dsh-novel-writer 1.5.0 → 2.0.0

package/README.en.md

@@ -0,0 +1,20 @@
+# dsh-novel-writer — Novel Writing Assistant Plugin (v2.0.0)
+
+A DSH (DeepSeek Harness) bundle plugin. **v2.0.0 ships a local semantic embedding engine (bge-small-zh-v1.5 ONNX, bundled, CPU inference, 0 token cost)** — semantic capabilities without any API fees.
+
+## Tools (14, each with an independent UI switch)
+novel_books / novel_chapters / novel_read / novel_keywords / novel_new_chapter / novel_import / novel_sentence_analysis / novel_sentence_config / novel_style_check / novel_plot / novel_settings / novel_summary / novel_continuity_check / **novel_semantic_search (new in v2.0.0)**
+
+## v2.0.0 — Semantic Layer
+- **Local engine** (lib/embedding.js): Xenova/bge-small-zh-v1.5 (quantized), 512-dim Chinese vectors, ~23MB, bundled; @huggingface/tokenizers + onnxruntime-web (WASM), local CPU, 0 token / 0 API cost;
+- **Lazy load** (~0.3s first call); automatic fallback to pure-rule mode if unavailable;
+- **Index cache**: per-book vectors at `.novel-writer/embedding/<book>.json` — repeated searches are instant;
+- **novel_semantic_search**: natural-language search across the whole book (foreshadowing clues, emotional scenes, setting mentions) — matches by meaning even when keywords differ;
+- **novel_style_check upgrade**: adds semantic similarity (target chapter vs rest of book) alongside rule-based fingerprint similarity;
+- **Switch: semanticEmbedding** — default ON (probe-based: auto-enabled when the model is present); turn it off in the sidebar「写作助手功能」panel to fully return to rule-only mode.
+
+## History (v0.3 → v1.6)
+Sentence-pattern analysis; emotion purification + AI review (strong/weak lexicon, pollution caveat); emotion quantification (Valence sliding window: variance/slope/conflict + implicit imagery + complexity score); worldview/genre/theme detection (modern/western/eastern + 15 genres + 35 themes); pragmatic-style review; settings tables (5) + foreshadowing registry + summaries + continuity audit + style check; analysis cache & report export (`.novel-writer/analysis/`).
+
+## Size
+~95MB total (23MB quantized model + 70MB runtime); zip 31MB. If disk is tight, disable the semantic switch and delete `lib/models/` + `node_modules/` to go back to rule-only.

package/README.md

@@ -1,36 +1,55 @@
-# dsh-novel-writer — 小说写作助手插件（v0.9.0：世界观层）
+# dsh-novel-writer — 小说写作助手插件（v2.0.0）
 
-DSH（DeepSeek Harness）小说写作助手 bundle 插件。宿主端零第三方依赖（仅 Node 内置模块），
-浏览器端仅依赖 Web GUI 自带的 react。
-**v0.9.0 在 v0.8.0 基础上补全「世界观层」：第五张设定表「世界观/用语规范」、文化基准自动判断（detect）、用语风格扫描、续写用语一致性检查**：分析缓存与报告导出、风格自检、伏笔登记表、
-关键词三字组/疑似人名、情感词典去噪、祈使句与环境分类改进、局域网开关可选放行。
+DSH（DeepSeek Harness）小说写作助手 bundle 插件。**v2.0.0 集成本地语义嵌入引擎（bge-small-zh-v1.5 ONNX，随插件分发，本地 CPU 推理，0 token 成本）**，在不增加任何 API 费用的前提下获得语义级能力。
 
-## 工具清单（13 个，全部带独立 UI 开关）
+## 工具清单（14 个，全部带独立 UI 开关）
 
-novel_books / novel_chapters / novel_read / novel_keywords / novel_new_chapter / novel_import / novel_sentence_analysis / novel_sentence_config / novel_style_check / novel_plot / **novel_settings（设定管理）** / **novel_summary（章节摘要）** / **novel_continuity_check（连贯性审计）**
+novel_books / novel_chapters / novel_read / novel_keywords / novel_new_chapter / novel_import / novel_sentence_analysis / novel_sentence_config / novel_style_check / novel_plot / novel_settings（设定管理）/ novel_summary（章节摘要）/ novel_continuity_check（连贯性审计）/ **novel_semantic_search（语义检索，v2.0.0 新增）**
 
-## v0.9.0 新增能力（世界观层）
-- **novel_settings 第五张表「世界观/用语规范」（category=worldview）**：
-  - `detect`：自动判断文化基准（western/eastern/mixed/unknown）+ 置信度 + 证据（中西词表按时代错置分类：器物/称谓/计量/宗教仪式/市井风貌/服饰/食物/制度）；
-  - 登记：name / basis / bannedWords（禁用词）/ recommended（替代词映射）/ ritual（仪式规范）；
-  - 文化基准由 AI 自行判断，不预设东西方。
-- **novel_continuity_check 用语风格扫描**：对照 worldview 禁用词表输出「用语冲突」候选（带建议替换词，未登记时用默认基准并提示 detect）；
-- **续写提示词**：新增世界观一致性检查步骤。
+## v2.0.0 新增能力（语义层）
 
-## v0.8.0 新增能力
+### 本地语义嵌入引擎（embedding.js）
+- 模型：Xenova/bge-small-zh-v1.5（512 维中文向量，约 91MB），随插件分发，开箱即用；
+- 运行时：**@huggingface/tokenizers**（官方轻量分词，294KB 零依赖）+ onnxruntime-web（WASM 推理），本地 CPU，0 token、0 API 费用；
+- 懒加载：首次调用才加载（约 0.3s）；加载失败自动回退纯规则，不影响任何既有功能；
+- 索引缓存：每书向量落盘 `.novel-writer/embedding/<书>.json`，重复检索秒开（47 万字约 3100 段，首次建索引约 20-30s，之后命中缓存）。
 
-- **novel_settings（四张表）**：人物卡/地点卡/道具清单/时间线，list/add/update/delete + scan 候选提取；
-- **novel_summary**：模型生成、插件存储的每章摘要（长书续写先读摘要）；
-- **novel_continuity_check**：对照设定表扫描全书 → 数字口径/人物缺场/别名/重复条目矛盾候选；
-- **novel_plot 字段化**：type/priority/relatedCharacters/locations/payoffCondition/mentionedIn/lastMentioned + scan 自动提及追踪；
-- **细节密度指标**：动作链密度/物件名词/感官词（千字归一）进 novel_sentence_analysis 与 novel_style_check；
-- **统一数据目录**：§BT§<书库根>/.novel-writer/§BT§ 下 plots/ settings/ summaries/ analysis/ audits/（伏笔旧位置自动迁移；分析缓存与关键词报告落盘到 analysis/）；
-- **UI 路径修复**：cordis 空字符串 root 回退 lastRoot、面板刷新按钮、打开/复制按钮不再禁用、显示 dataDir。
+### 新工具：novel_semantic_search（语义检索）
+用自然语言描述在全书中检索语义相关段落——找伏笔线索、情感场景、设定提及，**即使原文没有相同关键词也能命中**（如搜「压抑克制的时刻」能命中没有"压抑"二字的段落）。
 
-## v0.7.0 更新内容
+### novel_style_check 升级：语义级风格对比
+规则指纹相似度（句式/句长/情绪）之外，新增**语义相似度**（目标章 vs 全书其余部分向量余弦），双维度判定风格一致性。
 
-- **10 个工具全部带独立 UI 开关**（侧边栏「写作助手功能」面板）：novel_books / novel_chapters / novel_read / novel_keywords / novel_new_chapter / novel_import / novel_sentence_analysis / novel_sentence_config / novel_style_check / novel_plot，默认全开；
-- UI 更名：侧边栏入口/面板/横幅由「句式分析」改为「写作助手功能」；设置页卡片同步；
-- novel_sentence_config 升级：支持 tools 参数（如 `{ novel_plot: false }`）与返回各工具开关状态；
-- 关闭的工具调用时返回明确提示（模型可见）。
+### 功能开关：语义增强（semanticEmbedding）
+- **默认开启（探测式）**：模型存在即自动启用；用户可在侧边栏「写作助手功能」面板关闭，关闭后语义检索返回提示、风格检查跳过语义维度；
+- 关闭后插件完整回退到 v1.6.0 纯规则能力。
+
+## 历史能力（v0.3 → v1.6 累积）
+- 句式模式分析（九类句式分布/排列/段落/句长/情感曲线/风格指纹/密度）；
+- 情感净化 + AI 复核（强/弱情绪词分级、污染预警 caveat、强制抽查）；
+- 情感量化（Valence 滑动窗口：方差/斜率/矛盾指数 + 隐性意象 + 复杂度评分）；
+- 世界观/流派/题材检测（modern/western/eastern + 15 流派 + 35 题材，主副题材）；
+- 语用级审查（称谓/客套/仪式/语气，中西/现代禁用词）；
+- 设定管理五张表 + 伏笔登记 + 章节摘要 + 连贯性审计 + 风格自检；
+- 分析缓存与报告导出（`.novel-writer/analysis/`）。
+
+## 功能开关（侧边栏「写作助手功能」面板）
+- 总开关 enabled；autoAnalyze；14 个工具级开关；
+- 功能级：emotionCaveat（情感净化预警）/ genreTheme（题材与流派检测）/ emotionComplexity（情感量化）/ **semanticEmbedding（语义增强，v2.0.0）**。
+
+## 目录结构
 ```
+dsh-novel-writer-v2.0.0/
+├── lib/
+│   ├── index.js        # 宿主端（14 工具 + state 路由 + 开关门禁）
+│   ├── analysis.js     # 规则引擎（句式/情感净化/情感量化/题材）
+│   ├── embedding.js    # 本地语义嵌入引擎（bge-small-zh ONNX）
+│   ├── models/         # 模型文件（onnx/model.onnx 91MB + tokenizer）
+│   └── client.js       # 浏览器端（侧边栏面板 + 全部开关）
+├── node_modules/       # @huggingface/tokenizers + onnxruntime-web（随插件分发）
+├── test/               # 引擎 + client + e2e 测试
+└── skills/novel-writing/SKILL.md  # 模型使用指南
+```
+
+## 体积说明
+**约 95MB**（模型 23MB + 运行时 70MB），zip 31MB；已移除 transformers.js/onnxruntime-node/sharp 全家桶（省 85%）。如磁盘紧张可在「语义增强」开关关闭后删除 `lib/models/` 与 `node_modules/`，插件回到纯规则模式。

package/cordis.patch.yml

@@ -1,3 +1,4 @@
+# dsh-novel-writer v2.0.0 bundle patch: 14 tools (incl. novel_semantic_search) + local embedding engine.
 # dsh-novel-writer bundle patch: mounts the novel-writing assistant plugin
 # into the profile's loader tree. The entry's name is the plugin package
 # itself, resolved from the profile's node_modules.

package/lib/analysis.js

@@ -1,13 +1,11 @@
 /**
- * dsh-novel-writer — 句式模式分析引擎 (v1.5.0)
+ * dsh-novel-writer — 句式模式分析引擎 (v1.6.0 合并版)
  *
  * 融合 v0.3.0（深度分析）与 v0.4.0（轻量节奏参考）两版引擎：
  *   - 九类句式：陈述 / 环境 / 心理 / 对话 / 疑问 / 反问 / 感叹 / 祈使 / 省略留白
  *   - 排列规律：句式转移、2/3 连句模板、段首段尾句式、按章节的压缩排列序列（跑长编码）
  *   - 风格特征：句长分布、短长句占比、对话/心理/环境密度、主观性指数、风格指纹
  *   - 主观情感：轻量情感词典（喜/怒/哀/惧/惊）+ 强度副词加权，输出情感曲线
- *   - v1.5.0 情感净化：强/弱情绪词分级（clean 只计强情绪词）、污染源检测（R18/战斗/恐怖
- *     密度阈值）、caveat 预警与 aiAction AI 复核指令、confidence 可信度
  *   - 节奏建议：给模型的 guidance 文本（句式分布 + 高频组合 + 章节节奏序列）
  *   - 采样上限：maxSentences 保护超长文本（默认 20000 句）
  *
@@ -70,10 +68,6 @@ const EMOTION_WORDS = Object.freeze({
   fear: [...STRONG_EMOTION_WORDS.fear, ...WEAK_EMOTION_WORDS.fear],
   surprise: [...STRONG_EMOTION_WORDS.surprise, ...WEAK_EMOTION_WORDS.surprise]
 });
-/** 去重后的情感词表（模块级缓存，避免每次分析重复 Set 构建）。 */
-const EMOTION_WORDS_UNIQUE = Object.freeze(Object.fromEntries(
-  Object.entries(EMOTION_WORDS).map(([emotion, words]) => [emotion, [...new Set(words)]])
-));
 
 /** v1.5.0 情绪污染源词表：检测到高密度时降低情感可信度并提示 AI 复核。 */
 const EMOTION_POLLUTION = Object.freeze({
@@ -411,7 +405,302 @@ export function buildGuidance(ratios, topPatterns, chapterSequences) {
 }
 
 /**
- * 全书/单章句式模式分析主入口（v0.5.0 合并版）。
+ * v1.6.0 Valence 效价映射表（基于大连理工中文情感词汇本体库框架）：
+ * 七维（乐/好/怒/哀/惧/恶/惊）× 强度五档（1/3/4/7/9 → ±0.1/0.3/0.5/0.7/0.9）。
+ * 乐=正向，好=正向，怒/哀/惧/恶/惊=负向（惊=中性偏负）。纯规则查表，0 token。
+ */
+const VALENCE_WORDS = Object.freeze({
+  "欣喜": 0.9, "狂喜": 0.9, "欢天喜地": 0.9, "雀跃": 0.7, "高兴": 0.7, "开心": 0.7, "快乐": 0.7,
+  "喜悦": 0.7, "愉快": 0.5, "欢喜": 0.7, "兴奋": 0.7, "愉悦": 0.5, "欢快": 0.7, "乐": 0.5,
+  "欢": 0.5, "喜": 0.5, "笑": 0.3, "微笑": 0.3, "笑容": 0.3, "笑眯眯": 0.3, "哈哈": 0.3,
+  "欣慰": 0.5, "满足": 0.3, "痛快": 0.5, "爽快": 0.3, "甜": 0.5, "甜蜜": 0.5, "幸福": 0.9,
+  "美满": 0.9, "温馨": 0.5, "温暖": 0.3, "踏实": 0.1, "平静": 0.1, "释然": 0.4, "解脱": 0.4,
+  "喜欢": 0.7, "喜爱": 0.7, "欣赏": 0.5, "爱": 0.9, "疼爱": 0.7, "宠爱": 0.7, "仰慕": 0.7,
+  "尊敬": 0.5, "敬仰": 0.7, "崇拜": 0.7, "心动": 0.5, "眷恋": 0.7, "依恋": 0.7, "思念": 0.3,
+  "心疼": 0.3, "怜爱": 0.5, "温柔": 0.3, "珍惜": 0.5, "信赖": 0.5, "感恩": 0.7,
+  "愤怒": -0.9, "暴怒": -0.9, "怒发冲冠": -0.9, "火冒三丈": -0.9, "恼羞成怒": -0.9,
+  "生气": -0.7, "恼火": -0.7, "气愤": -0.7, "怒": -0.7, "恨": -0.9, "怨恨": -0.9, "憎恨": -0.9,
+  "厌恶": -0.7, "憎恶": -0.9, "不满": -0.5, "恼": -0.5, "气": -0.3, "发火": -0.7, "咬牙": -0.3,
+  "怒意": -0.7, "怒火": -0.9, "愤恨": -0.9, "恼羞": -0.7, "气冲冲": -0.7, "咬牙切齿": -0.5,
+  "悲伤": -0.7, "悲痛": -0.9, "悲痛欲绝": -0.9, "悲哀": -0.7, "哀伤": -0.7, "难过": -0.5,
+  "伤心": -0.7, "痛苦": -0.7, "心碎": -0.9, "绝望": -0.9, "哭泣": -0.7, "哭": -0.5, "泪": -0.3,
+  "眼泪": -0.3, "流泪": -0.5, "哽咽": -0.5, "抽泣": -0.5, "叹息": -0.3, "叹气": -0.3,
+  "惆怅": -0.5, "失落": -0.5, "忧伤": -0.5, "黯然": -0.5, "心酸": -0.7, "辛酸": -0.7,
+  "悲": -0.7, "凄凉": -0.7, "苦涩": -0.7, "苦闷": -0.5, "沮丧": -0.7, "消沉": -0.7,
+  "落寞": -0.5, "孤寂": -0.5, "郁闷": -0.3, "低落": -0.3, "压抑": -0.5, "心灰意冷": -0.9,
+  "恐惧": -0.9, "害怕": -0.7, "惊慌": -0.7, "不安": -0.3, "紧张": -0.3, "担心": -0.3,
+  "畏惧": -0.7, "惊恐": -0.9, "胆怯": -0.5, "发抖": -0.3, "哆嗦": -0.3, "心慌": -0.5,
+  "毛骨悚然": -0.9, "冷汗": -0.5, "忐忑": -0.5, "惶恐": -0.9, "心悸": -0.5, "惊惶": -0.7,
+  "胆战心惊": -0.9, "心虚": -0.5, "焦虑": -0.5, "恐慌": -0.9,
+  "恶心": -0.7, "厌恶": -0.7, "鄙视": -0.7, "轻蔑": -0.5, "嫌弃": -0.7, "反感": -0.5,
+  "憎恶": -0.9, "作呕": -0.7,
+  "惊讶": -0.1, "震惊": -0.5, "意外": -0.1, "吃惊": -0.3, "诧异": -0.3, "愕然": -0.3,
+  "愣住": -0.1, "目瞪口呆": -0.5, "难以置信": -0.5, "不可思议": -0.3, "惊愕": -0.5,
+  "惊奇": 0.1, "震撼": -0.3, "傻眼": -0.3, "呆住": -0.1, "惊呆": -0.5, "骇然": -0.5
+});
+
+/**
+ * v1.6.0 隐性情感载体映射表（意象/动作 → 效价 + 标签 + 脆弱标记）。
+ * 参考中国古典诗歌意象体系 + 现代微动作意象，规则查表 0 token。
+ */
+const IMPLICIT_CARRIERS = Object.freeze([
+  { words: ["雨", "阴雨", "细雨", "冷雨", "秋雨"], valence: -0.3, label: "压抑·萧瑟" },
+  { words: ["黄昏", "暮色", "残阳", "夕阳西下"], valence: -0.3, label: "迟暮·萧瑟" },
+  { words: ["枯枝", "落叶", "枯叶", "败叶"], valence: -0.3, label: "凋零·萧瑟" },
+  { words: ["冷风", "寒风", "北风", "秋风"], valence: -0.3, label: "寒冷·孤寂" },
+  { words: ["昏暗", "阴影", "灰暗", "幽暗"], valence: -0.3, label: "压抑" },
+  { words: ["孤雁", "孤鸿", "寒鸦"], valence: -0.25, label: "孤独" },
+  { words: ["残月", "冷月", "孤月"], valence: -0.25, label: "孤独·凄清" },
+  { words: ["梧桐", "芭蕉"], valence: -0.25, label: "愁绪" },
+  { words: ["寒蝉", "秋虫"], valence: -0.2, label: "凄切" },
+  { words: ["荒芜", "废墟", "断壁", "残垣"], valence: -0.4, label: "荒凉·衰败" },
+  { words: ["空荡", "空旷", "空落落"], valence: -0.3, label: "空虚" },
+  { words: ["暖光", "暖阳", "炉火", "烛火", "灯火"], valence: 0.2, label: "短暂温暖", fragile: true },
+  { words: ["茶烟", "炊烟", "轻烟"], valence: 0.2, label: "短暂温暖", fragile: true },
+  { words: ["余晖", "黄昏的光"], valence: 0.15, label: "短暂温暖", fragile: true },
+  { words: ["摩挲", "摩挲杯沿"], valence: -0.2, label: "焦虑·克制" },
+  { words: ["攥紧衣角", "攥着衣角", "握紧衣角"], valence: -0.2, label: "焦虑·克制" },
+  { words: ["咬唇", "咬住嘴唇", "咬着下唇"], valence: -0.2, label: "隐忍·克制" },
+  { words: ["低垂眼帘", "垂下眼帘", "垂下眼"], valence: -0.2, label: "隐忍·欲言又止" },
+  { words: ["沉默良久", "久久沉默"], valence: -0.2, label: "隐忍" },
+  { words: ["指尖发白", "指节发白", "攥紧拳头"], valence: -0.3, label: "压抑·愤怒" },
+  { words: ["颤抖的手", "手在抖"], valence: -0.3, label: "紧张·恐惧" },
+  { words: ["转身", "转过身"], valence: -0.3, label: "疏离·决绝" },
+  { words: ["走出", "推门而出", "大步离开"], valence: -0.3, label: "决绝" },
+  { words: ["背影", "远去的背影"], valence: -0.3, label: "疏离·失落" },
+  { words: ["回头", "回望"], valence: -0.2, label: "不舍·眷恋" },
+  { words: ["轻笑", "苦笑", "扯了扯嘴角"], valence: -0.15, label: "无奈·强颜" },
+  { words: ["摇头", "摇了摇头", "垂下头"], valence: -0.15, label: "无奈·妥协" },
+  { words: ["垂手", "放下手", "手垂落"], valence: -0.15, label: "无力·放弃" }
+]);
+
+/** v1.6.0 隐性载体扫描：意象/动作 → 负向/正向占比 + top 载体 + 脆弱标记。 */
+export function implicitEmotionScan(text) {
+  let negHits = 0, posHits = 0, fragileHits = 0;
+  const carrierCounts = new Map();
+  for (const carrier of IMPLICIT_CARRIERS) {
+    for (const word of carrier.words) {
+      const n = text.split(word).length - 1;
+      if (n > 0) {
+        if (carrier.valence < 0) negHits += n * -carrier.valence;
+        else { posHits += n * carrier.valence; if (carrier.fragile) fragileHits += n; }
+        carrierCounts.set(carrier.label + ":" + word, (carrierCounts.get(carrier.label + ":" + word) ?? 0) + n);
+      }
+    }
+  }
+  const total = negHits + posHits;
+  const top = [...carrierCounts.entries()].sort((x, y) => y[1] - x[1]).slice(0, 8).map(([k, n]) => ({ carrier: k.split(":")[1], label: k.split(":")[0], count: n }));
+  return {
+    negative: total === 0 ? 0 : Math.round((negHits / total) * 100) / 100,
+    positive: total === 0 ? 0 : Math.round((posHits / total) * 100) / 100,
+    fragile: fragileHits > 0,
+    totalHits: negHits + posHits,
+    topCarriers: top
+  };
+}
+
+/**
+ * v1.6.0 情感量化：Valence 滑动窗口时间序列 → 方差/斜率/矛盾指数。
+ */
+const EMOTION_CATS = ["joy", "anger", "sorrow", "fear", "surprise"];
+const EMOTION_MAX_ENTROPY = Math.log(EMOTION_CATS.length);
+
+/** v1.6.0 滑动窗口：每 100 字算平均效价 → 时间序列 + 正负词计数。 */
+export function valenceSeries(text, winChars = 100) {
+  const series = [];
+  const windowPosNeg = [];
+  const entries = Object.entries(VALENCE_WORDS).sort((x, y) => y[0].length - x[0].length);
+  for (let start = 0; start < text.length; start += winChars) {
+    const slice = text.slice(start, start + winChars);
+    let sum = 0, n = 0, pos = 0, neg = 0;
+    for (const [word, val] of entries) {
+      let from = 0;
+      while (from < slice.length) {
+        const idx = slice.indexOf(word, from);
+        if (idx === -1) break;
+        sum += val; n += 1;
+        if (val > 0) pos += 1; else if (val < 0) neg += 1;
+        from = idx + word.length;
+      }
+    }
+    series.push(n === 0 ? 0 : Math.round((sum / n) * 1000) / 1000);
+    windowPosNeg.push({ pos, neg });
+  }
+  const posWords = windowPosNeg.reduce((s, w) => s + w.pos, 0);
+  const negWords = windowPosNeg.reduce((s, w) => s + w.neg, 0);
+  return { series, posWords, negWords, windowCount: series.length, windowPosNeg };
+}
+
+/** v1.6.0 三指标：方差 V + 相邻撕裂 V_adj + 斜率 Δ（最小二乘+鲁棒版）+ 矛盾指数 C。 */
+export function valenceStats(text) {
+  const { series, posWords, negWords } = valenceSeries(text);
+  const n = series.length;
+  if (n === 0) return { windows: 0 };
+  const mean = series.reduce((x, y) => x + y, 0) / n;
+  const variance = series.reduce((s, x) => s + (x - mean) ** 2, 0) / n;
+  let adjSum = 0;
+  for (let i = 1; i < n; i += 1) adjSum += Math.abs(series[i] - series[i - 1]);
+  const adjVariance = n > 1 ? adjSum / (n - 1) : 0;
+  const iMean = (n - 1) / 2;
+  let num = 0, den = 0;
+  for (let i = 0; i < n; i += 1) {
+    num += (i - iMean) * (series[i] - mean);
+    den += (i - iMean) ** 2;
+  }
+  const slope = den === 0 ? 0 : num / den;
+  const delta = slope * (n - 1);
+  const third = Math.max(1, Math.floor(n / 3));
+  const headMean = series.slice(0, third).reduce((x, y) => x + y, 0) / third;
+  const tailMean = series.slice(-third).reduce((x, y) => x + y, 0) / third;
+  const { windowPosNeg } = valenceSeries(text);
+  const totalWords = posWords + negWords;
+  const posRatio = totalWords === 0 ? 0 : posWords / totalWords;
+  const negRatio = totalWords === 0 ? 0 : negWords / totalWords;
+  // 坑1方案：矛盾指数按"窗口内原始词"算再平均（避免全书平均掩盖"同窗交织"vs"分段喜悲"）
+  const windowConflicts = windowPosNeg
+    .filter((w) => w.pos + w.neg > 0)
+    .map((w) => 2 * Math.min(w.pos / (w.pos + w.neg), w.neg / (w.pos + w.neg)));
+  const conflict = windowConflicts.length === 0 ? 0 : windowConflicts.reduce((x, y) => x + y, 0) / windowConflicts.length;
+  return {
+    windows: n,
+    variance: Math.round(variance * 1000) / 1000,
+    adjVariance: Math.round(adjVariance * 1000) / 1000,
+    delta: Math.round(delta * 1000) / 1000,
+    deltaRobust: Math.round((tailMean - headMean) * 1000) / 1000,
+    conflict: Math.round(conflict * 1000) / 1000,
+    posRatio: Math.round(posRatio * 1000) / 1000,
+    negRatio: Math.round(negRatio * 1000) / 1000,
+    meanValence: Math.round(mean * 1000) / 1000
+  };
+}
+
+/** v1.6.0 显隐对比：显性均值 vs 隐性方向 → 表里不一。 */
+export function explicitImplicitCompare(explicitMean, implicit) {
+  if (!implicit || implicit.totalHits === 0) return { explicitImplicitConflict: false, explicitSign: "neutral", implicitSign: "neutral" };
+  const explicitSign = explicitMean > 0.15 ? "positive" : explicitMean < -0.15 ? "negative" : "neutral";
+  const implicitSign = implicit.negative >= 0.6 ? "negative" : implicit.positive >= 0.6 ? "positive" : "neutral";
+  return {
+    explicitImplicitConflict: explicitSign === "positive" && implicitSign === "negative",
+    explicitSign,
+    implicitSign
+  };
+}
+
+/** v1.6.0 复合情感共现（规则）：同段多情感类别 → 高频矛盾对。 */
+export function compositeEmotionPairs(blocks) {
+  const pairCounts = new Map();
+  for (const block of blocks) {
+    const present = new Set();
+    for (const sentence of block) {
+      const e = sentence.emotion;
+      for (const k of EMOTION_CATS) if ((e.cleanScores?.[k] ?? 0) > 0) present.add(k);
+    }
+    const list = [...present];
+    for (let i = 0; i < list.length; i += 1) {
+      for (let j = i + 1; j < list.length; j += 1) {
+        const key = [list[i], list[j]].sort().join("+");
+        pairCounts.set(key, (pairCounts.get(key) ?? 0) + 1);
+      }
+    }
+  }
+  const labels = {
+    "joy+sorrow": "悲喜交加", "joy+anger": "又爱又恨/喜怒交织", "sorrow+anger": "哀怒交加/愤懑",
+    "fear+sorrow": "悲伤恐惧", "joy+fear": "惊喜交加", "anger+fear": "惊惧愤怒", "sorrow+surprise": "愕然悲伤"
+  };
+  return [...pairCounts.entries()].map(([pair, count]) => ({ pair, count, label: labels[pair] ?? "复合情感" }))
+    .sort((x, y) => y.count - x.count).slice(0, 5);
+}
+
+/** v1.6.0 单章分布（五维 clean）→ 熵/多样性/主次。 */
+export function chapterEmotionStats(counts) {
+  const total = EMOTION_CATS.reduce((s, k) => s + (counts[k] ?? 0), 0);
+  if (total === 0) return null;
+  const p = EMOTION_CATS.map((k) => (counts[k] ?? 0) / total);
+  const entropy = -p.filter((x) => x > 0).reduce((s, x) => s + x * Math.log(x), 0);
+  const diversity = p.filter((x) => x >= 0.15).length;
+  const sorted = EMOTION_CATS.map((k, i) => ({ emotion: k, ratio: p[i] })).sort((x, y) => y.ratio - x.ratio);
+  return {
+    entropy: Math.round(entropy * 1000) / 1000,
+    diversity,
+    dominant: sorted[0].emotion,
+    dominantRatio: Math.round(sorted[0].ratio * 1000) / 1000,
+    secondary: sorted[1].emotion,
+    secondaryRatio: Math.round(sorted[1].ratio * 1000) / 1000
+  };
+}
+
+/** v1.6.0 全书聚合：复杂度评分 0-1（熵归一化0.5 + 多样性0.25 + 主次冲突0.25）+ 章间漂移。 */
+export function emotionComplexity(perChapter) {
+  const stats = perChapter.map((c) => ({ chapter: c.chapter, ...(chapterEmotionStats(c.counts) ?? {}) })).filter((s) => s.entropy !== void 0);
+  if (stats.length === 0) {
+    // 无情感词命中：不复杂（low），恒有值供模型读取
+    return {
+      score: 0, level: "low", entropy: 0, maxEntropy: Math.round(EMOTION_MAX_ENTROPY * 1000) / 1000,
+      diversity: 0, dominant: "neutral", dominantRatio: 0, secondary: "neutral", secondaryRatio: 0,
+      conflict: "", conflictStrength: 0,
+      chapterDrift: { meanEntropy: 0, entropyVariance: 0, swinging: false }
+    };
+  }
+  const totalCounts = {};
+  for (const c of perChapter) for (const k of EMOTION_CATS) totalCounts[k] = (totalCounts[k] ?? 0) + (c.counts[k] ?? 0);
+  const global = chapterEmotionStats(totalCounts);
+  const entropies = stats.map((s) => s.entropy);
+  const meanEntropy = entropies.reduce((x, y) => x + y, 0) / entropies.length;
+  const entropyVariance = entropies.reduce((s, e) => s + (e - meanEntropy) ** 2, 0) / entropies.length;
+  const meanDiversity = stats.reduce((s, x) => s + x.diversity, 0) / stats.length;
+  const entropyNorm = global ? global.entropy / EMOTION_MAX_ENTROPY : 0;
+  const conflictStrength = global ? global.secondaryRatio / Math.max(global.dominantRatio, 0.0001) : 0;
+  const score = Math.min(1, Math.max(0, entropyNorm * 0.5 + (meanDiversity / 5) * 0.25 + conflictStrength * 0.25));
+  const level = score >= 0.6 ? "high" : score >= 0.4 ? "medium" : "low";
+  return {
+    score: Math.round(score * 100) / 100,
+    level,
+    entropy: global ? global.entropy : 0,
+    maxEntropy: Math.round(EMOTION_MAX_ENTROPY * 1000) / 1000,
+    diversity: global ? global.diversity : 0,
+    dominant: global?.dominant ?? "neutral",
+    dominantRatio: global?.dominantRatio ?? 0,
+    secondary: global?.secondary ?? "neutral",
+    secondaryRatio: global?.secondaryRatio ?? 0,
+    conflict: global ? global.dominant + "↔" + global.secondary : "",
+    conflictStrength: Math.round(conflictStrength * 100) / 100,
+    chapterDrift: {
+      meanEntropy: Math.round(meanEntropy * 1000) / 1000,
+      entropyVariance: Math.round(entropyVariance * 10000) / 10000,
+      swinging: entropyVariance > 0.03
+    }
+  };
+}
+
+/**
+ * v1.6.0 情感量化入口：三指标 + 显隐对比 + 复杂度 + 复合共现（纯规则 0 token）。
+ */
+export function emotionalQuantification(text, perChapter, blocks) {
+  const stats = valenceStats(text);
+  const implicit = implicitEmotionScan(text);
+  const meanValence = stats.meanValence ?? 0;
+  // perChapter 为空（单章/无分章输入）时，用当前文本自身 clean 计数聚合，保证 complexity 恒有值
+  if (!Array.isArray(perChapter) || perChapter.length === 0) {
+    const counts = { joy: 0, anger: 0, sorrow: 0, fear: 0, surprise: 0 };
+    for (const block of blocks) {
+      for (const sentence of block) {
+        const cs = sentence.emotion.cleanScores ?? {};
+        for (const k of EMOTION_CATS) counts[k] += cs[k] ?? 0;
+      }
+    }
+    perChapter = [{ chapter: "全书", counts }];
+  }
+  return {
+    stats,
+    implicit,
+    compare: explicitImplicitCompare(meanValence, implicit),
+    complexity: emotionComplexity(perChapter),
+    composites: compositeEmotionPairs(blocks)
+  };
+}
+
+/**
+ * 全书/单章句式模式分析主入口（v1.6.0 合并版）。
  * @param text 正文文本。
  * @param options { top: 句式模板条数（默认 8）, maxSentences: 采样上限（默认 20000）, chapterTexts: [{chapter, text}] 可选分章输入 }
  * @returns 结构化分析结果（与 novel_sentence_analysis 输出 schema 一致）。
@@ -633,7 +922,7 @@ export function analyzeText(text, options = {}) {
       emotion: emotionName,
       label: EMOTION_LABELS[emotionName],
       count: round(emotionCounts[emotionName], 2),
-      words: EMOTION_WORDS_UNIQUE[emotionName].filter((word) => emotionWordCounts.has(word)).slice(0, 10)
+      words: [...new Set(EMOTION_WORDS[emotionName])].filter((word) => emotionWordCounts.has(word)).slice(0, 10)
     })),
     cleanScores: ["joy", "anger", "sorrow", "fear", "surprise"].map((emotionName) => ({
       emotion: emotionName,
@@ -651,7 +940,9 @@ export function analyzeText(text, options = {}) {
       })
       .sort((a, b) => b.count - a.count || a.word.localeCompare(b.word))
       .slice(0, 12),
-    curve: emotionCurve(blockMeta, curveSegments)
+    curve: emotionCurve(blockMeta, curveSegments),
+    // v1.6.0：情感量化（Valence 三指标 + 显隐对比 + 复杂度 + 复合共现）
+    quantification: emotionalQuantification(text, [], blockMeta.map((m) => m.sentences).filter((s) => s.length > 0))
   };
 
   // 主观性指数（启发式 0-100）