Skip to content

Latest commit

 

History

History
334 lines (265 loc) · 11.4 KB

File metadata and controls

334 lines (265 loc) · 11.4 KB

战略主体

战略主体(Strategic Actor)是 Omen 的核心抽象,用于描述并分析复杂战略博弈中的关键行动者。

Omen 提供可校验、可重复运行的自动化流程,能够从非结构化案例资料(人物、组织、历史事件等)中提取结构化战略主体,并输出图谱、时间线和画像结果,支持后续推演与复核。

核心价值

  • 本体建模:将人物、组织、事件映射到统一战略行动者语义层与本体,降低后续推演建模成本。
  • 标准输入层:把文本输入转化为结构化模型与状态快照,便于版本管理、回放和比较。
  • 数据源:战略主体产物直接作为图谱、时间线、画像和校验流程的统一数据源。
  • 工程化:全流程可本地运行、可重复执行,并提供确定性校验结果用于自动化门禁。

本体结构

actor_ontology 是战略主体的本体模型。它围绕核心的行动者承载“谁参与、发生了什么、彼此如何影响”,核心架构:

  • actors[] 每个元素代表一个参与者,分为两类:Actor(参与者)和 StrategicActor(行动者)。
  • products[] 定义了相关的产品或技术资产,如 productplatformtool 等。
  • influences[] 定义了参与者基于事件之间的关系,如 buildscompetes_withparticipates_inaffects 等。
  • events[] 记录了案例中的事实事件。

设计优势

  • 分层清晰:参与者、关系、事件三层可分别演化。
  • 事实与语义解耦:事实层可追溯,语义层可推理。
  • 面向工程落地:字段稳定,便于校验、回放、可视化。

关键概念

参与者

Actor 表示参与者,构成了战略互动网络中的角色。位于 actor_ontology.actors[],通过 type: Actor 区分,例如客户、团队、潜在用户等。

Actor 字段:

  • id:参与者唯一标识
  • name:参与者名称
  • role:参与者角色(如 customer/team)
  • profile:兴趣、影响力、对战略行动者的对齐程度等

示例片段

{
  "id": "actor-customer-pilot",
  "name": "10-person pilot team customer",
  "type": "Actor",
  "role": "customer",
  "profile": {
    "interest": "Improve R&D team standard, discover root causes, establish performance quantification standards.",
    "influence_level": "Medium (provided validation feedback)",
    "alignment_with_strategic_actor": "High (appreciated data-driven insights over subjective experience)"
  }
}

行动者

StrategicActor 是战略互动网络中的核心战略主体,位于 actor_ontology.actors[] 中,通过 type: StrategicActor 区分。

通过结构化 profile 深入描述行动者的背景、决策风格、价值主张和不可妥协约束等信息,这部分反映了“战略行动者是谁、如何决策、坚持什么”。

profile 字段:

  • background_facts:背景事实(经历、轨迹、关键经验)
  • strategic_style.decision_style:决策风格
  • strategic_style.value_proposition:价值主张
  • strategic_style.decision_preferences:决策偏好
  • strategic_style.non_negotiables:不可妥协约束

示例片段

{
  "id": "actor-founder-1",
  "name": "Chen Jiaxing",
  "type": "StrategicActor",
  "role": "founder",
  "profile": {
    "background_facts": {
      "key_experiences": [
        "1.5 years of exploration, development, and validation before launch"
      ]
    },
    "strategic_style": {
      "decision_style": "Data-driven, feedback-oriented",
      "value_proposition": "Replace process-driven tools with data-driven insights to achieve effective R&D management without disrupting developer focus.",
      "decision_preferences": [
        "Free community version to drive adoption",
        "Focus on product and technology, avoid distractions from non-product opinions",
        "Use own product for internal management"
      ],
      "non_negotiables": [
        "No manual task status updates required from developers",
        "No reliance on maturity models or industry standards for evaluation",
        "No process gates that sacrifice speed"
      ]
    }
  }
}

影响关系层

influences[] 描述“主体-主体/主体-事件/事件-产品”之间的关系,是图谱渲染与关系推断的输入层。

字段:

  • source / target:关系起点与终点
  • type:关系类型(如 buildscompetes_withparticipates_inaffects
  • description:关系解释

示例片段

[
  {
    "source": "actor-founder-1",
    "target": "product-x-developer",
    "type": "builds",
    "description": "Founder/lead actor builds and steers the core product asset."
  },
  {
    "source": "competitor-process-tools",
    "target": "actor-founder-1",
    "type": "influences",
    "description": "Competitor market presence influences founder strategic decisions."
  },
  {
    "source": "product-x-developer",
    "target": "competitor-process-tools",
    "type": "competes_with",
    "description": "Core product competes with traditional solutions."
  },
  {
    "source": "chen-jiaxing-1",
    "target": "product-x-developer",
    "type": "affects",
    "description": "Strategic decision affects the core product lifecycle."
  }
]

事件层

events[] 是案例中的事实事件序列,强调“可回溯事实”。

字段:

  • id / name / type:事件标识、名称和类型(如 launchpivotfunding_round
  • date:事件发生时间(格式 YYYY-MM
  • description:事件描述
  • actors_involved[]:事件的参与者列表
  • evidence_refs[]:事件描述中引用的原始证据(文章、演讲、访谈等)
  • is_strategy_related:是否与战略相关

示例片段

{
  "id": "chen-jiaxing-1",
  "name": "X-Developer Platform Launch",
  "type": "launch",
  "date": "2019-10",
  "description": "X-Developer platform launched after 1.5 years of exploration, development, and validation; free community version offered to help companies get started quickly.",
  "actors_involved": ["actor-founder-1"],
  "evidence_refs": [
    "# 从流程驱动到数据驱动,打造你的高效研发团队",
    "2019年10月24日",
    "经过一年半的探索、开发与验证,我们推出了一款研发效能分析与改进的平台:X-Developer,并提供了免费的社区版帮助企业快速起步。"
  ],
  "is_strategy_related": true
}

与 Strategy.events 的差异

actor_ontologystrategy_ontology 中都存在事件概念,但侧重点不同:

  • actor_ontology.events:围绕 StrategicActor 的事实层事件。偏“案例记录”,包含时间、证据、参与方,用于时间线与状态快照。
  • strategy_ontology:侧重“战略建模”的事件概念,偏“语义层事件”。通常存在于概念层/关系层(如 PlatformLaunchTeamAdoption),用于表达机制与推理关系。

可以理解为:前者回答“发生了什么”,后者回答“为什么这类事件在战略上有意义”。

示例片段

// strategy_ontology.json
"events": [
	{
		"name": "PlatformLaunch",
		"description": "Event of launching a platform after exploration and validation.",
		"category": "launch",
		"confidence": 1.0
	}
]

战略画像

战略画像是对 StrategicActor 的深度分析结果,对“决策风格-事件证据-战略行为”的综合归纳,用于:

  • 对外展示:快速说明该主体的战略人格
  • 内部复核:检查叙事与证据是否一致
  • 后续推演:作为上层解释与模拟的输入线索

输出文件为 analyze_persona.json,核心架构 persona_insight 说明:

  • narrative:行动者画像叙事
  • key_traits[]:关键特质与证据摘要
  • consistency_score:画像一致性评分 $[0,1]$

示例片段

{
  "persona_insight": {
    "narrative": "Chen Jiaxing emerges as a pragmatic, product-centric innovator whose strategic profile is defined by a deep-seated belief in empirical validation and developer-centric efficiency. The foundational 1.5-year period of exploration, development, and validation before launch is not merely a background fact but a core expression of his data-driven and feedback-oriented decision style...",
    "key_traits": [
      {
        "trait": "Empirically-Driven Innovator",
        "evidence_summary": "Invested 1.5 years in exploration and validation before launch, ensuring the product is rooted in data and real feedback, not assumptions."
      },
      {
        "trait": "Developer-Centric Pragmatist",
        "evidence_summary": "Non-negotiables explicitly forbid manual status updates and process gates that sacrifice speed, protecting developer focus as a core tenet."
      },
      {
        "trait": "Product-Led Strategist",
        "evidence_summary": "Decision preferences emphasize focusing on product/tech and using their own product internally; launch strategy uses a free version to drive organic adoption."
      }
    ],
    "consistency_score": 0.95
  }
}

战略状态快照

战略状态快照是某一时刻的事实层数据,生成物是 analyze_status.json,包含事件清单和关系描述,用于时间线和图谱的可视化展示。

核心结构:

  • timeline[]:事件时间线,每个事件包含 timenameeventdescriptionevidencestrategic 等字段
  • actor_graph:图谱节点和边,包含 nodes[]edges[]

示例片段

{
  "timeline": [
    {
      "id": "chen-jiaxing-1",
      "time": "2019-10",
      "name": "X-Developer Platform Launch",
      "event": "launch",
      "description": "X-Developer platform launched after 1.5 years of exploration, development, and validation; free community version offered...",
      "strategic": true
    }
  ],
  "actor_graph": {
    "nodes": [
      {
        "id": "actor-founder-1",
        "label": "Chen Jiaxing",
        "node_type": "strategic_actor",
        "role": "founder"
      },
      {
        "id": "product-x-developer",
        "label": "X-Developer",
        "node_type": "product"
      }
    ],
    "edges": [
      {
        "source": "actor-founder-1",
        "target": "product-x-developer",
        "label": "builds",
        "weight": 1.0
      }
    ]
  },
  "summary": {
    "timeline_event_count": 3,
    "actor_node_count": 9,
    "actor_edge_count": 12
  }
}

内置案例

仓库内置案例(cases/actors/):

案例文件 说明
chen-jiaxing.md X-Developer 创始人,数据驱动研发效能平台
elon-musk.md 特斯拉/SpaceX 创始人,多产业颠覆者
jack-ma.md 阿里巴巴创始人,电商生态构建者
jeff-bezos.md 亚马逊创始人,长期主义与飞轮效应
shenda.md 盛大集团创始人,多元化经营与平台转型案例
steve-jobs.md 苹果创始人,产品哲学与创新驱动

如何使用

如果你尚未安装 Omen,请先阅读快速开始,完成环境配置和模型接入。

首次使用:运行 CLI 生成一个简单案例:

omen analyze actor --doc chen-jiaxing

检查输出:查看生成的文件

ls output/actors/chen-jiaxing/

启动 UI:交互式探索图谱和时间线

streamlit run app/strategic_actor.py

尝试其他案例:如 elon-musksteve-jobs


进阶指南