geometrybase/hk-ipo
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Use Chinese for analyst reports

Browse Source 
Request:
- Make analyst reports Chinese by default and record the rule in the analyst skill.

Changes:
- Add a Simplified Chinese default-language rule to the analyst skill.
- Update the single-IPO report generator to emit Chinese Markdown sections, labels, actions, risks, triggers, and exit plans.
- Preserve ticker symbols, stage codes, rule ids, score buckets, and source paths as machine-readable identifiers.
- Regenerate the 06106 T0 report in Chinese.
- Document the Chinese report default in README and the rule change log.

Verification:
- Ran py_compile for scripts/generate_ipo_report.py.
- Generated a 06106 dry-run report and checked Chinese section headings.
- Regenerated reports/2026-06-15_06106_T0_prospectus_analysis.md.
- Ran git diff --check.

Next useful context:
- Future analyst prediction and review reports should be written in Simplified Chinese unless the user explicitly requests another language.
This commit is contained in:
geometrybase
2026-06-15 14:37:46 +00:00
parent 07d7a0064a
commit 907e30d9da
5 changed files with 214 additions and 188 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.codex/skills/analyst/SKILL.md
+6
View File
@@ -85,6 +85,7 @@ If the ticker is absent from `data/snapshots/analysis_model_v0_dataset.csv`, use
Generated prediction reports must remain stage-safe: Generated prediction reports must remain stage-safe:
- Analyst Markdown reports should be written in Simplified Chinese by default, while preserving ticker symbols, stage codes, rule ids, source paths, and stable error tags in their original machine-readable form.
- T0 reports use only prospectus-stage fields and T0 calibration. - T0 reports use only prospectus-stage fields and T0 calibration.
- T1 reports may add allotment demand fields and T1 calibration. - T1 reports may add allotment demand fields and T1 calibration.
- T2/D1 is the intended sell window; D5/D20/D60 returns are never shown as prediction inputs and are reserved for later review cards. - T2/D1 is the intended sell window; D5/D20/D60 returns are never shown as prediction inputs and are reserved for later review cards.
@@ -111,6 +112,11 @@ Every prediction card should include:
- explicit T2/D1 sell discipline - explicit T2/D1 sell discipline
- source paths - source paths
Language standard:
- Write analyst reports, prediction cards, and review cards in Simplified Chinese by default.
- Keep field identifiers, model versions, score buckets, ticker symbols, and source paths as code-formatted English identifiers when they are part of the project data contract.
Every review card should include: Every review card should include:
- linked prediction card - linked prediction card
README.md
+1 -1
View File
@@ -170,7 +170,7 @@ Use the analyst report generator after the archive and model dataset are current
The generator writes `reports/{date}_{ticker}_{stage}_analysis.md` by default. It auto-selects `T1_allotment` when structured allotment-demand facts exist; otherwise it generates a `T0_prospectus` report. Use `--stdout` for a dry run or `--output` to choose a specific Markdown path. The generator writes `reports/{date}_{ticker}_{stage}_analysis.md` by default. It auto-selects `T1_allotment` when structured allotment-demand facts exist; otherwise it generates a `T0_prospectus` report. Use `--stdout` for a dry run or `--output` to choose a specific Markdown path.
Prediction reports are stage-safe: T0 reports use only prospectus-stage facts and T0 calibration, while T1 reports add allotment demand and T1 calibration. Each report includes a concrete stage calendar for that ticker: T0 subscription window, T1 allotment-result date, T2 grey-market date/window, and D1 listing date. Reports should frame the trade as a T2/D1 exit. Post-listing D5/D20/D60 performance stays out of prediction reports and is reserved for review cards. Single-IPO analyst reports are written in Simplified Chinese by default, while ticker symbols, stage codes, rule ids, and source paths remain machine-readable. Prediction reports are stage-safe: T0 reports use only prospectus-stage facts and T0 calibration, while T1 reports add allotment demand and T1 calibration. Each report includes a concrete stage calendar for that ticker: T0 subscription window, T1 allotment-result date, T2 grey-market date/window, and D1 listing date. Reports should frame the trade as a T2/D1 exit. Post-listing D5/D20/D60 performance stays out of prediction reports and is reserved for review cards.
## Incremental Archive Sync ## Incremental Archive Sync
reports/2026-06-15_06106_T0_prospectus_analysis.md
+72 -72
View File
@@ -1,96 +1,96 @@
# 06106 IPO Analyst Report # 06106 IPO 分析报告
## Summary ## 摘要
- Ticker: `06106` - 股票代码：`06106`
- Company: Shanghai Seer Intelligent Technology Co., Ltd. - 公司：Shanghai Seer Intelligent Technology Co., Ltd.
- Stage: `T0_prospectus` - 分析阶段：`T0_prospectus`
- Report as of: `2026-06-15T15:00:00Z` - 报告生成时间：`2026-06-15T15:00:00Z`
- Model dataset as of: `2026-06-15T14:04:34Z` - 模型数据时间：`2026-06-15T14:04:34Z`
- Rule version: `ipo_score_v0` - 规则版本：`ipo_score_v0`
- Rule path: `rules/ipo_score_v0.yaml` - 规则路径：`rules/ipo_score_v0.yaml`
- Strategy horizon: short IPO subscription trade; intended exit is T2 grey market if reliable, otherwise D1. - 策略周期：短线 IPO 申购交易；优先在可靠 T2 暗盘卖出，否则默认 D1 卖出。
- Decision: `strong_watch` - 结论代码：`strong_watch`
- PM action: Strong watch at T0, still pending T1 demand confirmation for a T2/D1 exit. - 执行动作：T0 强关注，仍需等待 T1 认购热度确认后执行 T2/D1 退出纪律。
- T0 score: `11` - T0 分数：`11`
- Score bucket: `t0_gte_8` - 分数分桶：`t0_gte_8`
- Calibrated D1 positive probability: 76.4% from 72 historical D1 labels - 历史校准 D1 正收益概率：76.4%，样本数 72
## Stage Calendar ## 阶段日期表
| Stage | Concrete Date For This IPO | Meaning | | 阶段 | 本 IPO 对应日期 | 含义 |
| --- | --- | --- | | --- | --- | --- |
| `T0_prospectus` | 2026-06-15 to 2026-06-18 | Subscription window; use prospectus and offer terms only. | | `T0_prospectus` | 2026-06-15 至 2026-06-18 | 申购前/申购中阶段；只使用招股书和发行条款。 |
| `T1_allotment` | 2026-06-23 | Allotment results day; use public demand, placing demand, and allocation facts. | | `T1_allotment` | 2026-06-23 | 分配结果日；使用公开认购热度、国际配售热度和分配事实。 |
| `T2_grey_market` | 2026-06-23 after allotment results | Pre-listing grey-market sell window if a reliable executable source exists. | | `T2_grey_market` | 2026-06-23 分配结果公布后 | 上市前暗盘窗口；只有存在可靠且可执行的数据源时才作为卖出依据。 |
| `D1` | 2026-06-24 | First official trading day; default sell window when T2 data is unavailable or unreliable. | | `D1` | 2026-06-24 | 正式上市首日；T2 数据不可用或不可靠时的默认卖出窗口。 |
## Facts ## 基础事实
| Field | Value | | 字段 | 数值 |
| --- | --- | | --- | --- |
| Board | Main Board | | 板块 | Main Board |
| Status | open_for_subscription | | 状态 | open_for_subscription |
| Listing date | 2026-06-24 | | 上市日期 | 2026-06-24 |
| Application period | 2026-06-15 to 2026-06-18 | | 申购期 | 2026-06-15 至 2026-06-18 |
| Allotment result date | 2026-06-23 | | 分配结果日期 | 2026-06-23 |
| Listing method | n/a | | 上市方式 | 未记录 |
| Industry | Industrial intelligent robots / robot controllers | | 行业 | Industrial intelligent robots / robot controllers |
| Sponsors | n/a | | 保荐人 | 未记录 |
| Offer price | HK$101.60 | | 发行价 | HK$101.60 |
| Offer size | HK$1,066.5m | | 发行规模 | HK$1,066.5m |
| Market cap | HK$11,226.5m | | 市值 | HK$11,226.5m |
| Board lot | 50 | | 每手股数 | 50 |
| Minimum subscription | HK$5,131.24 | | 最低认购金额 | HK$5,131.24 |
| Initial public offer percentage | 5.0% | | 初始公开发售比例 | 5.0% |
| Over-allotment shares | 1,574,550 | | 超额配股权股数 | 1,574,550 |
## Short-Exit Model Inference ## 短线退出模型推断
- D1 positive probability: 76.4% - D1 正收益概率：76.4%
- D1 >= 10% probability: 47.2% - D1 涨幅不低于 10% 概率：47.2%
- Historical average D1 return for bucket: 28.6% - 同分桶历史 D1 平均收益：28.6%
- Historical median D1 return for bucket: 9.6% - 同分桶历史 D1 中位收益：9.6%
- T2 sell return is not modeled until an approved grey-market data source exists. - T2 暗盘卖出收益暂未建模，直到项目确认可靠暗盘数据源。
- D5/D20/D60 outcomes are review labels only, not holding targets. - D5/D20/D60 只作为复盘标签，不是持仓目标。
## Score Breakdown ## 评分拆解
| Component | Points | Reason | | 评分项 | 分数 | 原因代码 |
| --- | ---: | --- | | --- | ---: | --- |
| Offer size | 4 | `800m_to_2000m` | | 发行规模 | 4 | `800m_to_2000m` |
| Initial public offer percentage | 3 | `lte_5pct` | | 初始公开发售比例 | 3 | `lte_5pct` |
| Minimum subscription | 2 | `3500_to_10000` | | 最低认购金额 | 2 | `3500_to_10000` |
| Offer price | 1 | `gte_100` | | 发行价 | 1 | `gte_100` |
| Over-allotment option | 1 | `present` | | 超额配股权 | 1 | `present` |
## Bull Points ## 正面因素
- Offer size: +4 (`800m_to_2000m`). - 发行规模：+4 (`800m_to_2000m`)。
- Initial public offer percentage: +3 (`lte_5pct`). - 初始公开发售比例：+3 (`lte_5pct`)。
- Minimum subscription: +2 (`3500_to_10000`). - 最低认购金额：+2 (`3500_to_10000`)。
- Offer price: +1 (`gte_100`). - 发行价：+1 (`gte_100`)。
- Over-allotment option: +1 (`present`). - 超额配股权：+1 (`present`)。
## Risks And Gaps ## 风险与缺口
- No material negative scoring component. - 没有明显负向评分项。
- No required report field is blank for this stage. - 本阶段必需字段没有明显空缺。
- T2 grey-market signal is not used yet because the project has no approved reproducible source. - T2 暗盘信号暂未使用，因为项目还没有批准可复现的数据源。
- Post-listing D5/D20/D60 outcomes are labels for later review only and are not holding-period targets. - 上市后的 D5/D20/D60 表现只用于后续复盘，不是本模型的持仓周期目标。
## Triggers ## 触发条件
- Upgrade: stronger verified T1 demand, better allocation scarcity, or a new rule-backed positive catalyst. - 上调：T1 认购热度显著更强、分配稀缺性更好，或出现有规则支持的新正面催化。
- Downgrade: weak public or international demand, oversized supply, low-quality missing fields, or adverse market window. - 下调：公开或国际需求偏弱、供给过大、关键字段质量不足，或市场窗口明显转差。
## Exit Plan ## 退出计划
- If subscribed and allocated, plan to sell in T2 grey market when reliable executable data is available. - 如果申购并获配，且 T2 暗盘数据可靠且可执行，优先按 T2 暗盘卖出计划处理。
- If T2 is unavailable or unreliable, use D1 as the default exit window. - 如果 T2 不可用或不可靠，默认使用 D1 作为卖出窗口。
- Do not treat D5/D20/D60 as planned holding periods for this model. - 不把 D5/D20/D60 作为本模型的计划持仓周期。
- Record D1/D5/D20/D60 outcomes later as review labels, not as retroactive prediction inputs. - 后续记录 D1/D5/D20/D60 结果时，只作为复盘标签，不作为倒推预测输入。
## Source Paths ## 来源路径
- `data/raw/06106/prospectus_candidate_2026-06-15.pdf` - `data/raw/06106/prospectus_candidate_2026-06-15.pdf`
rules/rule_change_log.md
+23
View File
@@ -1,5 +1,28 @@
# Rule Change Log # Rule Change Log
## 2026-06-15 - Use Chinese for analyst reports
Request:
- Make analyst reports Chinese by default and record the rule in the analyst skill.
Change:
- Added a Simplified Chinese default language rule to the analyst skill.
- Updated the single-ticker report generator to write Chinese Markdown reports while preserving ticker symbols, stage codes, rule ids, score buckets, and source paths as machine-readable identifiers.
- Regenerated the 06106 T0 report in Chinese.
- Documented the Chinese report default in README.
Rationale:
- The analysis workflow is intended for Chinese-language IPO subscription decisions, while project identifiers still need to remain stable for scripts and audits.
Verification:
- Generated a 06106 dry-run report and checked the Chinese sections.
- Regenerated `reports/2026-06-15_06106_T0_prospectus_analysis.md`.
- Ran py_compile for the report generator and git diff --check.
## 2026-06-15 - Add concrete stage dates to reports ## 2026-06-15 - Add concrete stage dates to reports
Request: Request:
scripts/generate_ipo_report.py
+112 -115
View File
@@ -104,49 +104,49 @@ def as_bool(value: Any) -> bool:
def fmt_value(value: Any) -> str: def fmt_value(value: Any) -> str:
if value in {None, ""}: if value in {None, ""}:
return "n/a" return "未记录"
return str(value) return str(value)
def fmt_num(value: float | None, suffix: str = "", decimals: int = 1) -> str: def fmt_num(value: float | None, suffix: str = "", decimals: int = 1) -> str:
if value is None: if value is None:
return "n/a" return "未记录"
return f"{value:,.{decimals}f}{suffix}" return f"{value:,.{decimals}f}{suffix}"
def fmt_pct_rate(value: float | None) -> str: def fmt_pct_rate(value: float | None) -> str:
if value is None: if value is None:
return "n/a" return "未记录"
return f"{value * 100:.1f}%" return f"{value * 100:.1f}%"
def fmt_pct_points(value: float | None) -> str: def fmt_pct_points(value: float | None) -> str:
if value is None: if value is None:
return "n/a" return "未记录"
return f"{value:.1f}%" return f"{value:.1f}%"
def fmt_money_m(value: float | None) -> str: def fmt_money_m(value: float | None) -> str:
if value is None: if value is None:
return "n/a" return "未记录"
return f"HK${value:,.1f}m" return f"HK${value:,.1f}m"
def fmt_hkd(value: float | None) -> str: def fmt_hkd(value: float | None) -> str:
if value is None: if value is None:
return "n/a" return "未记录"
return f"HK${value:,.2f}" return f"HK${value:,.2f}"
def fmt_times(value: float | None) -> str: def fmt_times(value: float | None) -> str:
if value is None: if value is None:
return "n/a" return "未记录"
return f"{value:,.2f}x" return f"{value:,.2f}x"
def fmt_int(value: int | None) -> str: def fmt_int(value: int | None) -> str:
if value is None: if value is None:
return "n/a" return "未记录"
return f"{value:,}" return f"{value:,}"
@@ -214,37 +214,37 @@ def t0_decision_band(score: int) -> str:
def action_for_decision(decision: str) -> str: def action_for_decision(decision: str) -> str:
actions = { actions = {
"weak_or_avoid": "Avoid at T0 unless later T1 demand changes the setup.", "weak_or_avoid": "T0 阶段回避，除非后续 T1 认购热度明显改变格局。",
"neutral": "Wait for T1 allotment demand before subscribing.", "neutral": "暂等 T1 分配结果，不在 T0 阶段主动下重注。",
"positive_watch": "Watch positively, but wait for T1 confirmation before sizing for a T2/D1 exit.", "positive_watch": "正面观察，但需要等 T1 确认后再决定 T2/D1 退出仓位。",
"strong_watch": "Strong watch at T0, still pending T1 demand confirmation for a T2/D1 exit.", "strong_watch": "T0 强关注，仍需等待 T1 认购热度确认后执行 T2/D1 退出纪律。",
"avoid": "Avoid subscription.", "avoid": "回避申购。",
"avoid_or_wait": "Avoid or wait; do not size without a stronger catalyst.", "avoid_or_wait": "回避或等待；没有更强催化前不放大仓位。",
"watch_or_small": "Small subscription only if execution constraints support a T2/D1 exit.", "watch_or_small": "仅在执行条件支持 T2/D1 退出时小额参与。",
"selective_subscribe": "Selective subscription with disciplined T2/D1 sell sizing.", "selective_subscribe": "选择性申购，并严格按 T2/D1 卖出纪律控制仓位。",
"high_conviction_subscribe": "Subscribe, subject to allocation, liquidity, and T2/D1 sell discipline.", "high_conviction_subscribe": "积极申购，但仍受分配、流动性和 T2/D1 卖出纪律约束。",
} }
return actions[decision] return actions[decision]
def component_label(name: str) -> str: def component_label(name: str) -> str:
labels = { labels = {
"offer_size": "Offer size", "offer_size": "发行规模",
"public_pct": "Initial public offer percentage", "public_pct": "初始公开发售比例",
"min_subscription": "Minimum subscription", "min_subscription": "最低认购金额",
"offer_price": "Offer price", "offer_price": "发行价",
"over_allotment": "Over-allotment option", "over_allotment": "超额配股权",
"public_os": "Public oversubscription", "public_os": "公开认购倍数",
"international_os": "International oversubscription", "international_os": "国际配售认购倍数",
"valid_applications": "Valid applications", "valid_applications": "有效申请数",
"success_rate": "Application success rate", "success_rate": "申请成功率",
"hk_reallocation": "HK public offer reallocation", "hk_reallocation": "香港公开发售回拨",
} }
return labels.get(name, name.replace("_", " ").title()) return labels.get(name, name.replace("_", " ").title())
def components_table(components: list[ScoreComponent]) -> str: def components_table(components: list[ScoreComponent]) -> str:
lines = ["| Component | Points | Reason |", "| --- | ---: | --- |"] lines = ["| 评分项 | 分数 | 原因代码 |", "| --- | ---: | --- |"]
for component in components: for component in components:
lines.append(f"| {component_label(component.name)} | {component.points} | `{component.reason}` |") lines.append(f"| {component_label(component.name)} | {component.points} | `{component.reason}` |")
return "\n".join(lines) return "\n".join(lines)
@@ -252,36 +252,36 @@ def components_table(components: list[ScoreComponent]) -> str:
def facts_table(record: dict[str, str], stage: str) -> str: def facts_table(record: dict[str, str], stage: str) -> str:
rows = [ rows = [
("Board", fmt_value(record["board"])), ("板块", fmt_value(record["board"])),
("Status", fmt_value(record["status"])), ("状态", fmt_value(record["status"])),
("Listing date", fmt_value(record["listing_date"])), ("上市日期", fmt_value(record["listing_date"])),
("Application period", f"{fmt_value(record['application_start_date'])} to {fmt_value(record['application_end_date'])}"), ("申购期", f"{fmt_value(record['application_start_date'])} 至 {fmt_value(record['application_end_date'])}"),
("Allotment result date", fmt_value(record["allotment_results_expected_date"])), ("分配结果日期", fmt_value(record["allotment_results_expected_date"])),
("Listing method", fmt_value(record["listing_method"])), ("上市方式", fmt_value(record["listing_method"])),
("Industry", fmt_value(record["industry_label"])), ("行业", fmt_value(record["industry_label"])),
("Sponsors", fmt_value(record["sponsors"])), ("保荐人", fmt_value(record["sponsors"])),
("Offer price", fmt_hkd(as_float(record["offer_price_hkd"]))), ("发行价", fmt_hkd(as_float(record["offer_price_hkd"]))),
("Offer size", fmt_money_m(as_float(record["offer_size_hkd_m"]))), ("发行规模", fmt_money_m(as_float(record["offer_size_hkd_m"]))),
("Market cap", fmt_money_m(as_float(record["market_cap_hkd_m"]))), ("市值", fmt_money_m(as_float(record["market_cap_hkd_m"]))),
("Board lot", fmt_int(as_int(record["board_lot"]))), ("每手股数", fmt_int(as_int(record["board_lot"]))),
("Minimum subscription", fmt_hkd(as_float(record["min_subscription_amount_hkd"]))), ("最低认购金额", fmt_hkd(as_float(record["min_subscription_amount_hkd"]))),
("Initial public offer percentage", fmt_pct_points(as_float(record["public_offer_pct_initial"]) * 100 if record["public_offer_pct_initial"] else None)), ("初始公开发售比例", fmt_pct_points(as_float(record["public_offer_pct_initial"]) * 100 if record["public_offer_pct_initial"] else None)),
("Over-allotment shares", fmt_int(as_int(record["over_allotment_offer_shares"]))), ("超额配股权股数", fmt_int(as_int(record["over_allotment_offer_shares"]))),
] ]
if stage == T1_STAGE: if stage == T1_STAGE:
rows.extend( rows.extend(
[ [
("Public oversubscription", fmt_times(as_float(record["public_oversubscription_times"]))), ("公开认购倍数", fmt_times(as_float(record["public_oversubscription_times"]))),
("International oversubscription", fmt_times(as_float(record["international_oversubscription_times"]))), ("国际配售认购倍数", fmt_times(as_float(record["international_oversubscription_times"]))),
("Valid applications", fmt_int(as_int(record["valid_applications"]))), ("有效申请数", fmt_int(as_int(record["valid_applications"]))),
("Successful applications", fmt_int(as_int(record["successful_applications"]))), ("成功申请数", fmt_int(as_int(record["successful_applications"]))),
("Application success rate", fmt_pct_points(as_float(record["application_success_rate"]) * 100 if record["application_success_rate"] else None)), ("申请成功率", fmt_pct_points(as_float(record["application_success_rate"]) * 100 if record["application_success_rate"] else None)),
("International placees", fmt_int(as_int(record["international_placees"]))), ("国际配售承配人数", fmt_int(as_int(record["international_placees"]))),
("HK offer reallocation multiple", fmt_times(as_float(record["hk_offer_reallocation_multiple"]))), ("香港公开发售回拨倍数", fmt_times(as_float(record["hk_offer_reallocation_multiple"]))),
] ]
) )
lines = ["| Field | Value |", "| --- | --- |"] lines = ["| 字段 | 数值 |", "| --- | --- |"]
for label, value in rows: for label, value in rows:
lines.append(f"| {label} | {value} |") lines.append(f"| {label} | {value} |")
return "\n".join(lines) return "\n".join(lines)
@@ -292,36 +292,36 @@ def stage_calendar_table(record: dict[str, str]) -> str:
application_end = fmt_value(record["application_end_date"]) application_end = fmt_value(record["application_end_date"])
allotment_date = fmt_value(record["allotment_results_expected_date"]) allotment_date = fmt_value(record["allotment_results_expected_date"])
listing_date = fmt_value(record["listing_date"]) listing_date = fmt_value(record["listing_date"])
if allotment_date != "n/a": if allotment_date != "未记录":
t2_date = f"{allotment_date} after allotment results" t2_date = f"{allotment_date} 分配结果公布后"
elif listing_date != "n/a": elif listing_date != "未记录":
t2_date = "trading day before D1; exact date not archived" t2_date = "D1 前一个交易日；精确日期未归档"
else: else:
t2_date = "n/a" t2_date = "未记录"
rows = [ rows = [
( (
"T0_prospectus", "T0_prospectus",
f"{application_start} to {application_end}", f"{application_start} 至 {application_end}",
"Subscription window; use prospectus and offer terms only.", "申购前/申购中阶段；只使用招股书和发行条款。",
), ),
( (
"T1_allotment", "T1_allotment",
allotment_date, allotment_date,
"Allotment results day; use public demand, placing demand, and allocation facts.", "分配结果日；使用公开认购热度、国际配售热度和分配事实。",
), ),
( (
"T2_grey_market", "T2_grey_market",
t2_date, t2_date,
"Pre-listing grey-market sell window if a reliable executable source exists.", "上市前暗盘窗口；只有存在可靠且可执行的数据源时才作为卖出依据。",
), ),
( (
"D1", "D1",
listing_date, listing_date,
"First official trading day; default sell window when T2 data is unavailable or unreliable.", "正式上市首日；T2 数据不可用或不可靠时的默认卖出窗口。",
), ),
] ]
lines = ["| Stage | Concrete Date For This IPO | Meaning |", "| --- | --- | --- |"] lines = ["| 阶段 | 本 IPO 对应日期 | 含义 |", "| --- | --- | --- |"]
for stage, date_text, meaning in rows: for stage, date_text, meaning in rows:
lines.append(f"| `{stage}` | {date_text} | {meaning} |") lines.append(f"| `{stage}` | {date_text} | {meaning} |")
return "\n".join(lines) return "\n".join(lines)
@@ -340,29 +340,29 @@ def reason_lines(components: list[ScoreComponent], positive: bool) -> list[str]:
filtered = [component for component in components if (component.points > 0 if positive else component.points < 0)] filtered = [component for component in components if (component.points > 0 if positive else component.points < 0)]
filtered.sort(key=lambda component: component.points, reverse=positive) filtered.sort(key=lambda component: component.points, reverse=positive)
if not filtered: if not filtered:
return ["- No material positive scoring component." if positive else "- No material negative scoring component."] return ["- 没有明显正向评分项。" if positive else "- 没有明显负向评分项。"]
return [f"- {component_label(component.name)}: {component.points:+d} (`{component.reason}`)." for component in filtered[:5]] return [f"- {component_label(component.name)}：{component.points:+d} (`{component.reason}`)。" for component in filtered[:5]]
def missing_field_lines(record: dict[str, str], stage: str) -> list[str]: def missing_field_lines(record: dict[str, str], stage: str) -> list[str]:
required = [ required = [
("industry_label", "industry label"), ("industry_label", "行业"),
("market_cap_hkd_m", "market cap"), ("market_cap_hkd_m", "市值"),
("min_subscription_amount_hkd", "minimum subscription"), ("min_subscription_amount_hkd", "最低认购金额"),
] ]
if stage == T1_STAGE: if stage == T1_STAGE:
required.extend( required.extend(
[ [
("public_oversubscription_times", "public oversubscription"), ("public_oversubscription_times", "公开认购倍数"),
("international_oversubscription_times", "international oversubscription"), ("international_oversubscription_times", "国际配售认购倍数"),
("valid_applications", "valid applications"), ("valid_applications", "有效申请数"),
("successful_applications", "successful applications"), ("successful_applications", "成功申请数"),
] ]
) )
missing = [label for key, label in required if not record.get(key)] missing = [label for key, label in required if not record.get(key)]
if not missing: if not missing:
return ["- No required report field is blank for this stage."] return ["- 本阶段必需字段没有明显空缺。"]
return [f"- Missing or blank: {', '.join(missing)}."] return [f"- 缺失或空白字段：{', '.join(missing)}。"]
def build_report(record: dict[str, str], rows: list[dict[str, str]], stage: str, as_of: str) -> str: def build_report(record: dict[str, str], rows: list[dict[str, str]], stage: str, as_of: str) -> str:
@@ -376,85 +376,82 @@ def build_report(record: dict[str, str], rows: list[dict[str, str]], stage: str,
decision = t0_decision_band(score) decision = t0_decision_band(score)
components = parse_components(record["t0_score_breakdown"]) components = parse_components(record["t0_score_breakdown"])
metric = bucket_metric(rows, "t0_score_bucket", bucket, require_t1=False) metric = bucket_metric(rows, "t0_score_bucket", bucket, require_t1=False)
score_label = "T0 score"
else: else:
score = as_int(record["total_score"]) or 0 score = as_int(record["total_score"]) or 0
bucket = record["total_score_bucket"] bucket = record["total_score_bucket"]
decision = record["decision_band"] decision = record["decision_band"]
components = parse_components(record["t0_score_breakdown"]) + parse_components(record["t1_score_breakdown"]) components = parse_components(record["t0_score_breakdown"]) + parse_components(record["t1_score_breakdown"])
metric = bucket_metric(rows, "total_score_bucket", bucket, require_t1=True) metric = bucket_metric(rows, "total_score_bucket", bucket, require_t1=True)
score_label = "Total score"
paths = source_paths(record, stage) paths = source_paths(record, stage)
source_lines = [f"- `{path}`" for path in paths] or ["- No source path recorded for this stage."] source_lines = [f"- `{path}`" for path in paths] or ["- 本阶段没有记录来源路径。"]
lines = [ lines = [
f"# {ticker} IPO Analyst Report", f"# {ticker} IPO 分析报告",
"", "",
"## Summary", "## 摘要",
"", "",
f"- Ticker: `{ticker}`", f"- 股票代码：`{ticker}`",
f"- Company: {fmt_value(record['company_name_en'])}", f"- 公司：{fmt_value(record['company_name_en'])}",
f"- Stage: `{stage}`", f"- 分析阶段：`{stage}`",
f"- Report as of: `{as_of}`", f"- 报告生成时间：`{as_of}`",
f"- Model dataset as of: `{dataset_as_of}`", f"- 模型数据时间：`{dataset_as_of}`",
f"- Rule version: `{model_version}`", f"- 规则版本：`{model_version}`",
f"- Rule path: `{MODEL_RULE_PATH.as_posix()}`", f"- 规则路径：`{MODEL_RULE_PATH.as_posix()}`",
"- Strategy horizon: short IPO subscription trade; intended exit is T2 grey market if reliable, otherwise D1.", "- 策略周期：短线 IPO 申购交易；优先在可靠 T2 暗盘卖出，否则默认 D1 卖出。",
f"- Decision: `{decision}`", f"- 结论代码：`{decision}`",
f"- PM action: {action_for_decision(decision)}", f"- 执行动作：{action_for_decision(decision)}",
f"- {score_label}: `{score}`", f"- {'T0 分数' if stage == T0_STAGE else '总分'}：`{score}`",
f"- Score bucket: `{bucket}`", f"- 分数分桶：`{bucket}`",
f"- Calibrated D1 positive probability: {fmt_pct_rate(metric.d1_positive_rate)} from {metric.sample_size} historical D1 labels", f"- 历史校准 D1 正收益概率：{fmt_pct_rate(metric.d1_positive_rate)}，样本数 {metric.sample_size}",
"", "",
"## Stage Calendar", "## 阶段日期表",
"", "",
stage_calendar_table(record), stage_calendar_table(record),
"", "",
"## Facts", "## 基础事实",
"", "",
facts_table(record, stage), facts_table(record, stage),
"", "",
"## Short-Exit Model Inference", "## 短线退出模型推断",
"", "",
f"- D1 positive probability: {fmt_pct_rate(metric.d1_positive_rate)}", f"- D1 正收益概率：{fmt_pct_rate(metric.d1_positive_rate)}",
f"- D1 >= 10% probability: {fmt_pct_rate(metric.d1_strong_rate)}", f"- D1 涨幅不低于 10% 概率：{fmt_pct_rate(metric.d1_strong_rate)}",
f"- Historical average D1 return for bucket: {fmt_num(metric.average_d1_return_pct, '%')}", f"- 同分桶历史 D1 平均收益：{fmt_num(metric.average_d1_return_pct, '%')}",
f"- Historical median D1 return for bucket: {fmt_num(metric.median_d1_return_pct, '%')}", f"- 同分桶历史 D1 中位收益：{fmt_num(metric.median_d1_return_pct, '%')}",
"- T2 sell return is not modeled until an approved grey-market data source exists.", "- T2 暗盘卖出收益暂未建模，直到项目确认可靠暗盘数据源。",
"- D5/D20/D60 outcomes are review labels only, not holding targets.", "- D5/D20/D60 只作为复盘标签，不是持仓目标。",
"", "",
"## Score Breakdown", "## 评分拆解",
"", "",
components_table(components), components_table(components),
"", "",
"## Bull Points", "## 正面因素",
"", "",
*reason_lines(components, positive=True), *reason_lines(components, positive=True),
"", "",
"## Risks And Gaps", "## 风险与缺口",
"", "",
*reason_lines(components, positive=False), *reason_lines(components, positive=False),
*missing_field_lines(record, stage), *missing_field_lines(record, stage),
"- T2 grey-market signal is not used yet because the project has no approved reproducible source.", "- T2 暗盘信号暂未使用，因为项目还没有批准可复现的数据源。",
"- Post-listing D5/D20/D60 outcomes are labels for later review only and are not holding-period targets.", "- 上市后的 D5/D20/D60 表现只用于后续复盘，不是本模型的持仓周期目标。",
"", "",
"## Triggers", "## 触发条件",
"", "",
"- Upgrade: stronger verified T1 demand, better allocation scarcity, or a new rule-backed positive catalyst.", "- 上调：T1 认购热度显著更强、分配稀缺性更好，或出现有规则支持的新正面催化。",
"- Downgrade: weak public or international demand, oversized supply, low-quality missing fields, or adverse market window.", "- 下调：公开或国际需求偏弱、供给过大、关键字段质量不足，或市场窗口明显转差。",
"", "",
"## Exit Plan", "## 退出计划",
"", "",
"- If subscribed and allocated, plan to sell in T2 grey market when reliable executable data is available.", "- 如果申购并获配，且 T2 暗盘数据可靠且可执行，优先按 T2 暗盘卖出计划处理。",
"- If T2 is unavailable or unreliable, use D1 as the default exit window.", "- 如果 T2 不可用或不可靠，默认使用 D1 作为卖出窗口。",
"- Do not treat D5/D20/D60 as planned holding periods for this model.", "- 不把 D5/D20/D60 作为本模型的计划持仓周期。",
"- Record D1/D5/D20/D60 outcomes later as review labels, not as retroactive prediction inputs.", "- 后续记录 D1/D5/D20/D60 结果时，只作为复盘标签，不作为倒推预测输入。",
"", "",
"## Source Paths", "## 来源路径",
"", "",
*source_lines, *source_lines,
"",
] ]
return "\n".join(lines) return "\n".join(lines)