Morgan Stanley 风格投资演示生成器#

适用场景#

场景说明首次覆盖启动Initiation of Coverage — 公司/行业首次研究报告季度业绩更新Quarterly Earnings Update — 财报发布后的更新演示投资委员会 MemoInvestment Committee Memo — IC 审批材料路演演示Roadshow / Non-Deal Roadshow — 面向机构投资者管理层会议Management Meeting Prep — 与公司管理层交流材料并购推介M&A Pitch Book — 买方/卖方财务顾问材料

核心能力#

本技能可生成 18 种专业页面类型，覆盖投行研究演示的全流程：

页面类型一览#

#页面类型函数说明1封面页slide_cover()公司名、评级、目标价、行业观点、What's Changed 摘要2Key Takeawaysslide_key_takeaways()3-5 条核心要点，金色数字放大3Section Dividerslide_section_divider()章节分隔页，深蓝背景4Content 内容页slide_content()左文右图布局，支持饼图/数据卡片5Metric Blocksslide_metric_blocks()KPI 指标块（2x3 网格）6Rating Tableslide_rating_table()投资评级表格（Ticker/Rating/TP/Company）7Value Chainslide_value_chain()产业链/价值链可视化8What's Changedslide_whats_changed()变动追踪（评级/TP/观点变动）9Thesis in Chartsslide_thesis_in_charts()核心论点图表（2x2 网格）10Shovel Stocksslide_shovel_stocks()铲子股/概念股列表11Market Monitorslide_market_monitor()市场监测（四色卡片）12Financial Chartslide_financial_chart()财务图表（柱+折线叠加，含预测路径）13Executive Summaryslide_executive_summary()执行摘要（四象限：论点/TP/风险/催化剂）14Scenario Comparisonslide_scenario_comparison()DCF 情景对比（Bear/Base/Bull）15WACC Breakdownslide_wacc_breakdown()WACC 拆解表格16Valuation Bridgeslide_valuation_bridge()估值桥（从 NOPAT 到 Equity Value）17Sensitivity Heatmapslide_sensitivity_heatmap()敏感性热力图（WACC x TGR 矩阵）18Disclosureslide_disclosure()免责声明页

快速开始#

Python API#

python

from ms_investment_deck import make_deck, sample_data

# 使用内置样例数据生成演示文稿
data = sample_data()
make_deck(data, "output/my_deck.pptx", theme="classic", language="zh")

# 自定义数据
my_data = {
    "company_name": "示例公司",
    "title_cn": "投资标题",
    "title_en": "Investment Title",
    "subtitle": "副标题说明",
    "rating": "Overweight",
    "target_price": "$125",
    "current_price": "$98.5",
    "date_str": "2026-06-13",
    "analyst": "张三 · CFA",
    "research_type": "Foundation",
    "key_takeaways": [
        "要点一：核心投资逻辑",
        "要点二：关键数据支撑",
        "要点三：风险提示",
    ],
    "metrics": [
        {"value": "$600B", "unit": "Market Size 2026E", "label": "市场规模"},
        {"value": "+38%", "unit": "YoY", "label": "同比增速"},
    ],
    "content_pages": [
        {
            "title_cn": "行业分析",
            "title_en": "Industry Analysis",
            "body": ["分析正文段落一", "", "分析正文段落二"],
        },
    ],
}
make_deck(my_data, "output/custom.pptx")

CLI#

bash

# 使用内置样例数据生成
python scripts/run.py -o output/demo.pptx

# 指定语言
python scripts/run.py -o output/demo_en.pptx --lang en

# 指定主题
python scripts/run.py -o output/demo.pptx --theme classic

页面结构#

标准演示文稿的页面顺序（数据字段存在时自动插入）：

text

1.  封面页（Cover）                          — 始终生成
2.  What's Changed                            — whats_changed 字段存在时
3.  Key Takeaways                            — key_takeaways 字段存在时
4.  Section Divider: 分析内容                 — content_pages 存在时
5.  Content Pages (x N)                      — 每个内容页
6.  Section Divider: 关键指标                 — metrics 存在时
7.  Metric Blocks                            — KPI 指标块
8.  Financial Chart                         — financial_chart 存在时
9.  WACC Breakdown                           — scenarios 存在时
10. Valuation Bridge                         — scenarios 存在时
11. Sensitivity Heatmap                      — sensitivity 存在时
12. Section Divider: 投资评级总览             — rating_table 存在时
13. Rating Table                             — 评级表格
14. Section Divider: 铲子股清单             — shovel_stocks 存在时
15. Shovel Stocks                            — 铲子股列表
16. Market Monitor                           — market_monitor 存在时
17. Executive Summary                        — executive_summary 存在时
18. Thesis in Charts                         — thesis_charts 存在时
19. Scenario Comparison                      — scenarios 存在时
20. Value Chain                              — value_chain 存在时
21. Disclosure                               — 始终生成

数据字典#

顶层字段#

字段类型必填说明company_namestr否公司名称（封面显示）title_cnstr否中文标题title_enstr否英文标题subtitlestr否副标题ratingstr否投资评级：Overweight / Equal-weight / Underweighttarget_pricestr否目标价，如 "$125"current_pricestr否当前价格，如 "$98.5"date_strstr否日期字符串analyststr否分析师姓名research_typestr否研究类型：Foundation / Update / Flashindustry_viewstr否行业观点：Attractive / In-Line / Cautiousdisclosure_textstr否自定义免责声明文本

内容字段#

字段类型说明key_takeawaysList[str]3-5 条核心要点metricsList[dict]KPI 指标块数组content_pagesList[dict]内容页数组rating_tableList[List] 或 List[Dict]评级表格shovel_stocksList[dict]铲子股列表shovel_stocks_listList[dict]shovel_stocks 别名market_monitorList[dict]市场监测卡片financial_chartdict财务图表数据executive_summarydict执行摘要数据scenariosdictDCF 情景数据sensitivitydict敏感性矩阵数据whats_changedList[dict]变动摘要thesis_chartsList[dict]核心论点图表value_chainList[dict]产业链数据

metrics 条目格式#

python

{
    "value": "~$600B",       # 核心数字（金色放大）
    "unit": "AI CapEx 2026E", # 单位/时间
    "label": "全球 AI 相关资本开支",  # 说明文字
}

content_pages 条目格式#

python

{
    "title_cn": "中文标题",
    "title_en": "English Title",
    "section_cn": "章节中文名",   # 可选，用于 Section Divider
    "section_en": "Section Name",
    "body": ["段落一", "", "段落二"],  # 空字符串表示换行
    "chart_img": "",             # 图片路径（留空则显示占位卡片）
    "data": {"sectors_allocation": [...]},  # 可选，饼图数据
    "source": "Source: Company Data",
}

rating_table 格式#

python

# List[List] 格式
[
    ["NVDA.O", "Overweight", "$126.5", "NVIDIA", "AI compute 领导者"],
    ["TSM.N",  "Overweight", "$178.2", "Taiwan Semiconductor", "先进制程"],
]

# List[Dict] 格式（自动转换）
[
    {"ticker": "NVDA.O", "rating": "Overweight", "target_price": "$126.5",
     "company": "NVIDIA", "reason": "AI compute 领导者"},
]

shovel_stocks 条目格式#

python

{
    "rank": 1,
    "company": "Aurora Power Co.",
    "ticker": "AUR.N",
    "product": "高压直流输电与数据中心电力模块",
    "analyst": "S. Lee",
    "market_cap_mn": 3200.5,    # 市值（百万）
    "perf_1y_pct": 142.3,       # 一年涨幅（%）
    "rating": "Overweight",      # 可选，触发评级列
}

financial_chart 格式#

python

{
    "years": ["FY22", "FY23", "FY24", "FY25E", "FY26E", "FY27E", "FY28E"],
    "revenue": [1200.0, 1450.0, 1780.0, 2120.0, 2460.0, 2810.0, 3180.0],
    "ebitda":  [280.0, 348.0, 442.0, 550.0, 665.0, 780.0, 910.0],
    "margins": {
        "gross":  [55.0, 56.5, 58.2, 60.0, 61.5, 62.8, 63.9],
        "ebitda": [23.3, 24.0, 24.8, 25.9, 27.0, 27.8, 28.6],
        "net":    [14.2, 15.0, 16.1, 17.0, 17.8, 18.5, 19.1],
    },
    "takeaways": ["要点一", "要点二", "要点三"],
    "title_cn": "收入 / 利润率一览",   # 可选
    "title_en": "Revenue & EBITDA",    # 可选
}

executive_summary 格式#

python

{
    "thesis": "核心投资论点...",
    "tp_and_upside": "TP $125 (+27% upside)...",
    "key_risks": ["风险一", "风险二"],
    "catalysts": ["催化剂一", "催化剂二"],
    "title_cn": "执行摘要",   # 可选
    "title_en": "Executive Summary",  # 可选
}

scenarios 格式（DCF 情景分析）#

python

{
    "bear":  {"label": "Bear",  "wacc": 12.5, "tgr": 1.5, ...},
    "base":  {"label": "Base",  "wacc": 10.0, "tgr": 2.5, ...},
    "bull":  {"label": "Bull",  "wacc": 8.5,  "tgr": 3.5, ...},
}

sensitivity 格式#

python

{
    "wacc_range": [8.0, 9.0, 10.0, 11.0, 12.0, 13.0],
    "tgr_range":  [1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0],
    "matrix": [
        [120.0, 135.0, 152.0, 172.0, 198.0, 230.0],
        [110.0, 123.0, 138.0, 155.0, 176.0, 202.0],
        ...
    ],
}

whats_changed 条目格式#

python

{
    "action": "Initiated",       # Initiated / Reiterated / Upgraded / Downgraded
    "ticker": "NVDA.O",
    "company": "NVIDIA Corp",
    "prev_rating": "",
    "new_rating": "Overweight",
    "prev_tp": "",
    "new_tp": "$126.5",
    "reason": "AI compute 需求超预期",
}

thesis_charts 条目格式#

python

{
    "title": "AI CapEx 加速",
    "subtitle": "全球 AI 资本开支（$B）",
    "chart_type": "bar",       # bar / line / pie
    "data": {"labels": [...], "values": [...]},
    "takeaway": "2026E 预计达 $600B",
}

value_chain 条目格式#

python

{
    "stage": "上游",
    "items": [
        {"name": "HBM 内存", "companies": ["SK Hynix", "Samsung"], "share": "45%"},
    ],
}

主题配色#

当前内置 1 套经典主题（classic），采用 Morgan Stanley 品牌色系：

色彩Hex用途MS Navy#0B2C5C主色 / 标题栏 / 章节底色MS Gold#C8A951辅助强调 / 关键数字OW Green#1F7A3EOverweight 评级EW Orange#D97706Equal-weight 评级UW Red#B91C1CUnderweight 评级MS Brand Blue#00559F品牌蓝Chart Blue#3B81B9图表主色Chart Red#FB0301图表辅助色

API 参考#

主入口#

python

make_deck(data: Dict[str, Any], output_path: str,
          theme: str = "classic", language: str = "zh") -> str

构建并保存一份 Morgan Stanley 风格的研究演示文稿。所有数据字段均为可选，缺失则跳过对应页面。

参数：

• data — 结构化数据字典（参见数据字典）
• output_path — 输出 .pptx 文件路径
• theme — 主题名称（当前支持 "classic"）
• language — 语言："zh" / "en" / "bilingual"

返回： 成功保存的文件路径

示例数据#

python

sample_data() -> Dict[str, Any]

返回一份完整的样例数据字典，可直接传入 make_deck()。

python

sample_data_financial() -> Dict[str, Any]

返回一份财务图表专用样例数据（ financial_chart 字段）。

公共页面函数#

所有 slide_* 函数接受一个 pptx.slide.Slide 对象作为第一个参数，其余通过关键字参数传入。可直接调用以构建自定义页面序列。

函数关键参数slide_cover(slide, *, title_cn, title_en, subtitle, rating, target_price, current_price, company_name, analyst, date_str, research_type, industry_view, whats_changed, language)封面页slide_key_takeaways(slide, *, items, language, page_number, total_pages)核心要点slide_section_divider(slide, *, title_cn, title_en, category, language)章节分隔slide_content(slide, *, title_cn, title_en, body, chart_img, data, source, language, page_number, total_pages)内容页slide_metric_blocks(slide, *, metrics, language, page_number, total_pages)KPI 指标块slide_rating_table(slide, *, rows, language, page_number, total_pages)评级表格slide_value_chain(slide, *, data, language, page_number, total_pages)产业链slide_whats_changed(slide, *, data, language, page_number, total_pages)变动追踪slide_thesis_in_charts(slide, *, data, language, page_number, total_pages)论点图表slide_shovel_stocks(slide, *, stocks, language, page_number, total_pages)铲子股列表slide_market_monitor(slide, *, blocks, language, page_number, total_pages)市场监测slide_disclosure(slide, *, text, language, page_number, total_pages)免责声明slide_financial_chart(slide, *, years, revenue, ebitda, margins, takeaways, scenarios, title_cn, title_en, language, page_number, total_pages)财务图表slide_executive_summary(slide, *, thesis, tp_and_upside, key_risks, catalysts, title_cn, title_en, language, page_number, total_pages)执行摘要slide_scenario_comparison(slide, *, data, theme, language, page_number, total_pages)情景对比slide_wacc_breakdown(slide, *, data, theme, language, page_number, total_pages)WACC 拆解slide_valuation_bridge(slide, *, data, theme, language, page_number, total_pages)估值桥slide_sensitivity_heatmap(slide, *, data, theme, language, page_number, total_pages)敏感性热力图

依赖#

依赖版本说明Python>= 3.9运行时python-pptx>= 0.6.21PPTX 文件生成

安装依赖：

bash

pip install python-pptx

设计规范#

• 比例：16:9 宽屏（13.333" x 7.5"）
• 字体：英文 Arial，中文 微软雅黑 / Source Han Sans SC
• 关键数字：金色放大字号（36-48pt）
• 正文：13pt，深棕黑色 #251A1A
• 页码：右下角，9pt 灰色
• 标题栏：深蓝背景 + 白色文字
• 表格：交替行底色，评级颜色编码

---

Author: WANG DONG JIE ( @yjkj999999 | Clawhub)

Version: 1.1.0 | License: MIT | Category: Investment Presentation

适用于 CGMA/ACCA/CFA 持证人及投行/买方/卖方分析师。生成符合摩根士丹利品牌标准的投资演示文稿。