feat: 建立 DOCX 标准样式探针

docs/PROGRESS.md

@@ -153,12 +153,15 @@ DOCX 样式令牌，支持自动映射、清单覆盖和预设降级；Front Mat
 DOCX 映射。
 
 `v0.6.0` 阶段 11 的协议骨架已建立为独立
-`packages/docx-theme-engine`：定义 50 个稳定语义样式槽位、浏览器计算
+`packages/docx-theme-engine`：定义 56 个稳定语义样式槽位、浏览器计算
 样式快照、Word 目标样式令牌、属性来源、映射置信度和诊断协议。
 `docxStyle` 支持 `auto`、`auto-with-overrides`、`explicit` 三种模式，
 普通外部主题缺少 DOCX 声明时默认进入自动映射；现有主题的 `preset` 和
-扁平样式覆盖保持兼容。当前只建立公共契约与 workspace 构建边界，尚未
+扁平样式覆盖保持兼容。标准探针 DOM 已覆盖普通文档、公文、简报、项目
-实现样式探针页面、Chromium/Electron 采集器或接入 Pandoc 转换链路。
+报告和标书，Profile 专属槽位避免同名类在不同结构上下文中互相污染；
+平台无关采集器可以按稳定顺序读取全部 `getComputedStyle()` 并保留未
+命中状态。当前尚未实现 Playwright/Electron 页面适配器，也未将快照
+归一化为 Word 令牌或接入 Pandoc 转换链路。
 
 ## 2. 已完成
 
@@ -1085,8 +1088,8 @@ ECharts 第二阶段浏览器与 PDF 验证结果：
 - 阶段 10：已完成 Word 原生页眉页脚和第一轮全主题视觉审计；完整主题
   映射门禁未通过；
 - 阶段 11：已建立独立 `packages/docx-theme-engine`、标准语义槽位及
-  快照、令牌、来源、置信度和诊断协议；下一步实现样式探针与
+  快照、令牌、来源、置信度、诊断协议、标准探针 DOM 和平台无关采集器；
-  Chromium/Electron 计算样式采集；
+  下一步实现 Playwright/Electron 计算样式页面适配；
 - 阶段 12：建立统一语义文档模型，完成 Front Matter、封面、分节、
   表格宽度和分页控制；
 - 阶段 13：完成 Word/WPS 双向互存、外部主题兼容、体积和正式发布验收。

packages/docx-theme-engine/README.md

@@ -5,8 +5,10 @@ DOCX 主题映射引擎负责在浏览器主题 CSS 与 Word 样式之间建立
 
 当前阶段包含：
 
-- 标准 Markdown 与结构化文档语义槽位；
+- 56 个标准 Markdown 与结构化文档语义槽位；
+- 覆盖普通文档、公文、简报、项目报告和标书的标准探针 DOM；
 - Chromium/Electron 计算样式快照协议；
+- 不绑定具体浏览器实现的计算样式采集器；
 - Word 目标样式令牌；
 - 自动映射、显式覆盖和预设降级配置；
 - 样式来源、映射置信度和诊断协议；
@@ -20,6 +22,7 @@ DOCX 主题映射引擎负责在浏览器主题 CSS 与 Word 样式之间建立
 - 直接修改 OOXML；
 - 按主题 ID 维护专属转换分支。
 
-后续 Chromium 与 Electron 适配器只负责采集标准样式探针的
+后续 Chromium 与 Electron 适配器只负责挂载探针、应用主题 CSS，并将
-`getComputedStyle()` 结果。本包负责将快照归一化为 DOCX 样式令牌，
+平台的 `getComputedStyle()` 接入本包采集器。本包负责将快照归一化为
-`@md-to-pdf/docx-engine` 再消费令牌生成 `reference.docx` 和最终 OOXML。
+DOCX 样式令牌，`@md-to-pdf/docx-engine` 再消费令牌生成
+`reference.docx` 和最终 OOXML。

packages/docx-theme-engine/src/collector.ts

@@ -0,0 +1,104 @@
+import {
+  docxThemeStyleSnapshotSchema,
+  type DocxComputedStyle,
+  type DocxStyleSlotSnapshot,
+  type DocxThemeStyleSnapshot
+} from "./snapshot.js";
+import { DOCX_STYLE_SLOTS } from "./slots.js";
+
+export interface DocxStyleProbeRoot {
+  querySelector(selector: string): unknown | null;
+}
+
+export type DocxComputedStyleReader = (
+  element: unknown
+) => Readonly<DocxComputedStyle>;
+
+export type DocxComputedStyleDeclaration = {
+  readonly [Property in keyof DocxComputedStyle]: string;
+};
+
+export interface CollectDocxThemeStyleSnapshotOptions {
+  root: DocxStyleProbeRoot;
+  readComputedStyle: DocxComputedStyleReader;
+  themeId: string;
+  themeFingerprint: string;
+  viewport: {
+    widthPx: number;
+    heightPx: number;
+    deviceScaleFactor: number;
+  };
+  rootFontSizePx: number;
+}
+
+export function readDocxComputedStyle(
+  style: DocxComputedStyleDeclaration
+): DocxComputedStyle {
+  return {
+    fontFamily: style.fontFamily,
+    fontSize: style.fontSize,
+    fontWeight: style.fontWeight,
+    fontStyle: style.fontStyle,
+    color: style.color,
+    backgroundColor: style.backgroundColor,
+    lineHeight: style.lineHeight,
+    letterSpacing: style.letterSpacing,
+    textAlign: style.textAlign,
+    textIndent: style.textIndent,
+    textDecorationLine: style.textDecorationLine,
+    marginTop: style.marginTop,
+    marginRight: style.marginRight,
+    marginBottom: style.marginBottom,
+    marginLeft: style.marginLeft,
+    paddingTop: style.paddingTop,
+    paddingRight: style.paddingRight,
+    paddingBottom: style.paddingBottom,
+    paddingLeft: style.paddingLeft,
+    borderTop: style.borderTop,
+    borderRight: style.borderRight,
+    borderBottom: style.borderBottom,
+    borderLeft: style.borderLeft,
+    width: style.width,
+    maxWidth: style.maxWidth,
+    breakBefore: style.breakBefore,
+    breakAfter: style.breakAfter,
+    breakInside: style.breakInside,
+    display: style.display
+  };
+}
+
+export function collectDocxStyleSlotSnapshots(
+  root: DocxStyleProbeRoot,
+  readComputedStyle: DocxComputedStyleReader
+): DocxStyleSlotSnapshot[] {
+  return DOCX_STYLE_SLOTS.map((definition) => {
+    const element = root.querySelector(definition.selector);
+    if (!element) {
+      return {
+        slot: definition.name,
+        matched: false
+      };
+    }
+    return {
+      slot: definition.name,
+      matched: true,
+      computed: { ...readComputedStyle(element) }
+    };
+  });
+}
+
+export function collectDocxThemeStyleSnapshot(
+  options: CollectDocxThemeStyleSnapshotOptions
+): DocxThemeStyleSnapshot {
+  return docxThemeStyleSnapshotSchema.parse({
+    schemaVersion: 1,
+    themeId: options.themeId,
+    themeFingerprint: options.themeFingerprint,
+    viewport: options.viewport,
+    rootFontSizePx: options.rootFontSizePx,
+    slots: collectDocxStyleSlotSnapshots(
+      options.root,
+      options.readComputedStyle
+    )
+  });
+}

packages/docx-theme-engine/src/index.ts

@@ -1,4 +1,6 @@
+export * from "./collector.js";
 export * from "./configuration.js";
+export * from "./probe.js";
 export * from "./slots.js";
 export * from "./snapshot.js";
 export * from "./tokens.js";

packages/docx-theme-engine/src/probe.ts

@@ -0,0 +1,132 @@
+import {
+  DOCX_STYLE_SLOT_NAMES,
+  type DocxStyleSlotName
+} from "./slots.js";
+
+function attribute(slot: DocxStyleSlotName) {
+  return `data-docx-slot="${slot}"`;
+}
+
+const transparentPixel =
+  "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNk+A8AAQUBAScY42YAAAAASUVORK5CYII=";
+
+export function createDocxStyleProbeMarkup(): string {
+  return `
+<article id="write" ${attribute("document")}>
+  <header class="doc-metadata">
+    <h1 class="doc-metadata-title" ${attribute("document-title")}>文档标题</h1>
+    <p class="doc-author" ${attribute("document-author")}>示例作者</p>
+  </header>
+  <section class="docx-probe-markdown">
+    <h1 ${attribute("heading-1")}>一级标题</h1>
+    <h2 ${attribute("heading-2")}>二级标题</h2>
+    <h3 ${attribute("heading-3")}>三级标题</h3>
+    <h4 ${attribute("heading-4")}>四级标题</h4>
+    <h5 ${attribute("heading-5")}>五级标题</h5>
+    <h6 ${attribute("heading-6")}>六级标题</h6>
+    <p ${attribute("paragraph")}>
+      正文
+      <strong ${attribute("strong")}>粗体</strong>
+      <em ${attribute("emphasis")}>斜体</em>
+      <del ${attribute("strikethrough")}>删除线</del>
+      <code ${attribute("inline-code")}>inline()</code>
+      <a href="#probe" ${attribute("hyperlink")}>链接</a>
+    </p>
+    <ul ${attribute("unordered-list")}><li ${attribute("list-item")}>无序列表</li></ul>
+    <ol ${attribute("ordered-list")}><li>有序列表</li></ol>
+    <blockquote ${attribute("block-quote")}><p>引用内容</p></blockquote>
+    <pre class="md-fences" ${attribute("code-block")}><code>const value = 1;</code></pre>
+    <table ${attribute("table")}>
+      <thead><tr><th ${attribute("table-header")}>表头</th></tr></thead>
+      <tbody><tr><td ${attribute("table-cell")}>单元格</td></tr></tbody>
+    </table>
+    <figure ${attribute("figure")}>
+      <img src="${transparentPixel}" alt="样式探针" ${attribute("image")}>
+      <figcaption ${attribute("caption")}>图注</figcaption>
+    </figure>
+    <section class="footnotes" ${attribute("footnotes")}><p>脚注内容</p></section>
+  </section>
+  <section class="docx-probe-official">
+    <header class="doc-masthead doc-masthead-official" ${attribute("official-masthead")}>
+      <div class="doc-classification" ${attribute("official-classification")}>秘密　加急</div>
+      <p class="doc-issuer" ${attribute("official-issuer")}>示例发文机关</p>
+      <div class="doc-issue-row" ${attribute("official-issue-row")}>
+        <span class="doc-number" ${attribute("official-number")}>示例发〔2026〕1号</span>
+        <span class="doc-signatory" ${attribute("official-signatory")}><span>签发人：</span>张三</span>
+      </div>
+    </header>
+    <h1 class="doc-title" ${attribute("official-title")}>公文标题</h1>
+    <section class="doc-signature" ${attribute("official-signature")}><p>示例发文机关</p><time>2026年7月30日</time></section>
+    <footer class="doc-edition" ${attribute("official-edition")}><p class="doc-copy-to">抄送：示例单位</p><div class="doc-printing-row"><span>示例办公室</span><time>2026年7月30日印发</time></div></footer>
+  </section>
+  <section class="docx-probe-briefing">
+    <header class="doc-masthead doc-masthead-briefing" ${attribute("briefing-masthead")}>
+      <p class="doc-briefing-masthead">工作简报</p>
+      <div class="doc-briefing-meta" ${attribute("briefing-meta")}><span>第1期</span><span>示例单位</span><time>2026年7月30日</time></div>
+    </header>
+    <h1 class="doc-title" ${attribute("briefing-title")}>简报标题</h1>
+    <footer class="doc-briefing-contact" ${attribute("briefing-contact")}>联系人：示例</footer>
+  </section>
+  <header class="doc-cover doc-cover-project-report" ${attribute("project-report-cover")}>
+    <p class="doc-cover-project-name" ${attribute("project-report-project-name")}>示例建设项目</p>
+    <h1 class="doc-cover-title" ${attribute("project-report-title")}>可行性研究报告</h1>
+    <p class="doc-cover-version" ${attribute("project-report-version")}>送审稿 V1.0</p>
+    <p class="doc-cover-owner" ${attribute("project-report-owner")}><span>建设单位：</span>示例单位</p>
+    <p class="doc-cover-prepared-by" ${attribute("project-report-prepared-by")}><span>编制单位：</span>示例单位</p>
+    <time class="doc-cover-date" ${attribute("project-report-date")}>2026年7月</time>
+  </header>
+  <header class="doc-cover doc-cover-tender" ${attribute("tender-cover")}>
+    <p class="doc-cover-copy-mark" ${attribute("tender-copy-mark")}>正本</p>
+    <p class="doc-cover-project-name" ${attribute("tender-project-name")}>示例采购项目</p>
+    <p class="doc-cover-project-number" ${attribute("tender-project-number")}>ZB-2026-001</p>
+    <h1 class="doc-cover-title" ${attribute("tender-title")}>投标文件</h1>
+    <p class="doc-cover-volume" ${attribute("tender-volume")}>技术及商务文件</p>
+    <p class="doc-cover-bidder" ${attribute("tender-bidder")}><span>投标人：</span>示例公司</p>
+    <p class="doc-cover-representative" ${attribute("tender-representative")}><span>法定代表人或授权代表：</span>李四</p>
+    <time class="doc-cover-date" ${attribute("tender-date")}>2026年7月30日</time>
+  </header>
+</article>`.trim();
+}
+
+export function listDocxStyleProbeSlots(
+  markup = createDocxStyleProbeMarkup()
+): DocxStyleSlotName[] {
+  const matches = markup.matchAll(
+    /data-docx-slot="([^"]+)"/gu
+  );
+  return [...matches].map(
+    (match) => match[1] as DocxStyleSlotName
+  );
+}
+
+export function validateDocxStyleProbeMarkup(
+  markup = createDocxStyleProbeMarkup()
+): void {
+  const actual = listDocxStyleProbeSlots(markup);
+  const duplicates = actual.filter(
+    (name, index) => actual.indexOf(name) !== index
+  );
+  const actualSet = new Set(actual);
+  const missing = DOCX_STYLE_SLOT_NAMES.filter(
+    (name) => !actualSet.has(name)
+  );
+  const unexpected = actual.filter(
+    (name) =>
+      !DOCX_STYLE_SLOT_NAMES.includes(name as DocxStyleSlotName)
+  );
+  if (duplicates.length || missing.length || unexpected.length) {
+    throw new Error(
+      [
+        duplicates.length
+          ? `重复槽位：${[...new Set(duplicates)].join("、")}`
+          : "",
+        missing.length ? `缺少槽位：${missing.join("、")}` : "",
+        unexpected.length
+          ? `未知槽位：${[...new Set(unexpected)].join("、")}`
+          : ""
+      ]
+        .filter(Boolean)
+        .join("；")
+    );
+  }
+}

packages/docx-theme-engine/src/slots.ts

@@ -34,23 +34,29 @@ export const DOCX_STYLE_SLOT_NAMES = [
   "official-issue-row",
   "official-number",
   "official-signatory",
+  "official-title",
   "official-signature",
   "official-edition",
   "briefing-masthead",
   "briefing-meta",
+  "briefing-title",
   "briefing-contact",
-  "cover",
+  "project-report-cover",
-  "cover-copy-mark",
+  "project-report-project-name",
-  "cover-project-name",
+  "project-report-title",
-  "cover-project-number",
+  "project-report-version",
-  "cover-title",
+  "project-report-owner",
-  "cover-volume",
+  "project-report-prepared-by",
-  "cover-owner",
+  "project-report-date",
-  "cover-prepared-by",
+  "tender-cover",
-  "cover-bidder",
+  "tender-copy-mark",
-  "cover-representative",
+  "tender-project-name",
-  "cover-version",
+  "tender-project-number",
-  "cover-date"
+  "tender-title",
+  "tender-volume",
+  "tender-bidder",
+  "tender-representative",
+  "tender-date"
 ] as const;
 
 export const docxStyleSlotNameSchema = z.enum(
@@ -126,23 +132,29 @@ const kinds: Record<DocxStyleSlotName, DocxStyleSlotKind> = {
   "official-issue-row": "structure",
   "official-number": "structure",
   "official-signatory": "structure",
+  "official-title": "structure",
   "official-signature": "structure",
   "official-edition": "structure",
   "briefing-masthead": "structure",
   "briefing-meta": "structure",
+  "briefing-title": "structure",
   "briefing-contact": "structure",
-  cover: "structure",
+  "project-report-cover": "structure",
-  "cover-copy-mark": "structure",
+  "project-report-project-name": "structure",
-  "cover-project-name": "structure",
+  "project-report-title": "structure",
-  "cover-project-number": "structure",
+  "project-report-version": "structure",
-  "cover-title": "structure",
+  "project-report-owner": "structure",
-  "cover-volume": "structure",
+  "project-report-prepared-by": "structure",
-  "cover-owner": "structure",
+  "project-report-date": "structure",
-  "cover-prepared-by": "structure",
+  "tender-cover": "structure",
-  "cover-bidder": "structure",
+  "tender-copy-mark": "structure",
-  "cover-representative": "structure",
+  "tender-project-name": "structure",
-  "cover-version": "structure",
+  "tender-project-number": "structure",
-  "cover-date": "structure"
+  "tender-title": "structure",
+  "tender-volume": "structure",
+  "tender-bidder": "structure",
+  "tender-representative": "structure",
+  "tender-date": "structure"
 };
 
 export const DOCX_STYLE_SLOTS: readonly DocxStyleSlotDefinition[] =

packages/docx-theme-engine/tests/probe.test.ts

@@ -0,0 +1,122 @@
+import { describe, expect, it } from "vitest";
+import {
+  DOCX_STYLE_SLOT_NAMES,
+  collectDocxStyleSlotSnapshots,
+  collectDocxThemeStyleSnapshot,
+  createDocxStyleProbeMarkup,
+  listDocxStyleProbeSlots,
+  readDocxComputedStyle,
+  validateDocxStyleProbeMarkup,
+  type DocxComputedStyle,
+  type DocxStyleProbeRoot
+} from "../src/index.js";
+
+const computedStyle: DocxComputedStyle = {
+  fontFamily: '"Microsoft YaHei", sans-serif',
+  fontSize: "16px",
+  fontWeight: "400",
+  fontStyle: "normal",
+  color: "rgb(17, 17, 17)",
+  backgroundColor: "rgba(0, 0, 0, 0)",
+  lineHeight: "28px",
+  letterSpacing: "normal",
+  textAlign: "start",
+  textIndent: "32px",
+  textDecorationLine: "none",
+  marginTop: "0px",
+  marginRight: "0px",
+  marginBottom: "12px",
+  marginLeft: "0px",
+  paddingTop: "0px",
+  paddingRight: "0px",
+  paddingBottom: "0px",
+  paddingLeft: "0px",
+  borderTop: "0px none rgb(17, 17, 17)",
+  borderRight: "0px none rgb(17, 17, 17)",
+  borderBottom: "0px none rgb(17, 17, 17)",
+  borderLeft: "0px none rgb(17, 17, 17)",
+  width: "640px",
+  maxWidth: "none",
+  breakBefore: "auto",
+  breakAfter: "auto",
+  breakInside: "auto",
+  display: "block"
+};
+
+function createRoot(matchedSlots: readonly string[]): DocxStyleProbeRoot {
+  const selectors = new Set(
+    matchedSlots.map(
+      (slot) => `[data-docx-slot="${slot}"]`
+    )
+  );
+  return {
+    querySelector(selector) {
+      return selectors.has(selector) ? { selector } : null;
+    }
+  };
+}
+
+describe("DOCX 标准样式探针", () => {
+  it("每个标准槽位在探针中恰好出现一次", () => {
+    const markup = createDocxStyleProbeMarkup();
+    expect(() => validateDocxStyleProbeMarkup(markup)).not.toThrow();
+    const slots = listDocxStyleProbeSlots(markup);
+    expect(slots).toHaveLength(DOCX_STYLE_SLOT_NAMES.length);
+    expect(new Set(slots)).toEqual(
+      new Set(DOCX_STYLE_SLOT_NAMES)
+    );
+  });
+
+  it("拒绝缺失和重复槽位的探针", () => {
+    const markup = createDocxStyleProbeMarkup()
+      .replace('data-docx-slot="paragraph"', "")
+      .replace(
+        "</article>",
+        '<p data-docx-slot="strong">重复</p></article>'
+      );
+    expect(() => validateDocxStyleProbeMarkup(markup)).toThrow(
+      /重复槽位：strong；缺少槽位：paragraph/u
+    );
+  });
+});
+
+describe("DOCX 计算样式采集", () => {
+  it("按稳定槽位顺序采集并保留未命中状态", () => {
+    const slots = collectDocxStyleSlotSnapshots(
+      createRoot(["document", "paragraph"]),
+      () => readDocxComputedStyle(computedStyle)
+    );
+    expect(slots).toHaveLength(DOCX_STYLE_SLOT_NAMES.length);
+    expect(slots[0]).toMatchObject({
+      slot: "document",
+      matched: true,
+      computed: { fontSize: "16px" }
+    });
+    expect(
+      slots.find((entry) => entry.slot === "paragraph")
+    ).toMatchObject({ matched: true });
+    expect(
+      slots.find((entry) => entry.slot === "heading-1")
+    ).toEqual({ slot: "heading-1", matched: false });
+  });
+
+  it("生成可校验且与平台实现无关的主题快照", () => {
+    const snapshot = collectDocxThemeStyleSnapshot({
+      root: createRoot(DOCX_STYLE_SLOT_NAMES),
+      readComputedStyle: () =>
+        readDocxComputedStyle(computedStyle),
+      themeId: "external-clean",
+      themeFingerprint: "c".repeat(64),
+      viewport: {
+        widthPx: 794,
+        heightPx: 1123,
+        deviceScaleFactor: 1
+      },
+      rootFontSizePx: 16
+    });
+    expect(snapshot.slots).toHaveLength(
+      DOCX_STYLE_SLOT_NAMES.length
+    );
+    expect(snapshot.slots.every((slot) => slot.matched)).toBe(true);
+  });
+});