEZE / REPORT 常量、`--freeze` 后回放，标签 `astrologer_166c_2026_11_05`，合同轮换，上一份 `astrologer_166b_2026_11_04` 逐字节保留。校正身份仍是 `rectification-v5-matrix-scoring-15`，没有升 ALGORITHM_VERSION。与 `astrologer_166b_2026_11_04` 逐条逐字段比：密封 20 条全同，`metrics` 不变（头名率 0.40、前三率 0.45、平均分钟误差 6.95）；报时 900 条全同，`summary` 不变（±15 / 30 / 60 头名命中 4 / 3 / 0，前三 31 / 34 / 30，±15 交付覆盖 219，头名平均分钟误差 20.52）。相对 09-20 历史报告 changed_trial_count：密封 10 / 20、报时 731 / 900，与上一份相同。未调参。重冻提交 `39d7cb5e`。`jaimini.py`、`functional_benefics.py`、`rectification/*` 没有改动。 +- 冻结身份（ERR-110，本轮全部 shadbala.py 改动一次重冻）：`shadbala.py` 在冻结身份里（holdout_v3 `frozen_scoring.files` 与扩展身份）。本轮所有 shadbala.py 改动（BUG-1622、BUG-1623 的标签，BUG-1624 共用的水星规则函数）放在同一提交 `5ffef30a`；在本分支全部代码提交之后（`09fd3a42`）按 ERR-110 改 FREEZE / REPORT 常量、`--freeze` 后回放，标签 `astrologer_166c_2026_11_05`，合同轮换，上一份 `astrologer_166b_2026_11_04` 逐字节保留。校正身份仍是 `rectification-v5-matrix-scoring-15`，没有升 ALGORITHM_VERSION。与 `astrologer_166b_2026_11_04` 逐条逐字段比：密封 20 条全同，`metrics` 不变（头名率 0.40、前三率 0.45、平均分钟误差 6.95）；报时 900 条全同，`summary` 不变（±15 / 30 / 60 头名命中 4 / 3 / 0，前三 31 / 34 / 30，±15 交付覆盖 219，头名平均分钟误差 20.52）。相对 09-20 历史报告 changed_trial_count：密封 10 / 20、报时 731 / 900，与上一份相同。未调参。重冻提交 `39d7cb5e`。`jaimini.py`、`functional_benefics.py`、`rectification/*` 没有改动。 - 防复发：口径名写进输出；更正处注明出处性质。引用外部数值时先写清是书页、作者教材还是软件测试夹具。 - 相关记录：BUG-1583、BUG-1584、BUG-1575、ERR-081、#166。 - 复发自：ERR-081（夹具被写成出版例题；这次出在数据核对文档里，ERR-081 只约束了基准文件的命名，没有约束研究文档的措辞） - 修复版本：分支 `codex/astrologer-166c-20261105`，`5ffef30a`、`54e654db`；重冻 `39d7cb5e`（未推送）。 -## BUG-1614 | 水星自然吉凶口径名写成「BPHS 3.11、不含交点」，Drik 水星 / 木星相照的四分之一没有标口径 +## BUG-1623 | 水星自然吉凶口径名写成「BPHS 3.11、不含交点」，Drik 水星 / 木星相照的四分之一没有标口径 - 状态：resolved（只加标签、注释与测试，数值不变；未部署） - 首次发现 / 最近更新：2026-11-05（#166 评论 39249 第 2、4、7 项，对应问题清单第 23、25、24 条）/ 2026-11-05 @@ -22099,15 +22099,15 @@ - 用户现象：`method_variants.mercury_nature = bphs_3_11_any_malefic_same_sign_no_nodes`，读起来像 BPHS 规定排除交点；Drik 里水星、木星发出的相照按四分之一计，输出没有口径；26.11–12 特殊相照只引网络英译。 - 触发条件：任何盘（标签问题）。 - 根因：Santhanam 版 3.10–11（印刷页 27）的自然凶星包含太阳、火星、土星、亏月、罗睺、计都，随后说水星与凶星结合则凶；排除罗睺、计都是本仓的实现选择（Drik 只看七星），不是 BPHS；印刷页 28 的译者注（亏月与水星结合两者皆吉）是另一层，未采用。27.19（印刷页 284）水星 / 木星相照有三种读法：全额替换四分之一（1 倍）、Santhanam 英译字面「四分之一之外再加全额」（吉性 1.25 倍；原始 40 时木星或吉性水星 +10 + 40 = +50，凶性水星若附加项仍为正则 −10 + 40 = +30，凶性水星的符号待核）、Raman 的四分之一；占星师未裁定。26.11–12 已由占星师对照 Santhanam 印刷页 256 确认（火星 [180, 210) 为 60，木星 [120, 150)、[240, 270) 为 60 − 段内度数），左闭右开是工程约定。 -- 修复：`MERCURY_NATURE_CONVENTION` 改名为 `same_sign_any_natural_malefic_waning_moon_included_nodes_excluded_implementation_choice`（原值 / 新值 / 原因注释），行为不变；规则抽成唯一实现 `mercury_malefic_companions()`，`classify_drik_planets` 调用它（BUG-1615 的严格 Gajakesari 也调用它）。`DRIK_MERCURY_JUPITER_CONVENTION = "quarter_only_raman_pending"` 写进 `method_variants.drik_mercury_jupiter`，注释列三种读法与 40 的算例；本轮不加附加法。`_sphuta_drishti_virupas` 注释补 Santhanam 印刷页 256 与端点约定。 +- 修复：`MERCURY_NATURE_CONVENTION` 改名为 `same_sign_any_natural_malefic_waning_moon_included_nodes_excluded_implementation_choice`（原值 / 新值 / 原因注释），行为不变；规则抽成唯一实现 `mercury_malefic_companions()`，`classify_drik_planets` 调用它（BUG-1624 的严格 Gajakesari 也调用它）。`DRIK_MERCURY_JUPITER_CONVENTION = "quarter_only_raman_pending"` 写进 `method_variants.drik_mercury_jupiter`，注释列三种读法与 40 的算例；本轮不加附加法。`_sphuta_drishti_virupas` 注释补 Santhanam 印刷页 256 与端点约定。 - 验证：`tests/test_astrologer_rulings_166c.py`：亏月同座使水星为凶、盈月不使、交点不使；共用函数与 Drik 分类在 24 × 12 × 2 个位置上一致；木星 180° 相照太阳 Drik = 60 / 4 = 15；特殊相照端点 149.99 / 150 / 150.01、179.99 / 180 / 209.99 / 210 / 210.01、269.99 / 270 / 270.01，以及 `round(angle, 2)` 把 149.996 吸到 150（值 0）、149.994 仍为 30.01。`test_astrologer_rulings_166.py`、`test_shadbala_drik_deterministic.py`、`test_vp_jain_shadbala_benchmark.py` 照旧通过。 -- 冻结身份：同 BUG-1613（`astrologer_166c_2026_11_05`，回放逐条相同）。 +- 冻结身份：同 BUG-1622（`astrologer_166c_2026_11_05`，回放逐条相同）。 - 防复发：口径名不再写成来源已规定的样子；「实现选择」「待核」写进名字或注释。 - 相关记录：BUG-1585、BUG-1586、BUG-1576、#166。 - 复发自：无 - 修复版本：分支 `codex/astrologer-166c-20261105`，`5ffef30a`；重冻 `39d7cb5e`（未推送）。 -## BUG-1615 | 严格 Gajakesari 把与自然凶星同座的水星仍算作吉星支持 +## BUG-1624 | 严格 Gajakesari 把与自然凶星同座的水星仍算作吉星支持 - 状态：resolved（引擎 + 回归测试；未部署） - 首次发现 / 最近更新：2026-11-05（#166 评论 39249 第 6 项，对应 39236 第 6 项、问题清单 I 报告第 5 条）/ 2026-11-05 @@ -22117,13 +22117,13 @@ - 根因：BUG-1596 的严格定义把金星、水星无条件当吉星。PVR §11.6（印刷页 124–125，PDF 136–137）严格定义要求吉星结合或相照（二选一），三个主条件 AND；占星师：受凶影响的水星不得无条件提供吉星支持，应服从所选的自然吉凶规则。 - 修复：沙盒函数 `mercury_malefic_companions()` 懒加载并调用 `shadbala.mercury_malefic_companions`（与 Drik 同一规则、同一标签：与太阳、火星、土星或亏月同座即凶，交点不计；无循环导入），不另立规则。严格版把凶性水星从吉星支持里去掉；金星仍支持时触发项写明「水星与 X 同宫，按凶星计，不算吉星支持」。强木星变体：它判断「严格版不成立」时用同一规则，两条规则保持互斥；变体本身不用吉星支持，所以这条规则对它没有别的作用。保留并标待核：敌宫按自然友敌（不是五重友敌）、月亮不作支持星（占星师未认可）。 - 验证：`tests/test_astrologer_rulings_166c.py`：凶性水星是唯一支持时严格版不成立、只剩月木角宫关系；吉性水星仍成立；亏月同座使水星凶、严格版不成立；交点同座不影响；金星支持时写出被排除的水星；木星入旺巨蟹、唯一支持是凶性水星时严格版不成立、强木星变体成立。`test_astrologer_rulings_166_batch2.py`、`test_yoga_audit_i_20261104.py`（静态扫描：新辅助函数已在命名空间）、全部 `tests/test_yoga*.py` 通过。60 张标准测试盘（`references/planet_positions_60.json`）严格版成立 4 → 4，没有一张变；三张公开盘证据卡与 9 位名人回测捕获里没有 Gajakesari 成立，重新捕获后瑜伽输出无变化。 -- 冻结身份：yoga_engine.py、yoga_rules.json 不在冻结身份里；共用函数所在的 shadbala.py 改动见 BUG-1613 的重冻记录（回放逐条相同）。 +- 冻结身份：yoga_engine.py、yoga_rules.json 不在冻结身份里；共用函数所在的 shadbala.py 改动见 BUG-1622 的重冻记录（回放逐条相同）。 - 防复发：同一条自然吉凶规则只有一个实现，瑜伽沙盒调用它，不在规则表里重写。 -- 相关记录：BUG-1596、BUG-1614、BUG-1585、#166。 +- 相关记录：BUG-1596、BUG-1623、BUG-1585、#166。 - 复发自：无 - 修复版本：分支 `codex/astrologer-166c-20261105`，`e9ff1f61`（共用函数 `5ffef30a`）（未推送）。 -## BUG-1616 | Matrunasa 199 的效果文字写的是 198 的内容；198 的交点、凶星集合与版本页码没有标口径 +## BUG-1625 | Matrunasa 199 的效果文字写的是 198 的内容；198 的交点、凶星集合与版本页码没有标口径 - 状态：resolved（规则表文字与标签；检测不变；未部署） - 首次发现 / 最近更新：2026-11-05（#166 评论 39249 第 5 项）/ 2026-11-05 @@ -22139,7 +22139,7 @@ - 复发自：BUG-1597（拆分时漏改 199 的效果文字） - 修复版本：分支 `codex/astrologer-166c-20261105`，`8bba999f`（未推送）。 -## BUG-1617 | 天蝎 / 水瓶的 Arudha 只用固定宫主（火星、土星），没有 PVR 双宫主择强的对照 +## BUG-1626 | 天蝎 / 水瓶的 Arudha 只用固定宫主（火星、土星），没有 PVR 双宫主择强的对照 - 状态：resolved（新增参考模块，未进默认路径；未部署） - 首次发现 / 最近更新：2026-11-04（#166 第 28 条，BUG-1588 记为未实现）/ 2026-11-05（评论 39249 第 3 项） @@ -22149,7 +22149,7 @@ - 根因：只实现了固定宫主表；占星师 11-04 认为 PVR 口径 Arudha 应双主择强，但不要全系统替换，11-05 核对了比较顺序。 - 修复：`select_pvr_arudha_lord()` 按 PVR《Vedic Astrology: An Integrated Approach》§15.5.1（印刷页 201–203，PDF 213–215），前一级决出即停：1 只有一颗宫主在该星座内时取另一颗；2 同座伴星多者（不计自身，上升点不算，九曜含交点）；3 木星、水星、所在星座宫主三个角色以结合或星座相照支持多者（同一星承担两个角色计两次）；4 入旺；5 所在星座双体 > 固定 > 变动；6 座内度数高者，交点按 30° − 度数。完全并列、所在星座宫主又是天蝎 / 水瓶双主且递归回到自身、缺经度时返回 blocked 与原因，两颗候选都保留，不强行选一个。输出获选者或 blocked、两颗候选、决出层级、逐级 trace 与工程选择清单。工程选择（不是教材原文，待核）：并列 / 递归的 blocked 处理；宫主角色由候选星自身充当时照算（与 `jaimini._co_lord_strength` 一致）；交点入旺取罗睺金牛、计都天蝎。`pvr_arudha_reference()` 对天蝎 / 水瓶两宫并列给出默认 pada 与 PVR pada，pada 算法直接调用 `jaimini.calc_arudha_pada_for_house`（唯一的 Arudha 实现，含 §9.2 例外），只把宫主经度换成获选者。标签「PVR 双宫主比较（参考，未进默认）」，`in_default_paths = false`。`jaimini.py`（含 Chara 双主选择器 `_co_lord_strength`）一字未改。 - 验证：`tests/test_arudha_lord_selection_pvr.py` 16 条：第 1 级两个方向与两星都在本宫时继续；第 2 级交点计入伴星、上升点 / Ascendant 键不计；第 3 级木星同时任木星与宫主角色计 2；第 4 级入旺；第 5 级双体胜固定；第 6 级交点 30° − 度数改变胜负；完全并列 blocked；天蝎 / 水瓶互为宫主的递归 blocked；缺经度 blocked、非双主星座拒绝；参考行在获选者改变 pada 时标 `differs_from_default`、默认 pada 仍按火星；校正与默认路径的模块不引用本模块；CLI 带出参考字段且默认 `arudha_padas` 不变；`jaimini.py` 字节与冻结身份记录的 sha256 相同。 -- 冻结身份：新模块不在冻结身份里，也不被冻结文件引用；`jaimini.py` 未改。重冻（BUG-1613）回放逐条相同。占星师：「进入校正打分仍需独立验收」，本轮不接入。 +- 冻结身份：新模块不在冻结身份里，也不被冻结文件引用；`jaimini.py` 未改。重冻（BUG-1622）回放逐条相同。占星师：「进入校正打分仍需独立验收」，本轮不接入。 - 防复发：默认路径不引用本模块由测试锁住；取法逐级写进 trace，工程选择随结果输出。 - 相关记录：BUG-1588、BUG-1566、#166。 - 复发自：无 diff --git a/docs/research/data_audit_2026_11_04/astrologer_questions.md b/docs/research/data_audit_2026_11_04/astrologer_questions.md index 0b96e6b9..90d109ba 100644 --- a/docs/research/data_audit_2026_11_04/astrologer_questions.md +++ b/docs/research/data_audit_2026_11_04/astrologer_questions.md @@ -34,7 +34,7 @@ ## 五、Shadbala 位置力与相照力（J 报告，`audit_j_shadbala_sthana_dig_drik.md`） 16. 七分盘尊位力分值：Raman 22.5/7.5/3.75/1.875，还是 BPHS 27.3（梵文）20/10/4/2。 -17. D1 本垣按度数范围，还是整个星座都算（PyJHora、V.P. Jain 例题按整个星座）。〔2026-11-05 更正：「V.P. Jain 例题」实为 PyJHora 测试 `pvr_tests.py::shadbala_VPJainBook_tests` 转录的数值夹具，没有书名、版本和原页，不是书中例题（ERR-081 复发，BUG-1613）。〕 +17. D1 本垣按度数范围，还是整个星座都算（PyJHora、V.P. Jain 例题按整个星座）。〔2026-11-05 更正：「V.P. Jain 例题」实为 PyJHora 测试 `pvr_tests.py::shadbala_VPJainBook_tests` 转录的数值夹具，没有书名、版本和原页，不是书中例题（ERR-081 复发，BUG-1622）。〕 18. 月亮、水星本垣从 4°、16° 起，还是按 BPHS 从 3°、15° 起。 19. 本垣只在 D1 计，还是其他分盘也计。 20. 临时友敌取 D1，还是每张分盘各自计算。 @@ -64,14 +64,14 @@ ## 2026-11-05 占星师答复（#166 评论 39249）落地状态 -占星师说明本轮没有把七项都标成「已正式核定」。下表只记录本仓据此做了什么；「待核」的项保留现行算法，只加标签，不当作已核定的 BPHS 规则。生时校正分数不变（`astrologer_166c_2026_11_05` 重冻回放逐条相同，见 BUG-1613～1617）。 +占星师说明本轮没有把七项都标成「已正式核定」。下表只记录本仓据此做了什么；「待核」的项保留现行算法，只加标签，不当作已核定的 BPHS 规则。生时校正分数不变（`astrologer_166c_2026_11_05` 重冻回放逐条相同，见 BUG-1622～1626）。 | 39236 编号 | 对应本表 | 占星师结论 | 本仓处理 | |---|---|---|---| -| 1 | 17、19 | 45 分本版正文可确认，且译文说同值适用其余六盘；D1 整座 / 度数段、其他六盘怎么计、分盘度数与 D30 换算都待核。「V.P. Jain 例题」实为 PyJHora 测试转录的夹具 | 保留 D1 度数段，`method_variants.saptavargaja_moolatrikona = d1_degree_band_only_provisional`；D9 按分盘内度数判本垣标 `in_varga_degree_band_pending_review`；更正 V.P. Jain 出处（BUG-1613） | -| 2 | 23 | Santhanam 27.19（印刷页 284）字面是四分之一之外再加全额：吉性 1.25 倍；凶性水星的附加项符号待核；1 倍法、四分之一法各留标记 | 仍按四分之一，`method_variants.drik_mercury_jupiter = quarter_only_raman_pending`；本轮不加附加法（BUG-1614） | -| 3 | 28 | PVR §15.5.1（印刷页 201–203，PDF 213–215）六级比较可确认；上升点不算伴星；交点计入可作本教材口径；并列与递归保留两候选是工程处理；进校正打分须另行验收 | 新模块 `scripts/arudha_reference/pvr_dual_lord.py`，只作参考，不进默认 A10 / AL / UL、星盘页与校正（BUG-1617） | -| 4 | 25 | 正文自然凶星含罗睺、计都，排除交点不能称 BPHS；符合条件的亏月在集合内；印刷页 28 译注（亏月与水星结合两者皆吉）是另一层 | 行为不变，`MERCURY_NATURE_CONVENTION` 改名写明排除交点是实现选择、亏月计入；译注层未采用（BUG-1614） | -| 5 | 6（Matrunasa） | 198 三种情况为「或」；交点不给相照可暂时保留，但须标保守工程口径、待核；写明凶星集合；版本页码待核 | 检测不变，加标签；199 的效果文字原来写的是 198 的内容，已改（BUG-1616） | -| 6 | 5（Gajakesari） | 严格版：结合或相照二选一，三个主条件 AND；受凶影响的水星不得无条件算吉星支持；自然敌宫与不含月亮均待核 | 严格版不再把与自然凶星同座的水星算作支持（与第 4 项同一规则、同一函数）；自然敌宫、月亮不作支持保留并标待核（BUG-1615） | -| 7 | 24 | 26.11–12 两段与 Santhanam 印刷页 256 一致；左闭右开是工程约定 | 逻辑不变，补端点 ±0.01 测试与页码出处（BUG-1614） | +| 1 | 17、19 | 45 分本版正文可确认，且译文说同值适用其余六盘；D1 整座 / 度数段、其他六盘怎么计、分盘度数与 D30 换算都待核。「V.P. Jain 例题」实为 PyJHora 测试转录的夹具 | 保留 D1 度数段，`method_variants.saptavargaja_moolatrikona = d1_degree_band_only_provisional`；D9 按分盘内度数判本垣标 `in_varga_degree_band_pending_review`；更正 V.P. Jain 出处（BUG-1622） | +| 2 | 23 | Santhanam 27.19（印刷页 284）字面是四分之一之外再加全额：吉性 1.25 倍；凶性水星的附加项符号待核；1 倍法、四分之一法各留标记 | 仍按四分之一，`method_variants.drik_mercury_jupiter = quarter_only_raman_pending`；本轮不加附加法（BUG-1623） | +| 3 | 28 | PVR §15.5.1（印刷页 201–203，PDF 213–215）六级比较可确认；上升点不算伴星；交点计入可作本教材口径；并列与递归保留两候选是工程处理；进校正打分须另行验收 | 新模块 `scripts/arudha_reference/pvr_dual_lord.py`，只作参考，不进默认 A10 / AL / UL、星盘页与校正（BUG-1626） | +| 4 | 25 | 正文自然凶星含罗睺、计都，排除交点不能称 BPHS；符合条件的亏月在集合内；印刷页 28 译注（亏月与水星结合两者皆吉）是另一层 | 行为不变，`MERCURY_NATURE_CONVENTION` 改名写明排除交点是实现选择、亏月计入；译注层未采用（BUG-1623） | +| 5 | 6（Matrunasa） | 198 三种情况为「或」；交点不给相照可暂时保留，但须标保守工程口径、待核；写明凶星集合；版本页码待核 | 检测不变，加标签；199 的效果文字原来写的是 198 的内容，已改（BUG-1625） | +| 6 | 5（Gajakesari） | 严格版：结合或相照二选一，三个主条件 AND；受凶影响的水星不得无条件算吉星支持；自然敌宫与不含月亮均待核 | 严格版不再把与自然凶星同座的水星算作支持（与第 4 项同一规则、同一函数）；自然敌宫、月亮不作支持保留并标待核（BUG-1624） | +| 7 | 24 | 26.11–12 两段与 Santhanam 印刷页 256 一致；左闭右开是工程约定 | 逻辑不变，补端点 ±0.01 测试与页码出处（BUG-1623） | diff --git a/docs/research/data_audit_2026_11_04/audit_j_shadbala_sthana_dig_drik.md b/docs/research/data_audit_2026_11_04/audit_j_shadbala_sthana_dig_drik.md index 3a5d7fdf..8395b387 100644 --- a/docs/research/data_audit_2026_11_04/audit_j_shadbala_sthana_dig_drik.md +++ b/docs/research/data_audit_2026_11_04/audit_j_shadbala_sthana_dig_drik.md @@ -40,7 +40,7 @@ | Uchcha | 到落陷点的弧折回 0–180°，再除以 3 | BPHS 27.1；Raman §20 与例 3、例 13 | 无（PyJHora 140/140，jyotishganit 140/140） | 正确 | | Saptavargaja：七个分盘 | D1、D2（Parashara Hora）、D3（1-5-9）、D7、D9、D12、D30（Parashara） | BPHS 27 序言；Raman 例 6 | jyotishganit 138/140 | jyotishganit 的错：D2、D3、D7、D12、D30 用循环 Dn，不是 Parashara 分法；自然友敌查表恒返回 SAMA（`get_natural_relationship` 查的是 'friends' 等键，对行星名查不到）；根本三角常量错（月亮巨蟹、金星金牛、土星摩羯）；暂时友敌拿行星的分盘星座去比宫主的本命星座 | | Saptavargaja：分值 | 45 / 30 / 22.5 / 15 / 7.5 / 3.75 / 1.875 | Raman §30（第 22–23 页），沿用 Sripati；例 9 太阳 90.000 | BPHS 27.2–3、saravali 用 45 / 30 / 20 / 15 / 10 / 4 / 2 | **具名流派差 V1** | -| Saptavargaja：本命盘根本三角 | 按度数范围（太阳狮子 0–20° 等） | BPHS 3.51–54：「the rest is his own Bhava」 | PyJHora 与 VP Jain 夹具按整个星座（5/140；VP Jain 太阳、水星各 −15）〔2026-11-05 更正：原写「例题」；这是 PyJHora 测试转录的数值夹具，不是带书名页码的书中例题，BUG-1613〕 | **流派差 V2** | +| Saptavargaja：本命盘根本三角 | 按度数范围（太阳狮子 0–20° 等） | BPHS 3.51–54：「the rest is his own Bhava」 | PyJHora 与 VP Jain 夹具按整个星座（5/140；VP Jain 太阳、水星各 −15）〔2026-11-05 更正：原写「例题」；这是 PyJHora 测试转录的数值夹具，不是带书名页码的书中例题，BUG-1622〕 | **流派差 V2** | | Saptavargaja：月亮、水星根本三角的起点 | 月亮金牛 4°，水星处女 16° | 与 VedAstro 相同；疑出自 Raman《A Manual of Hindu Astrology》，未能核到页码 | BPHS 3.51–54 是 3° 和 15° | **流派差 V3** | | Saptavargaja：分盘里是否算根本三角 | 只在本命盘算 | Raman §30：「only when it is in its Moolatrikona Rasi」 | saravali、jyotishganit 在各分盘也按星座算 | **流派差 V4** | | Saptavargaja：暂时友敌 | 按本命盘位置 | Raman 例 6 | saravali 说「两种意见都有」 | **流派差 V5** | @@ -96,7 +96,7 @@ | 编号 | 选项（本仓现状写在前面） | 出处 | 影响大小 | |---|---|---|---| | V1 | Saptavargaja 分值：Raman / Sripati 22.5 / 7.5 / 3.75 / 1.875，还是 BPHS 20 / 10 / 4 / 2 | Raman §30；BPHS 27.2–3 梵文「bhūtābdhayaḥ khāgninakhāstithyo daśa yugāḥ karāḥ」 | 每个分盘 ±2.5；Saptavargaja 合计最多约 ±17 | -| V2 | 本命盘根本三角：按度数范围，还是按整个星座 | BPHS 3.51–54；PyJHora、VP Jain 夹具（PyJHora 测试转录，非书中例题，BUG-1613）按整个星座 | 每次 15；19 张盘里出现 5 次 | +| V2 | 本命盘根本三角：按度数范围，还是按整个星座 | BPHS 3.51–54；PyJHora、VP Jain 夹具（PyJHora 测试转录，非书中例题，BUG-1622）按整个星座 | 每次 15；19 张盘里出现 5 次 | | V3 | 月亮、水星根本三角的起点：4° / 16°，还是 BPHS 的 3° / 15° | BPHS 3.51–54；BUG-1575 | 只影响落在这 1° 里的盘 | | V4 | 分盘里算不算根本三角 | Raman §30；saravali | 每次 15 | | V5 | 暂时友敌按本命盘，还是按分盘 | Raman 例 6；saravali | 中等 | @@ -140,7 +140,7 @@ V8–V11 和 Sthana 中的任何一项一旦改动，生时校正分数都会变 | 项目 | 状态 | 说明 | |---|---|---| | MEVG / Global Web Evidence | Used（部分） | 第一层：BPHS（梵文 + Santhanam 英译）、Raman GBB 原书扫描本。第二层：PyJHora、VedAstro、jyotishganit 源码。第三层：saravali（只作旁证） | -| Real Case Calibration | blocked | 只有计算示例（Raman 标准盘、VP Jain 夹具〔PyJHora 测试转录，非书中例题，BUG-1613〕），没有结果校准 | +| Real Case Calibration | blocked | 只有计算示例（Raman 标准盘、VP Jain 夹具〔PyJHora 测试转录，非书中例题，BUG-1622〕），没有结果校准 | | Functional Benefic/Malefic | not used | 只讨论计算口径，不解读盘 | | 三引擎交叉 | Used | PyJHora（同岁差 140 项）、jyotishganit（140 项，跑本地 vendored 代码）、VedAstro（只有乔布斯盘线上值 + 2024 源码） | | 预检 | `pre_work_check.py` exit 0 | | diff --git a/docs/research/pre_work_error_ledger.md b/docs/research/pre_work_error_ledger.md index a5bc8645..aaa6ffb6 100644 --- a/docs/research/pre_work_error_ledger.md +++ b/docs/research/pre_work_error_ledger.md @@ -7,7 +7,7 @@ Purpose: read this file before substantial project work. It exists to stop repea - `pre_work_check.py` 第一次 exit 1：新 worktree 没有 `.venv`，报「No usable project Python found」，六项全部没跑。用 scratchpad venv（Python 3.12 + swisseph）设 `JYOTISH_PRE_WORK_PYTHON` 重跑 exit 0、六项全部 ok。那个 venv 没有 pip，pytest 等开发依赖要用 `uv pip install --python /bin/python -r requirements-dev.txt` 装进去。 - ERR-110 重冻 `astrologer_166c_2026_11_05`（只改 shadbala.py 标签与共用函数）：密封 20 条、报时 900 条与 `astrologer_166b_2026_11_04` 逐条相同。教训：① 隔天重冻时合同顶层 `evaluated_on` 也要改成新报告的日期，`test_contract_tracks_actual_current_scorer_and_dataset_audit` 会比；② 报时回放约 2.5 分钟，密封几秒；③ 所有冻结文件改动先放一个提交，重冻放在本分支全部代码提交之后，回放才对应最终代码。 - 捕获重生成（证据卡、名人回测、card facts）：除了本单新加的标签字段，`vedastro_cross_check.official_closure_reason` 在 `foreground_optional_evidence_timeout` 与 `official_raw_response_missing` 之间来回变、`elapsed_s` 每次不同，这是环境噪声，不是代码变化。瑜伽输出没变时不提交捕获。 -- 「V.P. Jain 例题」又出现在数据核对文档里（ERR-081 复发，BUG-1613）：引用外部数值时先写清来源是书页、作者教材还是软件测试夹具。 +- 「V.P. Jain 例题」又出现在数据核对文档里（ERR-081 复发，BUG-1622）：引用外部数值时先写清来源是书页、作者教材还是软件测试夹具。 ## 2026-11-04 · 数据核对 1a / 1b（worktree `/workspace/wt-data`） diff --git a/frontend/tests/rectification-confirmation-gate.test.ts b/frontend/tests/rectification-confirmation-gate.test.ts index 1c22d45c..2f8613ff 100644 --- a/frontend/tests/rectification-confirmation-gate.test.ts +++ b/frontend/tests/rectification-confirmation-gate.test.ts @@ -137,7 +137,7 @@ const diagnosticReport = JSON.parse(readFileSync( // 原因(18)：#166 第二批（BUG-1590～1598）改了 functional_benefics.py、jaimini.py、narayana_dasha.py、shadbala.py、varga.py，scoring-15；20 条回放与上一份逐条相同，ERR-110 另存新记录。确认门行为与阈值不变。 // 原值(19)：docs/research/sealed_holdout_rerun_astrologer_166b_2026_11_04.json // 新值(19)：docs/research/sealed_holdout_rerun_astrologer_166c_2026_11_05.json -// 原因(19)：#166 占星师 2026-11-05 答复（BUG-1613～1617）只改了 shadbala.py 的标签与共用水星规则函数，数值不变，仍是 scoring-15；20 条回放与上一份逐条相同，ERR-110 另存新记录。确认门行为与阈值不变。 +// 原因(19)：#166 占星师 2026-11-05 答复（BUG-1622～1626）只改了 shadbala.py 的标签与共用水星规则函数，数值不变，仍是 scoring-15；20 条回放与上一份逐条相同，ERR-110 另存新记录。确认门行为与阈值不变。 const currentTreeRerun = JSON.parse(readFileSync( new URL("../../docs/research/sealed_holdout_rerun_astrologer_166c_2026_11_05.json", import.meta.url), "utf8", diff --git a/references/competitive-analysis-2026-06-05.md b/references/competitive-analysis-2026-06-05.md index d6ce371e..99eb1f40 100644 --- a/references/competitive-analysis-2026-06-05.md +++ b/references/competitive-analysis-2026-06-05.md @@ -22,7 +22,7 @@ - **可借鉴**：建立 `benchmark/baselines/` 目录，固化 JSON 基线 3. **Shadbala 外部校准（我们正在做）** - - 已对齐 BV Raman 书例和 VP Jain 夹具（2026-11-05 更正：VP Jain 一项是 PyJHora 测试 `shadbala_VPJainBook_tests` 转录的数值夹具，没有书名、版本和页码，不是书例；ERR-081、BUG-1613） + - 已对齐 BV Raman 书例和 VP Jain 夹具（2026-11-05 更正：VP Jain 一项是 PyJHora 测试 `shadbala_VPJainBook_tests` 转录的数值夹具，没有书名、版本和页码，不是书例；ERR-081、BUG-1622） - `get_planet_mean_longitude()` 用于 Chesta Bala - `planet_aspect_relationship_table()` 用于 Drik Bala - **对比**：我们的 Shadbala 仍是 partial，需要外部校准 diff --git a/references/rectification_sealed_holdout.v1.json b/references/rectification_sealed_holdout.v1.json index 4b29bdce..a4d37227 100644 --- a/references/rectification_sealed_holdout.v1.json +++ b/references/rectification_sealed_holdout.v1.json @@ -201,7 +201,7 @@ "verified_minute_claim_allowed": false, "metric_gates_passed": false, "events_per_case": 3, - "boundary": "Previously exposed v3 shadow fixed-protocol rerun after the #166 astrologer reply 2026-11-05 (BUG-1613..1617): scripts/shadbala.py adds labels only (saptavargaja_moolatrikona d1_degree_band_only_provisional, drik_mercury_jupiter quarter_only_raman_pending, MERCURY_NATURE_CONVENTION renamed, 26.11-12 citation) and moves the Mercury-by-association rule into mercury_malefic_companions() shared with the strict Gajakesari yoga, values unchanged; scoring identity stays rectification-v5-matrix-scoring-15 (ERR-110). Versus the 2026-11-04 astrologer_166b_2026_11_04 report, all 20 trials and metrics are identical. Versus the 2026-09-20 historical report, changed_trial_count is 10. Not a release metric. No tuning. Prior reports and freezes preserved byte-for-byte.", + "boundary": "Previously exposed v3 shadow fixed-protocol rerun after the #166 astrologer reply 2026-11-05 (BUG-1622..1626): scripts/shadbala.py adds labels only (saptavargaja_moolatrikona d1_degree_band_only_provisional, drik_mercury_jupiter quarter_only_raman_pending, MERCURY_NATURE_CONVENTION renamed, 26.11-12 citation) and moves the Mercury-by-association rule into mercury_malefic_companions() shared with the strict Gajakesari yoga, values unchanged; scoring identity stays rectification-v5-matrix-scoring-15 (ERR-110). Versus the 2026-11-04 astrologer_166b_2026_11_04 report, all 20 trials and metrics are identical. Versus the 2026-09-20 historical report, changed_trial_count is 10. Not a release metric. No tuning. Prior reports and freezes preserved byte-for-byte.", "report_sha256": "fc3ab48bb7418e35f8f6e40b0aca8fa12c0e9004159e6a5a06b42a2c253cd3e5", "freeze_record_sha256": "cbda0f099b520c12b6d0c2c3cbd4e1b285fff5fd424cf75e15e06689f32390d6", "replay_started_at_utc": "2026-11-05T05:54:04.133903+00:00", @@ -209,7 +209,7 @@ "previous_fixed_protocol_rerun": { "report_path": "docs/research/sealed_holdout_rerun_astrologer_166b_2026_11_04.json", "freeze_record_path": "docs/research/sealed_holdout_rerun_astrologer_166b_2026_11_04.freeze.json", - "superseded_reason": "#166 astrologer reply 2026-11-05 (BUG-1613..1617): scripts/shadbala.py adds labels only (saptavargaja_moolatrikona d1_degree_band_only_provisional, drik_mercury_jupiter quarter_only_raman_pending, MERCURY_NATURE_CONVENTION renamed, 26.11-12 citation) and moves the Mercury-by-association rule into mercury_malefic_companions() shared with the strict Gajakesari yoga, values unchanged; scoring identity stays rectification-v5-matrix-scoring-15 (ERR-110) changed frozen identity files; the sealed replay output is unchanged, and ERR-110 stored a new record. Prior astrologer_166b_2026_11_04 report and freeze preserved byte-for-byte." + "superseded_reason": "#166 astrologer reply 2026-11-05 (BUG-1622..1626): scripts/shadbala.py adds labels only (saptavargaja_moolatrikona d1_degree_band_only_provisional, drik_mercury_jupiter quarter_only_raman_pending, MERCURY_NATURE_CONVENTION renamed, 26.11-12 citation) and moves the Mercury-by-association rule into mercury_malefic_companions() shared with the strict Gajakesari yoga, values unchanged; scoring identity stays rectification-v5-matrix-scoring-15 (ERR-110) changed frozen identity files; the sealed replay output is unchanged, and ERR-110 stored a new record. Prior astrologer_166b_2026_11_04 report and freeze preserved byte-for-byte." } }, "previous_fixed_protocol_rerun": { @@ -342,13 +342,13 @@ "must_not_claim_as_release_metrics": true, "replay_started_at_utc": "2026-11-05T05:54:10.220155+00:00", "replay_finished_at_utc": "2026-11-05T05:56:39.253808+00:00", - "boundary": "Previously exposed reported-offset sensitivity after the #166 astrologer reply 2026-11-05 (BUG-1613..1617): scripts/shadbala.py adds labels only (saptavargaja_moolatrikona d1_degree_band_only_provisional, drik_mercury_jupiter quarter_only_raman_pending, MERCURY_NATURE_CONVENTION renamed, 26.11-12 citation) and moves the Mercury-by-association rule into mercury_malefic_companions() shared with the strict Gajakesari yoga, values unchanged; scoring identity stays rectification-v5-matrix-scoring-15 (ERR-110). Versus the 2026-11-04 astrologer_166b_2026_11_04 report, all 900 trials and the summary are identical (top-1 hits at +-15/30/60 4/3/0 of 300; top-3 31/34/30; delivery coverage at +-15 219). Versus the 2026-09-20 historical report, changed_trial_count is 731 of 900. Not a release metric. No tuning. Prior reports and freezes preserved byte-for-byte.", + "boundary": "Previously exposed reported-offset sensitivity after the #166 astrologer reply 2026-11-05 (BUG-1622..1626): scripts/shadbala.py adds labels only (saptavargaja_moolatrikona d1_degree_band_only_provisional, drik_mercury_jupiter quarter_only_raman_pending, MERCURY_NATURE_CONVENTION renamed, 26.11-12 citation) and moves the Mercury-by-association rule into mercury_malefic_companions() shared with the strict Gajakesari yoga, values unchanged; scoring identity stays rectification-v5-matrix-scoring-15 (ERR-110). Versus the 2026-11-04 astrologer_166b_2026_11_04 report, all 900 trials and the summary are identical (top-1 hits at +-15/30/60 4/3/0 of 300; top-3 31/34/30; delivery coverage at +-15 219). Versus the 2026-09-20 historical report, changed_trial_count is 731 of 900. Not a release metric. No tuning. Prior reports and freezes preserved byte-for-byte.", "previous_reported_offset_replay": { "report_path": "docs/research/reported_offset_astrologer_166b_2026_11_04.json", "report_sha256": "8f195b6ffa258cf55372792c1f4536c77cab96f6abcf31f1e50c54dafd9185ca", "freeze_record_path": "docs/research/reported_offset_astrologer_166b_2026_11_04.freeze.json", "freeze_record_sha256": "c9bc41bb89ee5077d1c7952124bf1a9a9f15ec0cd19449cb2efe003559fc59fb", - "superseded_reason": "#166 astrologer reply 2026-11-05 (BUG-1613..1617): scripts/shadbala.py adds labels only (saptavargaja_moolatrikona d1_degree_band_only_provisional, drik_mercury_jupiter quarter_only_raman_pending, MERCURY_NATURE_CONVENTION renamed, 26.11-12 citation) and moves the Mercury-by-association rule into mercury_malefic_companions() shared with the strict Gajakesari yoga, values unchanged; scoring identity stays rectification-v5-matrix-scoring-15 (ERR-110) changed frozen identity files; the reported-offset replay output is unchanged, and ERR-110 stored a new record. Prior astrologer_166b_2026_11_04 report and freeze preserved byte-for-byte." + "superseded_reason": "#166 astrologer reply 2026-11-05 (BUG-1622..1626): scripts/shadbala.py adds labels only (saptavargaja_moolatrikona d1_degree_band_only_provisional, drik_mercury_jupiter quarter_only_raman_pending, MERCURY_NATURE_CONVENTION renamed, 26.11-12 citation) and moves the Mercury-by-association rule into mercury_malefic_companions() shared with the strict Gajakesari yoga, values unchanged; scoring identity stays rectification-v5-matrix-scoring-15 (ERR-110) changed frozen identity files; the reported-offset replay output is unchanged, and ERR-110 stored a new record. Prior astrologer_166b_2026_11_04 report and freeze preserved byte-for-byte." } }, "previous_reported_offset_replay": { diff --git a/references/yoga_rules.json b/references/yoga_rules.json index abf9cb80..3df09204 100644 --- a/references/yoga_rules.json +++ b/references/yoga_rules.json @@ -946,7 +946,7 @@ }, { "type": "custom", - "expr": "\n# #166 Q5 (BUG-1596): PVR strict definition. Jupiter in a kendra from the Moon (previous condition),\n# conjoined or aspected by a benefic, not debilitated, not combust, not in a natural enemy's sign.\n# The strong-Jupiter substitute is a separate rule (gajakesari_strong_jupiter_variant).\n# #166 2026-11-05 (BUG-1615): a natural-malefic Mercury (same sign as any natural malefic incl. the\n# waning Moon, nodes excluded: shadbala.MERCURY_NATURE_CONVENTION via mercury_malefic_companions())\n# is not a benefic supporter. Enemy sign = natural enmity (pending); the Moon is not a supporter (pending).\ntriggers = []\nmercury_with = mercury_malefic_companions()\nlinked = [b for b in ['Venus', 'Mercury'] if b in ctx.planets and (same_house(b, 'Jupiter') or aspect(b, 'Jupiter'))]\nbenefic_links = [b for b in linked if not (b == 'Mercury' and mercury_with)]\nok = bool(benefic_links) and not debil('Jupiter') and not combust('Jupiter') and not in_natural_enemy_sign('Jupiter')\nif ok:\n triggers = [('吉星关联：' + '、'.join(PLANET_ZH.get(p, p) for p in benefic_links), 'benefic association: ' + ', '.join(benefic_links))]\n if 'Mercury' in linked and mercury_with:\n triggers.append(('水星与' + '、'.join(PLANET_ZH.get(p, p) for p in mercury_with) + '同宫，按凶星计，不算吉星支持', 'Mercury is with ' + ', '.join(mercury_with) + ' (natural malefic), not counted'))\n__result__ = ok\n", + "expr": "\n# #166 Q5 (BUG-1596): PVR strict definition. Jupiter in a kendra from the Moon (previous condition),\n# conjoined or aspected by a benefic, not debilitated, not combust, not in a natural enemy's sign.\n# The strong-Jupiter substitute is a separate rule (gajakesari_strong_jupiter_variant).\n# #166 2026-11-05 (BUG-1624): a natural-malefic Mercury (same sign as any natural malefic incl. the\n# waning Moon, nodes excluded: shadbala.MERCURY_NATURE_CONVENTION via mercury_malefic_companions())\n# is not a benefic supporter. Enemy sign = natural enmity (pending); the Moon is not a supporter (pending).\ntriggers = []\nmercury_with = mercury_malefic_companions()\nlinked = [b for b in ['Venus', 'Mercury'] if b in ctx.planets and (same_house(b, 'Jupiter') or aspect(b, 'Jupiter'))]\nbenefic_links = [b for b in linked if not (b == 'Mercury' and mercury_with)]\nok = bool(benefic_links) and not debil('Jupiter') and not combust('Jupiter') and not in_natural_enemy_sign('Jupiter')\nif ok:\n triggers = [('吉星关联：' + '、'.join(PLANET_ZH.get(p, p) for p in benefic_links), 'benefic association: ' + ', '.join(benefic_links))]\n if 'Mercury' in linked and mercury_with:\n triggers.append(('水星与' + '、'.join(PLANET_ZH.get(p, p) for p in mercury_with) + '同宫，按凶星计，不算吉星支持', 'Mercury is with ' + ', '.join(mercury_with) + ' (natural malefic), not counted'))\n__result__ = ok\n", "combo_template": "严格定义：木星受吉星合相或相照，且不落陷、不燃烧、不落敌宫", "combo_template_en": "Strict definition: Jupiter conjoined or aspected by a benefic, not debilitated, combust or in an enemy sign" } @@ -965,7 +965,7 @@ "strength": "中", "strength_en": "moderate", "enabled": true, - "source_reference": "#166 Q5 (astrologer 2026-11-04): strict definition only; the strong-Jupiter variant and the bare Moon-Jupiter kendra relation are separate rules. #166 2026-11-05 (BUG-1615): PVR §11.6 (print pp. 124-125, PDF pp. 136-137): Moon kendra AND benefic conjunction-or-aspect AND Jupiter not debilitated / combust / in an enemy sign. A natural-malefic Mercury (shadbala.MERCURY_NATURE_CONVENTION) is not a supporter. Pending astrologer review: natural vs five-fold enmity for the enemy-sign test (natural kept); whether a benefic Moon may be a supporter (excluded, not approved).", + "source_reference": "#166 Q5 (astrologer 2026-11-04): strict definition only; the strong-Jupiter variant and the bare Moon-Jupiter kendra relation are separate rules. #166 2026-11-05 (BUG-1624): PVR §11.6 (print pp. 124-125, PDF pp. 136-137): Moon kendra AND benefic conjunction-or-aspect AND Jupiter not debilitated / combust / in an enemy sign. A natural-malefic Mercury (shadbala.MERCURY_NATURE_CONVENTION) is not a supporter. Pending astrologer review: natural vs five-fold enmity for the enemy-sign test (natural kept); whether a benefic Moon may be a supporter (excluded, not approved).", "bvr_chapter": "Chapter 1" }, { @@ -990,7 +990,7 @@ }, { "type": "custom", - "expr": "\n# #166 Q5: variant only when the strict definition does not hold.\n# #166 2026-11-05 (BUG-1615): the strict check below uses the same malefic-Mercury rule as the strict\n# rule, so the two stay mutually exclusive; the variant itself uses no benefic supporter.\nmercury_with = mercury_malefic_companions()\nbenefic_links = [b for b in ['Venus', 'Mercury'] if b in ctx.planets and not (b == 'Mercury' and mercury_with) and (same_house(b, 'Jupiter') or aspect(b, 'Jupiter'))]\nstrict = bool(benefic_links) and not debil('Jupiter') and not combust('Jupiter') and not in_natural_enemy_sign('Jupiter')\nstrong_jup = exalted('Jupiter') or own('Jupiter') or moola('Jupiter')\ntriggers = [(f\"木星强：{dignity('Jupiter')}\", f\"strong Jupiter: {dignity('Jupiter')}\")] if strong_jup and not strict else []\n__result__ = strong_jup and not strict and not debil('Jupiter')\n", + "expr": "\n# #166 Q5: variant only when the strict definition does not hold.\n# #166 2026-11-05 (BUG-1624): the strict check below uses the same malefic-Mercury rule as the strict\n# rule, so the two stay mutually exclusive; the variant itself uses no benefic supporter.\nmercury_with = mercury_malefic_companions()\nbenefic_links = [b for b in ['Venus', 'Mercury'] if b in ctx.planets and not (b == 'Mercury' and mercury_with) and (same_house(b, 'Jupiter') or aspect(b, 'Jupiter'))]\nstrict = bool(benefic_links) and not debil('Jupiter') and not combust('Jupiter') and not in_natural_enemy_sign('Jupiter')\nstrong_jup = exalted('Jupiter') or own('Jupiter') or moola('Jupiter')\ntriggers = [(f\"木星强：{dignity('Jupiter')}\", f\"strong Jupiter: {dignity('Jupiter')}\")] if strong_jup and not strict else []\n__result__ = strong_jup and not strict and not debil('Jupiter')\n", "combo_template": "变体：木星强（入旺 / 本宫 / 根本三角）代替吉星关联", "combo_template_en": "Variant: a strong Jupiter (exalted / own / moolatrikona) in place of benefic association" } @@ -1009,7 +1009,7 @@ "strength": "弱", "strength_en": "weak", "enabled": true, - "source_reference": "#166 Q5: a strong Jupiter (exalted / own / moolatrikona) substitutes for the benefic association; not the strict yoga. #166 2026-11-05 (BUG-1615): shown only when the strict definition (with the malefic-Mercury rule) does not hold.", + "source_reference": "#166 Q5: a strong Jupiter (exalted / own / moolatrikona) substitutes for the benefic association; not the strict yoga. #166 2026-11-05 (BUG-1624): shown only when the strict definition (with the malefic-Mercury rule) does not hold.", "bvr_chapter": "Chapter 1" }, { @@ -11578,7 +11578,7 @@ "category": "durbhaga", "logic": { "type": "custom", - "expr": "\n# #166 Q6 (BUG-1597): Raman 300 Combinations no. 198 only (no. 199 is bvr_matrunasa_199).\n# Hemmed by, associated with, or aspected by natural malefics (OR). Aspects are named Parashari graha\n# aspects (7th; Mars 4/8, Jupiter 5/9, Saturn 3/10); Rahu / Ketu count for conjunction and hemming,\n# not for aspect (implementation choice recorded in BUG-1597). No Jaimini rasi drishti.\n# #166 2026-11-05 (BUG-1616): nodes casting no aspect is a conservative engineering choice, pending\n# astrologer review (no source found granting or denying them the 7th). Malefic set used for hemming and\n# conjunction: MALEFICS = Sun, Mars, Saturn, Rahu, Ketu; aspects from Sun, Mars, Saturn only. Mercury is\n# not treated as a malefic here. Raman edition / page pending (pp. 188-190 in earlier notes is a locator only).\ntriggers = []\nif 'Moon' in ctx.planets:\n moon_h = house_of('Moon')\n prev_h = ((moon_h - 2) % 12) + 1\n next_h = (moon_h % 12) + 1\n prev_m = [p for p in MALEFICS if p in ctx.planets and house_of(p) == prev_h]\n next_m = [p for p in MALEFICS if p in ctx.planets and house_of(p) == next_h]\n if prev_m and next_m:\n triggers.append(('凶星夹制：' + '、'.join(PLANET_ZH.get(p, p) for p in prev_m + next_m), 'hemmed: ' + ', '.join(prev_m + next_m)))\n conj = [p for p in MALEFICS if p in ctx.planets and same_house('Moon', p)]\n if conj:\n triggers.append(('与月亮同宫：' + '、'.join(PLANET_ZH.get(p, p) for p in conj), 'with Moon: ' + ', '.join(conj)))\n asp = [p for p in ('Sun', 'Mars', 'Saturn') if p in ctx.planets and not same_house('Moon', p) and aspect(p, 'Moon')]\n if asp:\n triggers.append(('相照月亮：' + '、'.join(PLANET_ZH.get(p, p) for p in asp), 'aspects Moon: ' + ', '.join(asp)))\n__result__ = bool(triggers)\n", + "expr": "\n# #166 Q6 (BUG-1597): Raman 300 Combinations no. 198 only (no. 199 is bvr_matrunasa_199).\n# Hemmed by, associated with, or aspected by natural malefics (OR). Aspects are named Parashari graha\n# aspects (7th; Mars 4/8, Jupiter 5/9, Saturn 3/10); Rahu / Ketu count for conjunction and hemming,\n# not for aspect (implementation choice recorded in BUG-1597). No Jaimini rasi drishti.\n# #166 2026-11-05 (BUG-1625): nodes casting no aspect is a conservative engineering choice, pending\n# astrologer review (no source found granting or denying them the 7th). Malefic set used for hemming and\n# conjunction: MALEFICS = Sun, Mars, Saturn, Rahu, Ketu; aspects from Sun, Mars, Saturn only. Mercury is\n# not treated as a malefic here. Raman edition / page pending (pp. 188-190 in earlier notes is a locator only).\ntriggers = []\nif 'Moon' in ctx.planets:\n moon_h = house_of('Moon')\n prev_h = ((moon_h - 2) % 12) + 1\n next_h = (moon_h % 12) + 1\n prev_m = [p for p in MALEFICS if p in ctx.planets and house_of(p) == prev_h]\n next_m = [p for p in MALEFICS if p in ctx.planets and house_of(p) == next_h]\n if prev_m and next_m:\n triggers.append(('凶星夹制：' + '、'.join(PLANET_ZH.get(p, p) for p in prev_m + next_m), 'hemmed: ' + ', '.join(prev_m + next_m)))\n conj = [p for p in MALEFICS if p in ctx.planets and same_house('Moon', p)]\n if conj:\n triggers.append(('与月亮同宫：' + '、'.join(PLANET_ZH.get(p, p) for p in conj), 'with Moon: ' + ', '.join(conj)))\n asp = [p for p in ('Sun', 'Mars', 'Saturn') if p in ctx.planets and not same_house('Moon', p) and aspect(p, 'Moon')]\n if asp:\n triggers.append(('相照月亮：' + '、'.join(PLANET_ZH.get(p, p) for p in asp), 'aspects Moon: ' + ', '.join(asp)))\n__result__ = bool(triggers)\n", "combo_template": "Raman 198：月亮受凶星夹制、合相或相照", "combo_template_en": "Raman 198: Moon hemmed, conjoined or aspected by malefics" }, diff --git a/scripts/arudha_reference/pvr_dual_lord.py b/scripts/arudha_reference/pvr_dual_lord.py index 8a78fc9a..2ffff33b 100644 --- a/scripts/arudha_reference/pvr_dual_lord.py +++ b/scripts/arudha_reference/pvr_dual_lord.py @@ -1,7 +1,7 @@ #!/usr/bin/env python3 """PVR Arudha dual-lord selection for Scorpio (Mars / Ketu) and Aquarius (Saturn / Rahu). -Opt-in reference only (#166 Q28, astrologer reply 2026-11-05; BUG-1617). +Opt-in reference only (#166 Q28, astrologer reply 2026-11-05; BUG-1626). Label: "PVR 双宫主比较（参考，未进默认）". The default Arudha path (`jaimini.calc_arudha_padas`, fixed lords Mars / Saturn), the chart page, the consultation card and rectification scoring do NOT use this module. Moving it diff --git a/scripts/consultation_native_layers.py b/scripts/consultation_native_layers.py index 46d13cf9..08b5d552 100644 --- a/scripts/consultation_native_layers.py +++ b/scripts/consultation_native_layers.py @@ -165,7 +165,7 @@ def build_d9_summary(engine, chart: dict, modules: dict) -> dict: if not isinstance(pdata, dict) or not d9_row.get("sign") or not pdata.get("sign"): continue # A D9 Moolatrikona is read from the in-varga degree inside the D1 band - # (engine.VARGA_MOOLATRIKONA_BASIS): 待核 with the astrologer (#166, BUG-1613). + # (engine.VARGA_MOOLATRIKONA_BASIS): 待核 with the astrologer (#166, BUG-1622). d9_level = engine._get_dignity_level(name, d9_row["sign"], d9_row.get("degree"), d9_context) # D1: the chart's own status label (what the base section shows); its # level is read back from the same label table. @@ -196,7 +196,7 @@ def build_d9_summary(engine, chart: dict, modules: dict) -> dict: "vargottama": [name for name, row in rows.items() if row["vargottama"]], "reversals": reversals, "reversal_rule": REVERSAL_RULE, - # BUG-1613: how a D9 Moolatrikona is decided; pending astrologer review. + # BUG-1622: how a D9 Moolatrikona is decided; pending astrologer review. "moolatrikona_basis": engine.VARGA_MOOLATRIKONA_BASIS, } diff --git a/scripts/jyotish_engine.py b/scripts/jyotish_engine.py index 6645c71f..17e1dc8d 100644 --- a/scripts/jyotish_engine.py +++ b/scripts/jyotish_engine.py @@ -309,7 +309,7 @@ DEBILITATION = {'Sun': 'Libra', 'Moon': 'Scorpio', 'Mars': 'Cancer', 'Mercury': # 新值: Moon 金牛 [3, 30)、Mercury 处女 [15, 20)，与 shadbala._MOOLATRIKONA 一致 # 原因: BUG-1575 / BUG-1583，BPHS 3.51–54；#166 第 18 条（占星师 2026-11-04） MOOLATRIKONA = {'Sun': ('Leo', 0, 20), 'Moon': ('Taurus', 3, 30), 'Mars': ('Aries', 0, 12), 'Mercury': ('Virgo', 15, 20), 'Jupiter': ('Sagittarius', 0, 10), 'Venus': ('Libra', 0, 15), 'Saturn': ('Aquarius', 0, 20)} -# #166 2026-11-05 (BUG-1613): in a varga (the D9 expanded table, the D9 dignity in +# #166 2026-11-05 (BUG-1622): in a varga (the D9 expanded table, the D9 dignity in # _get_dignity_level, the consultation D9 summary) a planet is marked Moolatrikona when # its in-varga degree falls inside the D1 band above. Whether and how Moolatrikona # applies in the vargas (varga sign, or in-varga degree; the D30 mapping) is 待核 with @@ -14572,7 +14572,7 @@ def cmd_jaimini(args): if mode in ('all', 'arudha'): result['arudha_padas'] = calc_arudha_padas(asc_idx, planet_lons) result['graha_padas'] = calc_graha_padas(planet_lons) - # #166 Q28 (BUG-1617): PVR §15.5.1 dual-lord comparison for the houses on + # #166 Q28 (BUG-1626): PVR §15.5.1 dual-lord comparison for the houses on # Scorpio / Aquarius, next to the default fixed-lord padas. Reference only # (PVR 双宫主比较（参考，未进默认）); arudha_padas above stays the default. try: @@ -16493,7 +16493,7 @@ def cmd_full_reading(args): 'sign': d9_sign, 'house': ((d9_sign_idx - d9_asc_idx) % 12) + 1, 'degree_in_sign': d9_deg, - # D9 Moolatrikona by in-varga degree: VARGA_MOOLATRIKONA_BASIS (待核, BUG-1613). + # D9 Moolatrikona by in-varga degree: VARGA_MOOLATRIKONA_BASIS (待核, BUG-1622). 'dignity': _get_dignity_level(d9_pn, d9_sign, d9_deg, d9_context), } jaimini_result['darakaraka'] = analyze_darakaraka({ @@ -16993,7 +16993,7 @@ def cmd_full_reading(args): is_exalted = (EXALTATION.get(pn) == d9_sign) is_debilitated = (DEBILITATION.get(pn) == d9_sign) is_moola = False - # In-varga degree inside the D1 band: VARGA_MOOLATRIKONA_BASIS (待核, BUG-1613). + # In-varga degree inside the D1 band: VARGA_MOOLATRIKONA_BASIS (待核, BUG-1622). if pn in MOOLATRIKONA: mt_sign, mt_start, mt_end = MOOLATRIKONA[pn] if mt_sign == d9_sign and mt_start <= d9_deg < mt_end: diff --git a/scripts/research/reported_offset_sweep.py b/scripts/research/reported_offset_sweep.py index 5996ee65..cd432db1 100644 --- a/scripts/research/reported_offset_sweep.py +++ b/scripts/research/reported_offset_sweep.py @@ -56,7 +56,7 @@ MINUTE_STEP = 1 # 原因: #166 第二批（BUG-1590～1598）改了 functional_benefics.py、jaimini.py、narayana_dasha.py、shadbala.py、varga.py，校正身份升到 scoring-15；ERR-110 另存新记录 # 原值: docs/research/reported_offset_astrologer_166b_2026_11_04.{freeze.json,json} # 新值: docs/research/reported_offset_astrologer_166c_2026_11_05.{freeze.json,json} -# 原因: #166 占星师 2026-11-05 答复（BUG-1613～1617）只改了 shadbala.py 的标签、注释与共用的水星规则函数（数值不变），校正身份仍是 scoring-15；ERR-110 另存新记录 +# 原因: #166 占星师 2026-11-05 答复（BUG-1622～1626）只改了 shadbala.py 的标签、注释与共用的水星规则函数（数值不变），校正身份仍是 scoring-15；ERR-110 另存新记录 FREEZE = ROOT / "docs/research/reported_offset_astrologer_166c_2026_11_05.freeze.json" REPORT = ROOT / "docs/research/reported_offset_astrologer_166c_2026_11_05.json" LEGACY_REPORT = ROOT / "docs/research/reported_offset_2026_09_20.json" diff --git a/scripts/research/sealed_holdout_rerun.py b/scripts/research/sealed_holdout_rerun.py index 825be5fb..2f898067 100644 --- a/scripts/research/sealed_holdout_rerun.py +++ b/scripts/research/sealed_holdout_rerun.py @@ -53,7 +53,7 @@ DATASET = ROOT / "references/real_case_calibration/minute_rectification_holdout_ # 原因: #166 第二批（BUG-1590～1598）改了 functional_benefics.py、jaimini.py、narayana_dasha.py、shadbala.py、varga.py，校正身份升到 scoring-15；ERR-110 另存新记录 # 原值: docs/research/sealed_holdout_rerun_astrologer_166b_2026_11_04.{freeze.json,json} # 新值: docs/research/sealed_holdout_rerun_astrologer_166c_2026_11_05.{freeze.json,json} -# 原因: #166 占星师 2026-11-05 答复（BUG-1613～1617）只改了 shadbala.py 的标签、注释与共用的水星规则函数（数值不变），校正身份仍是 scoring-15；ERR-110 另存新记录 +# 原因: #166 占星师 2026-11-05 答复（BUG-1622～1626）只改了 shadbala.py 的标签、注释与共用的水星规则函数（数值不变），校正身份仍是 scoring-15；ERR-110 另存新记录 FREEZE = ROOT / "docs/research/sealed_holdout_rerun_astrologer_166c_2026_11_05.freeze.json" REPORT = ROOT / "docs/research/sealed_holdout_rerun_astrologer_166c_2026_11_05.json" LEGACY_REPORT = ROOT / "docs/research/sealed_holdout_rerun_2026_09_20.json" diff --git a/scripts/shadbala.py b/scripts/shadbala.py index 05eb96db..2f508fd3 100644 --- a/scripts/shadbala.py +++ b/scripts/shadbala.py @@ -128,7 +128,7 @@ SPECIAL_ASPECTS = { # BUG-1584, Q16: BPHS (Santhanam p. 265) Saptavargaja 45/30/20/15/10/4/2; Raman's # 22.5/15/7.5/3.75/1.875 is a separate convention. Compound relations from D1 (Q20). SAPTAVARGAJA_CONVENTION = "bphs_27_saptavargaja_20_15_10_4_2_d1_compound" -# BUG-1613 (#166 Q17 / Q19, astrologer 2026-11-05): how the 45-virupa +# BUG-1622 (#166 Q17 / Q19, astrologer 2026-11-05): how the 45-virupa # Moolatrikona slot of Saptavargaja is applied. Santhanam BPHS 27.2-4 (print # p. 265) confirms the value 45 and says the same values apply to the other six # vargas, so "45 only in D1" is NOT a rule that text states. Still 待核 (open): @@ -141,13 +141,13 @@ SAPTAVARGAJA_CONVENTION = "bphs_27_saptavargaja_20_15_10_4_2_d1_compound" SAPTAVARGAJA_MOOLATRIKONA_CONVENTION = "d1_degree_band_only_provisional" # BUG-1586, Q24: BPHS 26.9-12 piecewise special aspects replace the general # segments for Saturn, Mars and Jupiter (not Raman's add-on bonus). -# BUG-1614 (#166 Q24, astrologer 2026-11-05): 26.11-12 (Mars [180, 210) = 60; +# BUG-1623 (#166 Q24, astrologer 2026-11-05): 26.11-12 (Mars [180, 210) = 60; # Jupiter [120, 150) and [240, 270) = 60 - offset) match Santhanam BPHS vol. 1, # print p. 256. Each branch returns the final single-aspect value (no general # value on top). Left-closed / right-open segments are an engineering # convention for the endpoints, not a rule the book spells out. DRIK_SPECIAL_ASPECT_CONVENTION = "bphs_26_9_12_piecewise" -# BUG-1614 (#166 Q23, astrologer 2026-11-05): aspects CAST BY Mercury and +# BUG-1623 (#166 Q23, astrologer 2026-11-05): aspects CAST BY Mercury and # Jupiter in Drik Bala. Three readings, none adopted as BPHS: # 1. full value replaces the quarter: 1x; # 2. Santhanam's literal English of 27.19 (print p. 284): the usual +/- quarter @@ -162,11 +162,11 @@ DRIK_MERCURY_JUPITER_CONVENTION = "quarter_only_raman_pending" # natural malefic (Sun, Mars, Saturn, waning Moon); Rahu / Ketu are not counted. # 原值: "bphs_3_11_any_malefic_same_sign_no_nodes" # 新值: 下面这个名字（行为不变） -# 原因: BUG-1614（#166 第 25 条，占星师 2026-11-05）：Santhanam 3.10-11（印刷页 27） +# 原因: BUG-1623（#166 第 25 条，占星师 2026-11-05）：Santhanam 3.10-11（印刷页 27） # 的自然凶星包含罗睺、计都，排除交点是本仓的实现选择，不是 BPHS 规则；符合条件的亏月 # 在集合内。印刷页 28 的译者注（亏月与水星结合时两者皆吉）是另一层，没有采用。 # 同星座、无容许度是具名的离散实现。mercury_malefic_companions() 是这条规则唯一的 -# 实现，Drik 分类与严格 Gajakesari（yoga_engine）都调用它（BUG-1615）。 +# 实现，Drik 分类与严格 Gajakesari（yoga_engine）都调用它（BUG-1624）。 MERCURY_NATURE_CONVENTION = "same_sign_any_natural_malefic_waning_moon_included_nodes_excluded_implementation_choice" # BUG-1587, Q21: Kendradi Bala by whole-sign house from the Lagna sign. KENDRADI_CONVENTION = "whole_sign_from_lagna" @@ -282,10 +282,10 @@ def mercury_malefic_companions(planets: Dict) -> list[str]: """Natural malefics sharing Mercury's sign (MERCURY_NATURE_CONVENTION). Not empty = Mercury is a natural malefic by association. Rahu / Ketu are - not counted (implementation choice, BUG-1614); the waning Moon is. + not counted (implementation choice, BUG-1623); the waning Moon is. ``planets`` rows carry ``sign`` and ``degree`` (in-sign or absolute), as for calc_shadbala and the yoga engine. Shared by Drik Bala and the strict - Gajakesari supporter check (BUG-1615), so the rule lives in one place. + Gajakesari supporter check (BUG-1624), so the rule lives in one place. """ if "Mercury" not in planets: return [] @@ -367,7 +367,7 @@ def _sphuta_drishti_virupas(angle: float, aspecting_planet: str, variant: str = Sources: jyotishvidya.com ch. 26 English; 26.11-12 (Mars, Jupiter) also checked against Santhanam BPHS vol. 1 print p. 256 (astrologer 2026-11-05, - BUG-1614). Segments are [start, end): 150, 210 and 270 belong to the next + BUG-1623). Segments are [start, end): 150, 210 and 270 belong to the next branch (engineering convention for endpoints). ``round(angle, 2)`` below snaps offsets under 0.005 deg onto the endpoint. """ @@ -425,7 +425,7 @@ def calc_drik_bala_precise(pname: str, all_planets: Dict, variant: str = "bphs") Aspects cast by Mercury and Jupiter take the same quarter as every other aspect (DRIK_MERCURY_JUPITER_CONVENTION, Raman reading; BPHS 27.19 1x / - 1.25x readings are 待核, BUG-1614). + 1.25x readings are 待核, BUG-1623). """ if pname not in all_planets: return 0.0 @@ -508,7 +508,7 @@ def calc_sthana_bala_precise(pname: str, all_planets: Dict, house: int, variant: mt_sign, mt_start, mt_end = _MOOLATRIKONA[pname] if variant == "raman_gbb": mt_start = _RAMAN_MOOLATRIKONA_START.get(pname, mt_start) - # SAPTAVARGAJA_MOOLATRIKONA_CONVENTION (provisional, 待核, BUG-1613): 45 only + # SAPTAVARGAJA_MOOLATRIKONA_CONVENTION (provisional, 待核, BUG-1622): 45 only # in D1 and only inside the degree band; BPHS 27.2-4 does not state that # restriction, the other six vargas get own sign / compound relation. if division == 1 and sign_idx == mt_sign and mt_start <= degree < mt_end: @@ -877,7 +877,7 @@ def calc_shadbala(planets: Dict, asc_sign: str, birth_hour: float, 'moon_paksha': MOON_PAKSHA_CONVENTION if context else 'legacy_approximation', 'sun_ayana': SUN_AYANA_CONVENTION if context else 'legacy_approximation', 'saptavargaja': SAPTAVARGAJA_CONVENTION, - # #166 2026-11-05 (BUG-1613 / BUG-1614): provisional / pending labels, values unchanged. + # #166 2026-11-05 (BUG-1622 / BUG-1623): provisional / pending labels, values unchanged. 'saptavargaja_moolatrikona': SAPTAVARGAJA_MOOLATRIKONA_CONVENTION, 'drik_special_aspects': DRIK_SPECIAL_ASPECT_CONVENTION, 'drik_mercury_jupiter': DRIK_MERCURY_JUPITER_CONVENTION, diff --git a/scripts/yoga_engine.py b/scripts/yoga_engine.py index ac69f3b3..0600a698 100644 --- a/scripts/yoga_engine.py +++ b/scripts/yoga_engine.py @@ -749,7 +749,7 @@ def _compiled_exec_code(expr): def _shadbala_mercury_rule(): """shadbala.mercury_malefic_companions, imported on first use. - #166 (BUG-1615): the strict Gajakesari supporter check reuses the one + #166 (BUG-1624): the strict Gajakesari supporter check reuses the one implementation of the Mercury-by-association rule (BPHS 3.10-11 as labelled by shadbala.MERCURY_NATURE_CONVENTION) instead of re-deciding it. Imported lazily because shadbala pulls in swisseph; no import cycle @@ -1735,7 +1735,7 @@ class YogaEngine: return sign in SIGN_LORDS and SIGN_LORDS[sign] in NATURAL_ENEMIES.get(p, []) def mercury_malefic_companions(): - """BUG-1615: natural malefics sharing Mercury's sign; non-empty = Mercury is a natural malefic. + """BUG-1624: natural malefics sharing Mercury's sign; non-empty = Mercury is a natural malefic. Same rule and label as Drik Bala (shadbala.MERCURY_NATURE_CONVENTION): Sun, Mars, Saturn and the waning Moon count, Rahu / Ketu do not. diff --git a/tests/test_arudha_lord_selection_pvr.py b/tests/test_arudha_lord_selection_pvr.py index 52ff1466..1928931a 100644 --- a/tests/test_arudha_lord_selection_pvr.py +++ b/tests/test_arudha_lord_selection_pvr.py @@ -1,4 +1,4 @@ -"""#166 Q28 (BUG-1617): PVR §15.5.1 Arudha dual-lord selection, opt-in reference only.""" +"""#166 Q28 (BUG-1626): PVR §15.5.1 Arudha dual-lord selection, opt-in reference only.""" from __future__ import annotations diff --git a/tests/test_astrologer_rulings_166c.py b/tests/test_astrologer_rulings_166c.py index 2fa99196..71c1333b 100644 --- a/tests/test_astrologer_rulings_166c.py +++ b/tests/test_astrologer_rulings_166c.py @@ -1,8 +1,8 @@ """#166 follow-up: astrologer reply of 2026-11-05 (Gitea #166 comment 39249). Labels and boundary tests only for Shadbala (no value changes); the strict -Gajakesari malefic-Mercury rule (BUG-1615), the Matrunasa labels (BUG-1616) -and the opt-in PVR Arudha dual-lord selector (BUG-1617) have their own tests +Gajakesari malefic-Mercury rule (BUG-1624), the Matrunasa labels (BUG-1625) +and the opt-in PVR Arudha dual-lord selector (BUG-1626) have their own tests further down / in tests/test_arudha_lord_selection_pvr.py. """ @@ -26,7 +26,7 @@ def _planets(**lons: float) -> dict: return {name: {"degree": lon} for name, lon in lons.items()} -# ── BUG-1613: Saptavargaja Moolatrikona is a provisional D1-degree-band rule ── +# ── BUG-1622: Saptavargaja Moolatrikona is a provisional D1-degree-band rule ── def test_saptavargaja_moolatrikona_label_is_provisional_and_reported() -> None: assert shadbala.SAPTAVARGAJA_MOOLATRIKONA_CONVENTION == "d1_degree_band_only_provisional" @@ -51,7 +51,7 @@ def test_saptavargaja_moolatrikona_behaviour_unchanged_d1_band_only() -> None: assert all(row[f"sapta_d{d}"] != 45.0 for d in (2, 3, 7, 9, 12, 30)) -# ── BUG-1614: Drik Mercury / Jupiter quarter label; Mercury nature relabel ──── +# ── BUG-1623: Drik Mercury / Jupiter quarter label; Mercury nature relabel ──── def test_drik_mercury_jupiter_label_is_pending_and_still_a_quarter() -> None: assert shadbala.DRIK_MERCURY_JUPITER_CONVENTION == "quarter_only_raman_pending" @@ -136,7 +136,7 @@ def test_special_aspect_offsets_below_half_hundredth_snap_to_endpoint() -> None: assert shadbala._sphuta_drishti_virupas(149.994, "Jupiter") == 30.01 -# ── BUG-1615: strict Gajakesari, a natural-malefic Mercury is not a supporter ── +# ── BUG-1624: strict Gajakesari, a natural-malefic Mercury is not a supporter ── from yoga_engine import YogaEngine # noqa: E402 @@ -196,7 +196,7 @@ def test_gajakesari_venus_supports_and_the_excluded_mercury_is_named() -> None: def test_strong_jupiter_variant_uses_the_same_mercury_rule() -> None: # Moon Aries, Jupiter exalted in Cancer with a malefic Mercury (Sun in Cancer) as its only link: - # the strict yoga fails, so the variant shows (before BUG-1615 the strict yoga held instead). + # the strict yoga fails, so the variant shows (before BUG-1624 the strict yoga held instead). positions = {"Moon": ("Aries", 10), "Jupiter": ("Cancer", 5), "Sun": ("Cancer", 28), "Mercury": ("Cancer", 20), "Venus": ("Gemini", 10), "Mars": ("Capricorn", 5), "Saturn": ("Sagittarius", 5)} rows = _yogas(positions) @@ -204,7 +204,7 @@ def test_strong_jupiter_variant_uses_the_same_mercury_rule() -> None: assert "Strong-Jupiter Gajakesari Variant" in rows -# ── BUG-1616: Matrunasa 198 labels, 199 effects describe 199 ────────────────── +# ── BUG-1625: Matrunasa 198 labels, 199 effects describe 199 ────────────────── import json # noqa: E402 @@ -243,7 +243,7 @@ def test_matrunasa_199_effects_describe_the_navamsa_chain_not_198() -> None: assert hit["triggers_zh"] == ["月亮 → 水星 → 金星，落第 6 宫"] -# ── BUG-1613: D9 Moolatrikona by in-varga degree is labelled pending ───────── +# ── BUG-1622: D9 Moolatrikona by in-varga degree is labelled pending ───────── def test_varga_moolatrikona_basis_label() -> None: import jyotish_engine diff --git a/tests/test_sealed_holdout_contract_freshness.py b/tests/test_sealed_holdout_contract_freshness.py index f8cecb5d..16368028 100644 --- a/tests/test_sealed_holdout_contract_freshness.py +++ b/tests/test_sealed_holdout_contract_freshness.py @@ -99,7 +99,7 @@ IDENTITY_FORBIDDEN_PREFIXES = ( "docs/research/sealed_holdout_rerun_astrologer_166b_2026_11_04", "docs/research/reported_offset_astrologer_166b_2026_11_04", # 原值: 排除到 astrologer_166b_2026_11_04；新值: 再排除 astrologer_166c_2026_11_05（追加） - # 原因: #166 占星师 2026-11-05 答复（BUG-1613～1617）改了 shadbala.py 的标签与共用水星规则函数，数值不变，ERR-110 另存新记录 + # 原因: #166 占星师 2026-11-05 答复（BUG-1622～1626）改了 shadbala.py 的标签与共用水星规则函数，数值不变，ERR-110 另存新记录 "docs/research/sealed_holdout_rerun_astrologer_166c_2026_11_05", "docs/research/reported_offset_astrologer_166c_2026_11_05", "docs/research/history/", diff --git a/tests/test_shadbala_service_api.py b/tests/test_shadbala_service_api.py index 8c89b4bb..2f9d371c 100644 --- a/tests/test_shadbala_service_api.py +++ b/tests/test_shadbala_service_api.py @@ -108,7 +108,7 @@ def test_birth_data_request_returns_precise_shadbala(api_base) -> None: "saptavargaja": "bphs_27_saptavargaja_20_15_10_4_2_d1_compound", # 原值: 无 saptavargaja_moolatrikona、drik_mercury_jupiter；mercury_nature "bphs_3_11_any_malefic_same_sign_no_nodes" # 新值: 补两项待核标签；mercury_nature 改名写明排除交点是实现选择（数值不变） - # 原因: BUG-1613 / BUG-1614，#166 占星师 2026-11-05 答复（第 17 / 19、23、25 条） + # 原因: BUG-1622 / BUG-1623，#166 占星师 2026-11-05 答复（第 17 / 19、23、25 条） "saptavargaja_moolatrikona": "d1_degree_band_only_provisional", "drik_special_aspects": "bphs_26_9_12_piecewise", "drik_mercury_jupiter": "quarter_only_raman_pending",