feat: 将主题令牌接入 DOCX 动态模板

This commit is contained in:
SkyJourney
2026-07-31 08:43:26 +08:00
parent 459a079f3b
commit 6e40374200
33 changed files with 1478 additions and 56 deletions
@@ -3,6 +3,7 @@ import http from "node:http";
import path from "node:path";
import { fileURLToPath } from "node:url";
import { createApplicationService } from "@md-to-pdf/application";
import { defaultExportConfig } from "@md-to-pdf/core";
import {
DOCX_STYLE_SLOT_NAMES,
createDocxThemeStyleCaptureScript,
@@ -120,8 +121,17 @@ try {
const tokenSets = [];
for (const theme of bundledThemes) {
console.error(`[Electron DOCX theme styles] capturing ${theme.id}`);
const themeCss = await application.getThemeCss(theme.id);
assert(themeCss !== undefined, `主题 ${theme.id} 缺少 CSS`);
const prepared = await application.prepareDocxExport({
markdown: `# ${theme.name}`,
fileName: `${theme.id}.md`,
language: "zh-CN",
exportConfig: {
...defaultExportConfig,
themeId: theme.id
}
});
const themeCss = prepared.theme.css;
const manifest = prepared.theme.manifest;
const themeFingerprint =
await createDocxThemeStyleFingerprint(theme.id, themeCss);
const request = {
@@ -150,7 +160,7 @@ try {
tokenSets.push(
resolveDocxThemeTokens({
snapshot,
config: normalizeDocxThemeMappingConfig(theme)
config: normalizeDocxThemeMappingConfig(manifest)
})
);
}
@@ -17,6 +17,7 @@ import {
import path from "node:path";
import {
DocxExportService,
DocxThemeTokenService,
MAXIMUM_MARKDOWN_LENGTH,
readDocxExportRuntimeLimits,
type ApplicationService
@@ -159,6 +160,9 @@ export class DesktopApplicationController {
const pandocRuntime = new PandocRuntime({
desktopResourcesPath: options.desktopResourcesPath
});
this.#docxMediaEngine = new ElectronDocxMediaEngine({
renderUrl: options.docxRenderUrl
});
this.#docxExportService = new DocxExportService({
application: this.#applicationService,
runtime: pandocRuntime,
@@ -169,11 +173,11 @@ export class DesktopApplicationController {
import.meta.url
)
}),
themeTokens: new DocxThemeTokenService(
this.#docxMediaEngine
),
limits: readDocxExportRuntimeLimits()
});
this.#docxMediaEngine = new ElectronDocxMediaEngine({
renderUrl: options.docxRenderUrl
});
}
async initialize() {
@@ -21,6 +21,7 @@ export interface ElectronDocxMediaEngineOptions {
export class ElectronDocxMediaEngine
implements DocxMediaCaptureAdapter, DocxThemeStyleCaptureAdapter
{
readonly themeStyleBaseUrl: string;
private readonly renderUrl: string;
private windowPromise: Promise<BrowserWindow> | undefined;
private queue = Promise.resolve();
@@ -28,6 +29,7 @@ export class ElectronDocxMediaEngine
constructor(options: ElectronDocxMediaEngineOptions) {
this.renderUrl = options.renderUrl;
this.themeStyleBaseUrl = this.renderUrl;
}
capture(
@@ -3,12 +3,22 @@ import http from "node:http";
import path from "node:path";
import { fileURLToPath } from "node:url";
import { createApplicationService } from "@md-to-pdf/application";
import {
DOCX_PANDOC_VERSION,
defaultExportConfig
} from "@md-to-pdf/core";
import {
DOCX_STYLE_SLOT_NAMES,
createDocxThemeStyleFingerprint,
normalizeDocxThemeMappingConfig,
resolveDocxThemeTokens
} from "@md-to-pdf/docx-theme-engine";
import {
PandocRuntime,
createDynamicReferenceDocx,
readReferenceDocxPackage,
resolveTokenFonts
} from "@md-to-pdf/docx-engine";
import { chromium } from "playwright";
import { captureDocxThemeStyleWithPlaywrightPage } from "../dist/playwright-docx-theme-style.js";
@@ -19,6 +29,10 @@ const outputDirectory = path.join(
"output",
"docx-theme-styles"
);
const referenceDirectory = path.join(
outputDirectory,
"references"
);
function assert(condition, message) {
if (!condition) {
@@ -26,6 +40,81 @@ function assert(condition, message) {
}
}
function styleBlock(stylesXml, styleId) {
const escaped = styleId.replace(/[.*+?^${}()|[\]\\]/gu, "\\$&");
return stylesXml.match(
new RegExp(
`<w:style[^>]*w:styleId="${escaped}"[^>]*>[\\s\\S]*?<\\/w:style>`,
"u"
)
)?.[0];
}
function inspectReference(reference, tokens) {
const entries = readReferenceDocxPackage(reference.content).entries;
const stylesXml = new TextDecoder().decode(
entries.get("word/styles.xml")
);
const fontTableXml = new TextDecoder().decode(
entries.get("word/fontTable.xml")
);
const normal = styleBlock(stylesXml, "Normal");
const paragraph = tokens.slots.find(
(slot) => slot.slot === "paragraph"
)?.style;
assert(normal, `主题 ${tokens.themeId} 缺少 Normal 样式`);
const fonts = resolveTokenFonts(paragraph, {
latin: "Arial",
eastAsia: "SimSun"
});
if (paragraph?.fontCandidates.length) {
assert(
normal.includes(`w:ascii="${fonts.latin}"`),
`主题 ${tokens.themeId} 的 Latin 正文字体未写入模板`
);
assert(
normal.includes(`w:eastAsia="${fonts.eastAsia}"`),
`主题 ${tokens.themeId} 的东亚正文字体未写入模板`
);
assert(
fontTableXml.includes(`w:name="${fonts.eastAsia}"`),
`主题 ${tokens.themeId} 的东亚字体未写入字体表`
);
}
if (paragraph?.fontSizePt !== undefined) {
assert(
normal.includes(
`w:sz w:val="${Math.round(paragraph.fontSizePt * 2)}"`
),
`主题 ${tokens.themeId} 的正文字号未写入模板`
);
}
for (const styleId of [
"Heading1",
"SourceCode",
"Table",
"MdOfficialTitle",
"MdBriefingTitle",
"MdProjectReportTitle",
"MdTenderTitle"
]) {
assert(
stylesXml.includes(`w:styleId="${styleId}"`),
`主题 ${tokens.themeId} 缺少样式 ${styleId}`
);
}
return {
bytes: reference.content.byteLength,
templateFingerprint: reference.templateFingerprint,
cacheKey: reference.cacheKey,
stylePreset: reference.stylePreset,
partCount: reference.partCount,
paragraphFonts: fonts,
customStyleCount:
stylesXml.match(/w:styleId="Md[A-Za-z]+"/gu)?.length ?? 0
};
}
const application = createApplicationService({
bundledRoot: path.join(repositoryDirectory, "themes"),
localRoot: path.join(repositoryDirectory, ".local", "themes")
@@ -88,6 +177,20 @@ if (!address || typeof address === "string") {
}
const baseUrl = `http://127.0.0.1:${address.port}/`;
const browser = await chromium.launch({ headless: true });
const runtime = new PandocRuntime(
process.env.DOCX_PANDOC_PATH?.trim()
? { configuredPath: process.env.DOCX_PANDOC_PATH.trim() }
: {}
);
const capability = await runtime.probe();
if (capability.capability.status !== "available") {
throw new Error(capability.capability.message);
}
assert(
capability.capability.detectedVersion === DOCX_PANDOC_VERSION,
`Pandoc 版本不匹配:期望 ${DOCX_PANDOC_VERSION},实际 ${capability.capability.detectedVersion}`
);
const baseline = await runtime.getDefaultReferenceDocx();
try {
const context = await browser.newContext({
@@ -104,10 +207,20 @@ try {
);
const snapshots = [];
const tokenSets = [];
const references = [];
for (const theme of bundledThemes) {
console.error(`[DOCX theme styles] capturing ${theme.id}`);
const themeCss = await application.getThemeCss(theme.id);
assert(themeCss !== undefined, `主题 ${theme.id} 缺少 CSS`);
const prepared = await application.prepareDocxExport({
markdown: `# ${theme.name}`,
fileName: `${theme.id}.md`,
language: "zh-CN",
exportConfig: {
...defaultExportConfig,
themeId: theme.id
}
});
const themeCss = prepared.theme.css;
const manifest = prepared.theme.manifest;
const themeFingerprint =
await createDocxThemeStyleFingerprint(theme.id, themeCss);
const snapshot =
@@ -129,12 +242,27 @@ try {
.join("、")}`
);
snapshots.push(snapshot);
tokenSets.push(
resolveDocxThemeTokens({
snapshot,
config: normalizeDocxThemeMappingConfig(theme)
})
const tokens = resolveDocxThemeTokens({
snapshot,
config: normalizeDocxThemeMappingConfig(manifest)
});
tokenSets.push(tokens);
const reference = createDynamicReferenceDocx(baseline, {
exportConfig: prepared.request.exportConfig,
theme: manifest,
fileName: prepared.request.fileName,
metadata: prepared.document.metadata,
themeTokens: tokens
});
fs.mkdirSync(referenceDirectory, { recursive: true });
fs.writeFileSync(
path.join(referenceDirectory, `${theme.id}.docx`),
reference.content
);
references.push({
themeId: theme.id,
...inspectReference(reference, tokens)
});
}
assert(
snapshots.length === 14,
@@ -172,7 +300,9 @@ try {
themeCount: snapshots.length,
slotCount: DOCX_STYLE_SLOT_NAMES.length,
snapshots,
tokenSets
tokenSets,
pandocVersion: capability.capability.detectedVersion,
references
},
null,
2
@@ -190,7 +320,9 @@ try {
diagnostics: tokenSets.reduce(
(count, tokens) => count + tokens.diagnostics.length,
0
)
),
pandocVersion: capability.capability.detectedVersion,
references: references.length
})
);
await context.close();
+7 -2
View File
@@ -14,6 +14,7 @@ import {
ApplicationRequestError,
DocxExportService,
DocxExportServiceError,
DocxThemeTokenService,
createApplicationService,
readDocxExportRuntimeLimits,
type ApplicationService
@@ -133,6 +134,7 @@ function createDocxServerTiming(
`queue;dur=${milliseconds(timings.queueMs)}`,
`runtime-probe;dur=${milliseconds(timings.probeMs)}`,
`prepare;dur=${milliseconds(timings.prepareMs)}`,
`theme-style;dur=${milliseconds(timings.themeStyleMs)}`,
`media;dur=${milliseconds(timings.mediaMs)}`,
`reference;dur=${milliseconds(timings.referenceMs)}`,
`pandoc;dur=${milliseconds(timings.pandocMs)}`,
@@ -166,6 +168,8 @@ export function buildApp(options: BuildAppOptions = {}) {
const pandocRuntime = options.docxExportService
? undefined
: new PandocRuntime();
const docxMediaAdapter =
options.docxMediaAdapter ?? createDocxMediaEngine();
const docxExportService =
options.docxExportService ??
new DocxExportService({
@@ -174,10 +178,11 @@ export function buildApp(options: BuildAppOptions = {}) {
converter: new PandocDocxConverter({
runtime: pandocRuntime!
}),
themeTokens: new DocxThemeTokenService(
docxMediaAdapter
),
limits: readDocxExportRuntimeLimits()
});
const docxMediaAdapter =
options.docxMediaAdapter ?? createDocxMediaEngine();
if (options.prewarmPdfBrowser) {
app.addHook("onReady", async () => {
+2
View File
@@ -38,6 +38,7 @@ async function closeContext(context: BrowserContext) {
export class PlaywrightDocxMediaEngine
implements ServerDocxMediaCaptureAdapter
{
readonly themeStyleBaseUrl: string;
private readonly renderOrigin: string;
private readonly renderUrl: string;
private readonly launchBrowser: () => Promise<Browser>;
@@ -54,6 +55,7 @@ export class PlaywrightDocxMediaEngine
"/preview-frame.html?target=continuous",
this.renderOrigin
).href;
this.themeStyleBaseUrl = this.renderUrl;
this.launchBrowser =
options.launchBrowser ??
(() =>
+12 -6
View File
@@ -42,11 +42,12 @@ function generatedDocx(): DocxExportResult {
queueMs: 1,
probeMs: 2,
prepareMs: 3,
mediaMs: 4,
referenceMs: 5,
pandocMs: 6,
validationMs: 7,
totalMs: 28
themeStyleMs: 4,
mediaMs: 5,
referenceMs: 6,
pandocMs: 7,
validationMs: 8,
totalMs: 36
}
};
}
@@ -72,6 +73,8 @@ function createDocxService(
function createMediaAdapter() {
return {
themeStyleBaseUrl:
"http://localhost:5173/preview-frame.html",
capture: vi.fn<
DocxMediaCaptureAdapter["capture"]
>(async () => ({
@@ -82,6 +85,9 @@ function createMediaAdapter() {
},
captures: []
})),
captureThemeStyle: vi.fn(async () => {
throw new Error("此测试使用注入的 DOCX 导出服务");
}),
close: vi.fn(async () => undefined)
} satisfies ServerDocxMediaCaptureAdapter;
}
@@ -173,7 +179,7 @@ describe("DOCX HTTP API", () => {
"runtime-probe;dur=2.0"
);
expect(response.headers["server-timing"]).toContain(
"pandoc;dur=6.0"
"pandoc;dur=7.0"
);
expect(response.rawPayload).toEqual(
Buffer.from([80, 75, 3, 4])
+1
View File
@@ -133,6 +133,7 @@ describe("DOCX 导出客户端", () => {
queueMs: 0,
probeMs: 0,
prepareMs: 0,
themeStyleMs: 0,
mediaMs: 0,
referenceMs: 0,
pandocMs: 0,
+16 -1
View File
@@ -193,6 +193,20 @@ Front Matter 和正文首个 H1 构建主题无关的语义树,再由 Renderer
`reference.docx` 或最终 DOCX;下一步阶段 12B 将把已归一化的主题令牌
接入动态 Word 模板。
`v0.6.0` 阶段 12B 已将主题令牌接入真实 DOCX 生产链。应用层现在复用
Server Playwright 与 Desktop Electron 适配器,按主题 CSS SHA-256
指纹缓存计算样式快照,并与媒体捕获并行生成 56 槽位令牌;`explicit`
模式直接使用清单覆盖与预设降级,无需启动浏览器。完整令牌进入 Pandoc
转换输入和动态模板缓存键,主题变化不会复用旧模板;诊断进入导出警告,
并新增 `themeStyleMs` 耗时。DOCX 引擎通过固定槽位映射写入标准 Markdown
样式和 31 个 `Md*` 结构样式,同时覆盖字体、字号、颜色、段落、代码、
引用、表格、图注、链接、字体表与 Office 主题字体,全程不按主题 ID
分支。Playwright Chromium 151 与 Electron Chromium 150 对 14 套主题
各采集 784 个槽位,完整令牌差异为 0;固定 Pandoc 3.9.0.2 的默认模板
成功生成并验证 14 份动态 `reference.docx`,每份均含 31 个结构样式。
当前结构样式尚未绑定到 Pandoc AST,封面、版头、分节与标题去重留给
阶段 12C。
## 2. 已完成
### 2.1 项目骨架
@@ -1124,7 +1138,8 @@ ECharts 第二阶段浏览器与 PDF 验证结果:
近似诊断已通过 14 主题双引擎矩阵;下一步由统一语义文档模型消费令牌;
- 阶段 12A:已建立独立统一语义文档模型,Renderer 与 DOCX 准备结果
共用同一语义树,现有 HTML DOM 保持兼容;
- 阶段 12B主题令牌接入动态 `reference.docx`
- 阶段 12B已将跨端主题令牌接入动态 `reference.docx`14 主题双引擎
令牌和真实 Pandoc 模板矩阵通过;
- 阶段 12C:完成 Pandoc 结构映射、封面分节、标题去重、表格宽度和
分页控制;
- 阶段 12D:完成 14 套主题的真实 Word/WPS 视觉与结构回归;
+2
View File
@@ -9916,6 +9916,7 @@
"dependencies": {
"@md-to-pdf/core": "0.1.0",
"@md-to-pdf/docx-engine": "0.1.0",
"@md-to-pdf/docx-theme-engine": "0.1.0",
"@md-to-pdf/renderer": "0.1.0"
},
"devDependencies": {
@@ -9938,6 +9939,7 @@
"version": "0.1.0",
"dependencies": {
"@md-to-pdf/core": "0.1.0",
"@md-to-pdf/docx-theme-engine": "0.1.0",
"@xmldom/xmldom": "0.9.10",
"fflate": "0.8.3"
},
+1 -1
View File
@@ -30,7 +30,7 @@
"verify:docx-conversion": "npm run build -w @md-to-pdf/core && npm run build -w @md-to-pdf/docx-engine && npm run verify:conversion -w @md-to-pdf/docx-engine",
"verify:docx-matrix": "npm run build -w @md-to-pdf/core && npm run build -w @md-to-pdf/docx-engine && npm run verify:acceptance -w @md-to-pdf/docx-engine",
"verify:docx-themes": "npm run build:web-runtime && npm run verify:themes -w @md-to-pdf/docx-engine",
"verify:docx-theme-styles": "npm run build -w @md-to-pdf/core && npm run build -w @md-to-pdf/semantic-document && npm run build -w @md-to-pdf/docx-theme-engine && npm run build -w @md-to-pdf/renderer && npm run build -w @md-to-pdf/application && npm run build -w @md-to-pdf/server && npm run verify:docx-theme-styles -w @md-to-pdf/server && npm run verify:docx-theme-styles -w @md-to-pdf/desktop",
"verify:docx-theme-styles": "npm run build -w @md-to-pdf/core && npm run build -w @md-to-pdf/semantic-document && npm run build -w @md-to-pdf/docx-theme-engine && npm run build -w @md-to-pdf/docx-engine && npm run build -w @md-to-pdf/renderer && npm run build -w @md-to-pdf/application && npm run build -w @md-to-pdf/server && npm run verify:docx-theme-styles -w @md-to-pdf/server && npm run verify:docx-theme-styles -w @md-to-pdf/desktop",
"verify:docx-acceptance": "npm run build:web-runtime && npm run build -w @md-to-pdf/server && npm run build -w @md-to-pdf/desktop && npm run verify:pandoc -w @md-to-pdf/docx-engine && npm run verify:conversion -w @md-to-pdf/docx-engine && npm run verify:acceptance -w @md-to-pdf/docx-engine && npm run verify:docx-http -w @md-to-pdf/server && npm run test -w @md-to-pdf/desktop -- desktop-docx-save.test.ts",
"verify:docx-http": "npm run build:web-runtime && npm run build -w @md-to-pdf/server && npm run verify:docx-http -w @md-to-pdf/server",
"test": "npm run test -w @md-to-pdf/markdown-echarts && npm run test -w @md-to-pdf/core && npm run build -w @md-to-pdf/core && npm run test -w @md-to-pdf/semantic-document && npm run build -w @md-to-pdf/semantic-document && npm run test -w @md-to-pdf/docx-theme-engine && npm run test -w @md-to-pdf/docx-engine && npm run test -w @md-to-pdf/renderer && npm run test -w @md-to-pdf/application && npm run build -w @md-to-pdf/application && npm run test -w @md-to-pdf/preview-engine && npm run build -w @md-to-pdf/preview-engine && npm run test -w @md-to-pdf/web && npm run test -w @md-to-pdf/server && npm run test -w @md-to-pdf/desktop",
+9 -3
View File
@@ -13,6 +13,7 @@ src/
application-service.ts 渲染、主题和资源用例入口
docx-export-service.ts DOCX 并发、超时、错误和完整导出编排
docx-media-service.ts DOCX 媒体尺寸计算、PNG 校验与清单组装
docx-theme-token-service.ts 主题指纹、快照缓存和 Word 令牌准备
image-resources.ts 本地、Base64 与受限远程图片处理
theme-registry.ts 内置及自定义主题扫描、校验与缓存
index.ts 公共导出入口
@@ -21,6 +22,7 @@ tests/
bundled-themes.test.ts
docx-export-service.test.ts
docx-media-service.test.ts
docx-theme-token-service.test.ts
image-resources.test.ts
```
@@ -56,10 +58,14 @@ Desktop 才会以 Markdown 所在目录为边界解析相对资源。
大小、总大小和媒体 ID。最终按文档顺序输出稳定的 `media-001.png`
清单;它不依赖 Playwright 或 Electron,平台代码只负责 Chromium 捕获。
`DocxExportService` 串联 capability 探测、请求准备、平台媒体捕获、
Pandoc 转换和最终 OOXML 校验。默认并发为 1、队列为 4、总超时为 90 秒;
`DocxThemeTokenService` 复用 Server Playwright 或 Desktop Electron
适配器采集主题计算样式,按主题 CSS 指纹缓存快照并生成 56 槽位令牌。
`explicit` 模式不启动浏览器,直接使用清单覆盖和预设降级。
`DocxExportService` 串联 capability 探测、请求准备、平台媒体捕获、主题
令牌准备、Pandoc 转换和最终 OOXML 校验。媒体与主题令牌并行准备。默认并发为 1、队列为 4、总超时为 90 秒;
可以通过 `DOCX_CONCURRENCY``DOCX_MAX_QUEUE``DOCX_TIMEOUT_MS`
配置。排队、探测、准备、媒体、模板、Pandoc、校验和总耗时使用共享协议
配置。排队、探测、准备、主题样式、媒体、模板、Pandoc、校验和总耗时使用共享协议
返回;关闭服务时会取消活动任务、拒绝排队任务并等待清理完成。
内置主题来自仓库 `themes/`,当前名称为 Typora Github、
+1
View File
@@ -23,6 +23,7 @@
"dependencies": {
"@md-to-pdf/core": "0.1.0",
"@md-to-pdf/docx-engine": "0.1.0",
"@md-to-pdf/docx-theme-engine": "0.1.0",
"@md-to-pdf/renderer": "0.1.0"
},
"devDependencies": {
@@ -20,6 +20,10 @@ import {
prepareDocxMedia,
type DocxMediaCaptureAdapter
} from "./docx-media-service.js";
import type {
DocxThemeTokenProvider,
PreparedDocxThemeTokens
} from "./docx-theme-token-service.js";
import type { ImageResolutionContext } from "./image-resources.js";
const DEFAULT_DOCX_CONCURRENCY = 1;
@@ -47,6 +51,7 @@ export interface DocxExportServiceOptions {
application: Pick<ApplicationService, "prepareDocxExport">;
runtime: DocxCapabilityProvider;
converter: DocxConversionPort;
themeTokens: DocxThemeTokenProvider;
limits?: Partial<DocxExportRuntimeLimits>;
}
@@ -359,24 +364,51 @@ export class DocxExportService {
const prepareMs = performance.now() - prepareStarted;
controller.signal.throwIfAborted();
const mediaStarted = performance.now();
const media = await prepareDocxMedia(
prepared,
options.mediaAdapter,
controller.signal
);
const mediaMs = performance.now() - mediaStarted;
let themeStyleMs = 0;
let mediaMs = 0;
const [themeTokens, media] = await Promise.all([
(async () => {
const started = performance.now();
try {
return await this.options.themeTokens.prepare(
prepared,
controller.signal
);
} finally {
themeStyleMs = performance.now() - started;
}
})(),
(async () => {
const started = performance.now();
try {
return await prepareDocxMedia(
prepared,
options.mediaAdapter,
controller.signal
);
} finally {
mediaMs = performance.now() - started;
}
})()
]);
controller.signal.throwIfAborted();
const conversion = await this.options.converter.convert(
this.createConversionInput(prepared, media),
this.createConversionInput(
prepared,
themeTokens,
media
),
controller.signal
);
return {
docx: conversion.docx,
fileName: createDocxFileName(prepared.request.fileName),
diagnostics: {
warnings: media.warnings,
warnings: [
...media.warnings,
...themeTokens.warnings
],
echartsErrors: media.echartsErrors,
mermaidErrors: media.mermaidErrors
},
@@ -384,6 +416,7 @@ export class DocxExportService {
queueMs,
probeMs,
prepareMs,
themeStyleMs,
mediaMs,
referenceMs: conversion.timings.referenceMs,
pandocMs: conversion.timings.pandocMs,
@@ -446,6 +479,7 @@ export class DocxExportService {
private createConversionInput(
prepared: PreparedDocxExport,
themeTokens: PreparedDocxThemeTokens,
media: Awaited<ReturnType<typeof prepareDocxMedia>>
): PandocDocxConversionInput {
return {
@@ -454,6 +488,7 @@ export class DocxExportService {
language: prepared.request.language,
exportConfig: prepared.request.exportConfig,
theme: prepared.theme.manifest,
themeTokens: themeTokens.tokens,
metadata: prepared.document.metadata,
media
};
@@ -0,0 +1,105 @@
import type {
DocxThemeTokenSet,
DocxThemeStyleCaptureAdapter
} from "@md-to-pdf/docx-theme-engine";
import {
DOCX_STYLE_SLOTS,
DocxThemeStyleSnapshotCache,
createDocxThemeStyleFingerprint,
normalizeDocxThemeMappingConfig,
resolveDocxThemeTokens,
type DocxThemeStyleSnapshot
} from "@md-to-pdf/docx-theme-engine";
import type { PreparedDocxExport } from "./application-service.js";
export interface PreparedDocxThemeTokens {
tokens: DocxThemeTokenSet;
warnings: string[];
}
export interface DocxThemeTokenProvider {
prepare(
prepared: PreparedDocxExport,
signal?: AbortSignal
): Promise<PreparedDocxThemeTokens>;
}
function createExplicitSnapshot(
themeId: string,
themeFingerprint: string
): DocxThemeStyleSnapshot {
return {
schemaVersion: 1,
themeId,
themeFingerprint,
viewport: {
widthPx: 794,
heightPx: 1123,
deviceScaleFactor: 1
},
rootFontSizePx: 16,
slots: DOCX_STYLE_SLOTS.map(({ name }) => ({
slot: name,
matched: false
}))
};
}
export class DocxThemeTokenService
implements DocxThemeTokenProvider
{
private readonly snapshots: DocxThemeStyleSnapshotCache;
constructor(
private readonly adapter: DocxThemeStyleCaptureAdapter
) {
this.snapshots = new DocxThemeStyleSnapshotCache(adapter);
}
async prepare(
prepared: PreparedDocxExport,
signal?: AbortSignal
): Promise<PreparedDocxThemeTokens> {
signal?.throwIfAborted();
const config = normalizeDocxThemeMappingConfig(
prepared.theme.manifest
);
const themeFingerprint =
await createDocxThemeStyleFingerprint(
prepared.theme.manifest.id,
prepared.theme.css
);
signal?.throwIfAborted();
const snapshot =
config.mode === "explicit"
? createExplicitSnapshot(
prepared.theme.manifest.id,
themeFingerprint
)
: await this.snapshots.capture(
{
themeId: prepared.theme.manifest.id,
themeFingerprint,
themeCss: prepared.theme.css,
baseUrl: this.adapter.themeStyleBaseUrl
},
signal
);
const tokens = resolveDocxThemeTokens({
snapshot,
config
});
return {
tokens,
warnings: tokens.diagnostics
.filter(
(diagnostic) => diagnostic.severity !== "info"
)
.map((diagnostic) => diagnostic.message)
};
}
invalidate(themeId?: string) {
this.snapshots.invalidate(themeId);
}
}
+5
View File
@@ -30,6 +30,11 @@ export {
type DocxMediaCaptureRequest,
type DocxMediaRenderDimensions
} from "./docx-media-service.js";
export {
DocxThemeTokenService,
type DocxThemeTokenProvider,
type PreparedDocxThemeTokens
} from "./docx-theme-token-service.js";
export {
DocxExportService,
DocxExportServiceError,
@@ -10,6 +10,7 @@ import type {
PandocDocxConversionResult,
PandocRuntimeResolution
} from "@md-to-pdf/docx-engine";
import type { DocxThemeTokenSet } from "@md-to-pdf/docx-theme-engine";
import {
DocxExportService,
readDocxExportRuntimeLimits,
@@ -84,6 +85,16 @@ const availableCapability: DocxCapability = {
detectedVersion: "3.9.0.2"
};
const themeTokens: DocxThemeTokenSet = {
schemaVersion: 1,
themeId: "test-theme",
themeFingerprint: "c".repeat(64),
mode: "auto-with-overrides",
basePreset: "technical",
slots: [],
diagnostics: []
};
function resolution(
capability: DocxCapability = availableCapability
): PandocRuntimeResolution {
@@ -151,13 +162,22 @@ function createService(options: {
options.convert ??
vi.fn().mockResolvedValue(conversionResult())
},
themeTokens: {
prepare: vi.fn().mockResolvedValue({
tokens: themeTokens,
warnings: []
})
},
limits: options.limits
});
}
describe("DOCX 共享导出服务", () => {
it("串联准备、媒体和转换并返回完整耗时", async () => {
const service = createService({});
const convert = vi
.fn()
.mockResolvedValue(conversionResult());
const service = createService({ convert });
const result = await service.generate(
{ markdown: "# 测试" },
{ mediaAdapter: emptyAdapter }
@@ -170,7 +190,14 @@ describe("DOCX 共享导出服务", () => {
echartsErrors: [],
mermaidErrors: []
});
expect(convert).toHaveBeenCalledWith(
expect.objectContaining({
themeTokens
}),
expect.any(AbortSignal)
);
expect(result.timings).toMatchObject({
themeStyleMs: expect.any(Number),
referenceMs: 3,
pandocMs: 4,
validationMs: 5
@@ -0,0 +1,142 @@
import { describe, expect, it, vi } from "vitest";
import {
defaultExportConfig,
type ThemeManifest
} from "@md-to-pdf/core";
import {
DOCX_STYLE_SLOTS,
type DocxThemeStyleCaptureAdapter
} from "@md-to-pdf/docx-theme-engine";
import {
DocxThemeTokenService,
type PreparedDocxExport
} from "../src/index.js";
function manifest(
docxStyle: ThemeManifest["docxStyle"] = {
mode: "auto",
basePreset: "technical"
}
): ThemeManifest {
return {
manifestVersion: 1,
id: "test-theme",
name: "测试主题",
version: "1.0.0",
description: "测试",
author: "测试",
license: "内部许可",
entry: "theme.css",
domPreset: "generic",
defaultFontSize: "16px",
supportedFeatures: [],
category: "general",
compatibleProfiles: [],
docxStyle,
bundled: true
};
}
function prepared(
docxStyle?: ThemeManifest["docxStyle"]
): PreparedDocxExport {
return {
request: {
markdown: "# 测试",
fileName: "测试.md",
language: "zh-CN",
resources: [],
exportConfig: {
...defaultExportConfig,
themeId: "test-theme"
}
},
document: {
rendererVersion: 1,
articleHtml:
'<article id="write"><h1>测试</h1></article>',
bodyHtml: "<h1>测试</h1>",
metadata: {
title: "测试",
author: "",
subject: "",
keywords: [],
language: "zh-CN"
},
semanticDocument: {
schemaVersion: 1,
titlePolicy: {
metadataTitle: "suppress",
firstBodyHeading: "keep"
},
regions: []
},
features: [],
warnings: []
},
theme: {
manifest: manifest(docxStyle),
source: "bundled",
css: "#write { font-family: Test; }"
}
};
}
function adapter(): DocxThemeStyleCaptureAdapter {
return {
themeStyleBaseUrl:
"http://localhost:5173/preview-frame.html",
captureThemeStyle: vi.fn(async (request) => ({
schemaVersion: 1,
themeId: request.themeId,
themeFingerprint: request.themeFingerprint,
viewport: {
widthPx: 794,
heightPx: 1123,
deviceScaleFactor: 1
},
rootFontSizePx: 16,
slots: DOCX_STYLE_SLOTS.map(({ name }) => ({
slot: name,
matched: false
}))
}))
};
}
describe("DOCX 主题令牌准备服务", () => {
it("按主题指纹缓存跨端样式快照", async () => {
const captureAdapter = adapter();
const service = new DocxThemeTokenService(captureAdapter);
const first = await service.prepare(prepared());
const second = await service.prepare(prepared());
expect(captureAdapter.captureThemeStyle).toHaveBeenCalledTimes(1);
expect(first.tokens).toEqual(second.tokens);
expect(first.tokens.themeId).toBe("test-theme");
expect(first.tokens.themeFingerprint).toMatch(/^[a-f0-9]{64}$/u);
expect(first.tokens.slots).toHaveLength(DOCX_STYLE_SLOTS.length);
});
it("显式模式跳过浏览器采集并只产生信息级降级诊断", async () => {
const captureAdapter = adapter();
const service = new DocxThemeTokenService(captureAdapter);
const result = await service.prepare(
prepared({
mode: "explicit",
basePreset: "technical"
})
);
expect(captureAdapter.captureThemeStyle).not.toHaveBeenCalled();
expect(result.tokens.mode).toBe("explicit");
expect(result.warnings).toEqual([]);
expect(
result.tokens.diagnostics.every(
(diagnostic) => diagnostic.severity === "info"
)
).toBe(true);
});
});
+1
View File
@@ -187,6 +187,7 @@ export interface DocxGenerationTimings {
queueMs: number;
probeMs: number;
prepareMs: number;
themeStyleMs: number;
mediaMs: number;
referenceMs: number;
pandocMs: number;
+11 -3
View File
@@ -1,8 +1,8 @@
# @md-to-pdf/docx-engine
纯 Node.js 的 DOCX 模板和 Pandoc 编排核心。当前阶段负责读取固定 Pandoc
版本的默认 `reference.docx`、验证 ZIP 安全边界,将主题 DOCX 样式声明
解析为完整样式预设,并生成纸张、页边距、字体、段落、代码、表格、
版本的默认 `reference.docx`、验证 ZIP 安全边界,将归一化主题令牌和
兼容样式预设映射为 Word 样式,并生成纸张、页边距、字体、段落、代码、表格、
页眉、页脚和页码均已映射的动态模板。Pandoc 转换器通过静态 Lua Filter
将普通图片、Mermaid 和 ECharts 节点替换为受控 PNG,同时保留 Markdown
正文、标题、列表、表格、代码和公式的可编辑文档结构。
@@ -12,7 +12,9 @@
- 不依赖 Fastify、Electron 或浏览器 UI
- ZIP 使用纯 JavaScript `fflate`
- OOXML 使用 `@xmldom/xmldom`,不拼接未经转义的用户 XML
- 不解析任意主题 CSS主题通过 `docxStyle` 提供受限结构化覆盖;
- 不解析任意主题 CSS只消费 `@md-to-pdf/docx-theme-engine` 的受限
语义令牌;`docxStyle` 预设在槽位缺失时提供兼容降级;
- 槽位只映射到稳定的标准或 `Md*` Word 样式,不包含主题 ID 分支;
- 生成结果会复验全部 XML、包内关系、内容类型、最终节和关键样式;
- 每次转换使用独立临时目录、隔离 Pandoc data 目录并在 `finally` 清理;
- 媒体映射检查顺序、PNG、像素、单图/总大小和物理显示尺寸;
@@ -37,6 +39,7 @@ npm run build -w @md-to-pdf/docx-engine
npm run verify:docx-reference
npm run verify:docx-conversion
npm run verify:docx-acceptance
npm run verify:docx-theme-styles
```
`verify:docx-reference` 要求本机 `PATH` 中存在 Pandoc 3.9.0.2,也可以通过
@@ -51,3 +54,8 @@ npm run verify:docx-acceptance
`output/docx-acceptance/`,供 Word/WPS 互操作验收使用。根级命令还会
依次执行动态模板、媒体转换、Server HTTP 和 Desktop 原生保存验收;
仅需重跑三配置矩阵时可使用 `npm run verify:docx-matrix`
`verify:docx-theme-styles` 使用 Playwright 与 Electron 分别采集 14 套
内置主题的 56 个槽位,检查跨引擎令牌一致性,并使用固定 Pandoc 默认
模板生成 14 份动态 `reference.docx`。模板矩阵检查标准 Markdown 样式、
结构化 `Md*` 样式、正文字体、字号、字体表和缓存指纹。
+1
View File
@@ -27,6 +27,7 @@
},
"dependencies": {
"@md-to-pdf/core": "0.1.0",
"@md-to-pdf/docx-theme-engine": "0.1.0",
"@xmldom/xmldom": "0.9.10",
"fflate": "0.8.3"
},
+1
View File
@@ -12,4 +12,5 @@ export * from "./reference-package.js";
export * from "./section-transform.js";
export * from "./style-presets.js";
export * from "./styles-transform.js";
export * from "./token-style-map.js";
export * from "./validator.js";
@@ -17,6 +17,7 @@ import {
type PreparedDocxMedia,
type ThemeManifest
} from "@md-to-pdf/core";
import type { DocxThemeTokenSet } from "@md-to-pdf/docx-theme-engine";
import {
createDynamicReferenceCacheKey,
createDynamicReferenceDocx,
@@ -57,6 +58,7 @@ export interface PandocDocxConversionInput {
language: string;
exportConfig: ExportConfig;
theme: ThemeManifest;
themeTokens: DocxThemeTokenSet;
metadata: MarkdownDocumentMetadata;
media: PreparedDocxMedia;
}
@@ -133,6 +135,7 @@ function referenceOptions(
return {
exportConfig: input.exportConfig,
theme: input.theme,
themeTokens: input.themeTokens,
fileName: input.fileName,
metadata: {
title: input.metadata.title,
+26 -4
View File
@@ -23,6 +23,10 @@ import {
transformStylesXml,
transformThemeFontsXml
} from "./styles-transform.js";
import {
createDocxTokenSlotMap,
resolveTokenFonts
} from "./token-style-map.js";
import { resolveDocxThemeStyle } from "./style-presets.js";
import { validateDynamicReferenceDocx } from "./validator.js";
@@ -67,6 +71,7 @@ export function createDynamicReferenceCacheKey(
docxStyle: options.theme.docxStyle,
pageDefaults: options.theme.pageDefaults
},
themeTokens: options.themeTokens,
paper: options.exportConfig.paper,
pageDecorationsMode:
options.exportConfig.pageDecorationsMode,
@@ -85,29 +90,46 @@ export function createDynamicReferenceDocx(
): DynamicReferenceDocxResult {
const reference = readReferenceDocxPackage(baseline);
const style = resolveDocxThemeStyle(options.theme);
const tokenSlots = options.themeTokens
? createDocxTokenSlotMap(options.themeTokens)
: undefined;
const bodyFonts = resolveTokenFonts(
tokenSlots?.get("paragraph")?.style ??
tokenSlots?.get("document")?.style,
style.body.fonts
);
const page = resolveReferencePageOptions(options);
const withDecorations = createHeaderFooterParts(
reference.entries,
options,
page.header,
page.footer,
style.body.fonts.eastAsia,
bodyFonts.eastAsia,
page.dimensions.width - page.margins.left - page.margins.right
);
const entries = withDecorations.entries;
entries.set(
"word/styles.xml",
transformStylesXml(entries.get("word/styles.xml")!, style)
transformStylesXml(
entries.get("word/styles.xml")!,
style,
options.themeTokens
)
);
entries.set(
"word/fontTable.xml",
transformFontTableXml(entries.get("word/fontTable.xml")!, style)
transformFontTableXml(
entries.get("word/fontTable.xml")!,
style,
options.themeTokens
)
);
entries.set(
"word/theme/theme1.xml",
transformThemeFontsXml(
entries.get("word/theme/theme1.xml")!,
style
style,
options.themeTokens
)
);
entries.set(
@@ -7,12 +7,14 @@ import {
type MarkdownDocumentMetadata,
type ThemeManifest
} from "@md-to-pdf/core";
import type { DocxThemeTokenSet } from "@md-to-pdf/docx-theme-engine";
export interface DynamicReferenceDocxOptions {
exportConfig: ExportConfig;
theme: ThemeManifest;
fileName: string;
metadata: Pick<MarkdownDocumentMetadata, "title" | "author">;
themeTokens?: DocxThemeTokenSet;
}
export function resolveReferencePageOptions(
+482 -7
View File
@@ -1,5 +1,16 @@
import type { DocxFontFamily } from "@md-to-pdf/core";
import type {
DocxSlotStyleToken,
DocxStyleSlotName,
DocxThemeTokenSet
} from "@md-to-pdf/docx-theme-engine";
import type { ResolvedDocxThemeStyle } from "./style-presets.js";
import {
DOCX_SLOT_WORD_STYLE_BINDINGS,
collectDocxTokenFonts,
createDocxTokenSlotMap,
resolveTokenFonts
} from "./token-style-map.js";
import {
DRAWING_NAMESPACE,
WORD_NAMESPACE,
@@ -152,6 +163,242 @@ function ensureStyle(
return style;
}
function setWordAttribute(
element: XmlElement,
name: string,
value: string
) {
element.setAttributeNS(WORD_NAMESPACE, `w:${name}`, value);
}
function toggleProperty(
parent: XmlElement,
localName: string,
enabled: boolean | undefined
) {
if (enabled === undefined) {
return;
}
removeDirectChildren(parent, WORD_NAMESPACE, localName);
if (enabled) {
appendElement(parent, WORD_NAMESPACE, `w:${localName}`);
}
}
function fallbackFontsForSlot(
slot: DocxStyleSlotName,
fallback: ResolvedDocxThemeStyle
) {
if (slot.startsWith("heading-") || slot.endsWith("-title")) {
return fallback.headings.fonts;
}
if (slot === "inline-code" || slot === "code-block") {
return fallback.code.fonts;
}
if (
slot === "table" ||
slot === "table-header" ||
slot === "table-cell"
) {
return fallback.table.fonts;
}
if (slot === "caption") {
return fallback.caption.fonts;
}
return fallback.body.fonts;
}
function applyTokenRunStyle(
parent: XmlElement,
token: DocxSlotStyleToken,
fallbackFonts: DocxFontFamily
) {
if (token.fontCandidates.length) {
setFonts(parent, resolveTokenFonts(token, fallbackFonts));
}
if (token.fontSizePt !== undefined) {
removeDirectChildren(parent, WORD_NAMESPACE, "sz", "szCs");
const size = pointsToHalfPoints(token.fontSizePt);
appendElement(parent, WORD_NAMESPACE, "w:sz", {
"w:val": size
});
appendElement(parent, WORD_NAMESPACE, "w:szCs", {
"w:val": size
});
}
if (token.color !== undefined) {
removeDirectChildren(parent, WORD_NAMESPACE, "color");
appendElement(parent, WORD_NAMESPACE, "w:color", {
"w:val": colorValue(token.color)
});
}
if (token.backgroundColor !== undefined) {
removeDirectChildren(parent, WORD_NAMESPACE, "shd");
appendElement(parent, WORD_NAMESPACE, "w:shd", {
"w:val": "clear",
"w:color": "auto",
"w:fill": colorValue(token.backgroundColor)
});
}
if (token.underline !== undefined) {
removeDirectChildren(parent, WORD_NAMESPACE, "u");
if (token.underline) {
appendElement(parent, WORD_NAMESPACE, "w:u", {
"w:val": "single"
});
}
}
if (token.letterSpacingPt !== undefined) {
removeDirectChildren(parent, WORD_NAMESPACE, "spacing");
appendElement(parent, WORD_NAMESPACE, "w:spacing", {
"w:val": pointsToTwips(token.letterSpacingPt)
});
}
toggleProperty(parent, "b", token.bold);
toggleProperty(parent, "bCs", token.bold);
toggleProperty(parent, "i", token.italic);
toggleProperty(parent, "iCs", token.italic);
toggleProperty(parent, "strike", token.strikethrough);
toggleProperty(parent, "vanish", token.hidden);
}
function applyTokenBorders(
parent: XmlElement,
token: DocxSlotStyleToken
) {
if (!token.borders) {
return;
}
removeDirectChildren(parent, WORD_NAMESPACE, "pBdr");
const borders = appendElement(
parent,
WORD_NAMESPACE,
"w:pBdr"
);
for (const side of [
"top",
"right",
"bottom",
"left"
] as const) {
const border = token.borders[side];
if (!border) {
continue;
}
appendElement(borders, WORD_NAMESPACE, `w:${side}`, {
"w:val": border.style,
"w:sz": String(
Math.max(0, Math.min(160, Math.round(border.widthPt * 8)))
),
"w:space": String(
Math.max(
0,
Math.min(
31,
Math.round(token.paddingPt?.[side] ?? 0)
)
)
),
"w:color": colorValue(border.color)
});
}
}
function applyTokenParagraphStyle(
parent: XmlElement,
token: DocxSlotStyleToken
) {
if (
token.spacingBeforePt !== undefined ||
token.spacingAfterPt !== undefined ||
token.lineSpacing !== undefined
) {
const spacing = ensureDirectElement(
parent,
WORD_NAMESPACE,
"w:spacing"
);
if (token.spacingBeforePt !== undefined) {
setWordAttribute(
spacing,
"before",
pointsToTwips(token.spacingBeforePt)
);
}
if (token.spacingAfterPt !== undefined) {
setWordAttribute(
spacing,
"after",
pointsToTwips(token.spacingAfterPt)
);
}
if (token.lineSpacing !== undefined) {
setWordAttribute(
spacing,
"line",
String(Math.round(token.lineSpacing * 240))
);
setWordAttribute(spacing, "lineRule", "auto");
}
}
if (
token.firstLineIndentPt !== undefined ||
token.leftIndentPt !== undefined ||
token.rightIndentPt !== undefined
) {
const indent = ensureDirectElement(
parent,
WORD_NAMESPACE,
"w:ind"
);
if (token.firstLineIndentPt !== undefined) {
setWordAttribute(
indent,
"firstLine",
pointsToTwips(token.firstLineIndentPt)
);
indent.removeAttributeNS(WORD_NAMESPACE, "firstLineChars");
}
if (token.leftIndentPt !== undefined) {
setWordAttribute(
indent,
"left",
pointsToTwips(token.leftIndentPt)
);
indent.removeAttributeNS(WORD_NAMESPACE, "leftChars");
}
if (token.rightIndentPt !== undefined) {
setWordAttribute(
indent,
"right",
pointsToTwips(token.rightIndentPt)
);
}
}
if (token.alignment !== undefined) {
removeDirectChildren(parent, WORD_NAMESPACE, "jc");
appendElement(parent, WORD_NAMESPACE, "w:jc", {
"w:val": token.alignment
});
}
if (token.backgroundColor !== undefined) {
removeDirectChildren(parent, WORD_NAMESPACE, "shd");
appendElement(parent, WORD_NAMESPACE, "w:shd", {
"w:val": "clear",
"w:color": "auto",
"w:fill": colorValue(token.backgroundColor)
});
}
applyTokenBorders(parent, token);
toggleProperty(parent, "keepLines", token.keepLines);
toggleProperty(parent, "keepNext", token.keepWithNext);
toggleProperty(
parent,
"pageBreakBefore",
token.pageBreakBefore
);
}
function applyBodyStyle(
styles: XmlElement,
style: ResolvedDocxThemeStyle
@@ -589,9 +836,215 @@ function applyTableAndCaption(
);
}
function applyTableTokenStyles(
styles: XmlElement,
slots: ReadonlyMap<
DocxStyleSlotName,
{ style: DocxSlotStyleToken }
>,
fallback: ResolvedDocxThemeStyle
) {
const tableToken = slots.get("table")?.style;
const cellToken =
slots.get("table-cell")?.style ?? tableToken;
const headerToken =
slots.get("table-header")?.style ?? cellToken;
if (!tableToken && !cellToken && !headerToken) {
return;
}
const table = ensureStyle(
styles,
"Table",
"table",
"TableNormal"
);
const tblPr = ensureDirectElement(
table,
WORD_NAMESPACE,
"w:tblPr"
);
if (tableToken?.widthPercent !== undefined) {
removeDirectChildren(tblPr, WORD_NAMESPACE, "tblW");
appendElement(tblPr, WORD_NAMESPACE, "w:tblW", {
"w:w": String(
Math.round(tableToken.widthPercent * 50)
),
"w:type": "pct"
});
}
const padding = cellToken?.paddingPt ?? tableToken?.paddingPt;
if (padding) {
removeDirectChildren(tblPr, WORD_NAMESPACE, "tblCellMar");
const margins = appendElement(
tblPr,
WORD_NAMESPACE,
"w:tblCellMar"
);
for (const side of [
"top",
"right",
"bottom",
"left"
] as const) {
appendElement(margins, WORD_NAMESPACE, `w:${side}`, {
"w:w": pointsToTwips(padding[side]),
"w:type": "dxa"
});
}
}
if (tableToken?.borders) {
removeDirectChildren(tblPr, WORD_NAMESPACE, "tblBorders");
const borders = appendElement(
tblPr,
WORD_NAMESPACE,
"w:tblBorders"
);
for (const side of [
"top",
"right",
"bottom",
"left"
] as const) {
const border = tableToken.borders[side];
if (!border) {
continue;
}
appendElement(borders, WORD_NAMESPACE, `w:${side}`, {
"w:val": border.style,
"w:sz": String(Math.round(border.widthPt * 8)),
"w:space": "0",
"w:color": colorValue(border.color)
});
}
}
if (cellToken) {
applyTokenRunStyle(
ensureDirectElement(table, WORD_NAMESPACE, "w:rPr"),
cellToken,
fallback.table.fonts
);
}
let firstRow = directChildren(
table,
WORD_NAMESPACE,
"tblStylePr"
).find(
(element) =>
element.getAttributeNS(WORD_NAMESPACE, "type") === "firstRow"
);
if (!firstRow) {
firstRow = appendElement(
table,
WORD_NAMESPACE,
"w:tblStylePr",
{ "w:type": "firstRow" }
);
}
if (headerToken) {
applyTokenRunStyle(
ensureDirectElement(firstRow, WORD_NAMESPACE, "w:rPr"),
headerToken,
fallback.table.fonts
);
if (headerToken.backgroundColor !== undefined) {
const tcPr = ensureDirectElement(
firstRow,
WORD_NAMESPACE,
"w:tcPr"
);
removeDirectChildren(tcPr, WORD_NAMESPACE, "shd");
appendElement(tcPr, WORD_NAMESPACE, "w:shd", {
"w:val": "clear",
"w:color": "auto",
"w:fill": colorValue(
headerToken.backgroundColor
)
});
}
}
}
function applyTokenStyles(
styles: XmlElement,
tokens: DocxThemeTokenSet,
fallback: ResolvedDocxThemeStyle
) {
const slots = createDocxTokenSlotMap(tokens);
const documentToken = slots.get("document")?.style;
if (documentToken) {
const defaults = firstDirectChild(
styles,
WORD_NAMESPACE,
"docDefaults"
);
if (defaults) {
applyTokenRunStyle(
ensureDirectElement(
ensureDirectElement(
defaults,
WORD_NAMESPACE,
"w:rPrDefault"
),
WORD_NAMESPACE,
"w:rPr"
),
documentToken,
fallback.body.fonts
);
applyTokenParagraphStyle(
ensureDirectElement(
ensureDirectElement(
defaults,
WORD_NAMESPACE,
"w:pPrDefault"
),
WORD_NAMESPACE,
"w:pPr"
),
documentToken
);
}
}
for (const [slot, entry] of slots) {
const bindings = DOCX_SLOT_WORD_STYLE_BINDINGS[slot];
if (!bindings) {
continue;
}
for (const binding of bindings) {
const target = ensureStyle(
styles,
binding.styleId,
binding.type,
binding.basedOn
);
if (binding.type === "paragraph") {
applyTokenParagraphStyle(
ensureDirectElement(
target,
WORD_NAMESPACE,
"w:pPr"
),
entry.style
);
}
applyTokenRunStyle(
ensureDirectElement(
target,
WORD_NAMESPACE,
"w:rPr"
),
entry.style,
fallbackFontsForSlot(slot, fallback)
);
}
}
applyTableTokenStyles(styles, slots, fallback);
}
export function transformStylesXml(
content: Uint8Array,
style: ResolvedDocxThemeStyle
style: ResolvedDocxThemeStyle,
tokens?: DocxThemeTokenSet
) {
const document = parseXmlPart(content, "word/styles.xml");
const styles = document.documentElement!;
@@ -605,12 +1058,16 @@ export function transformStylesXml(
applyHeadings(styles, style);
applyCodeAndQuote(styles, style);
applyTableAndCaption(styles, style);
if (tokens) {
applyTokenStyles(styles, tokens, style);
}
return serializeXmlPart(document);
}
export function transformFontTableXml(
content: Uint8Array,
style: ResolvedDocxThemeStyle
style: ResolvedDocxThemeStyle,
tokens?: DocxThemeTokenSet
) {
const document = parseXmlPart(content, "word/fontTable.xml");
const root = document.documentElement!;
@@ -628,6 +1085,11 @@ export function transformFontTableXml(
requiredFonts.add(fonts.complexScript);
}
}
if (tokens) {
for (const font of collectDocxTokenFonts(tokens)) {
requiredFonts.add(font);
}
}
const existing = new Set(
Array.from(
root.getElementsByTagNameNS(WORD_NAMESPACE, "font")
@@ -648,9 +1110,22 @@ export function transformFontTableXml(
export function transformThemeFontsXml(
content: Uint8Array,
style: ResolvedDocxThemeStyle
style: ResolvedDocxThemeStyle,
tokens?: DocxThemeTokenSet
) {
const document = parseXmlPart(content, "word/theme/theme1.xml");
const slots = tokens
? createDocxTokenSlotMap(tokens)
: undefined;
const headingFonts = resolveTokenFonts(
slots?.get("heading-1")?.style,
style.headings.fonts
);
const bodyFonts = resolveTokenFonts(
slots?.get("paragraph")?.style ??
slots?.get("document")?.style,
style.body.fonts
);
for (const schemeName of ["majorFont", "minorFont"]) {
const scheme = document.getElementsByTagNameNS(
DRAWING_NAMESPACE,
@@ -670,14 +1145,14 @@ export function transformThemeFontsXml(
latin?.setAttribute(
"typeface",
schemeName === "majorFont"
? style.headings.fonts.latin
: style.body.fonts.latin
? headingFonts.latin
: bodyFonts.latin
);
eastAsia?.setAttribute(
"typeface",
schemeName === "majorFont"
? style.headings.fonts.eastAsia
: style.body.fonts.eastAsia
? headingFonts.eastAsia
: bodyFonts.eastAsia
);
}
return serializeXmlPart(document);
+192
View File
@@ -0,0 +1,192 @@
import type { DocxFontFamily } from "@md-to-pdf/core";
import {
docxThemeTokenSetSchema,
type DocxResolvedStyleSlot,
type DocxSlotStyleToken,
type DocxStyleSlotName,
type DocxThemeTokenSet
} from "@md-to-pdf/docx-theme-engine";
export interface DocxWordStyleBinding {
styleId: string;
type: "paragraph" | "character";
basedOn?: string;
}
function paragraph(
styleId: string,
basedOn = "Normal"
): DocxWordStyleBinding {
return { styleId, type: "paragraph", basedOn };
}
function character(
styleId: string,
basedOn = "DefaultParagraphFont"
): DocxWordStyleBinding {
return { styleId, type: "character", basedOn };
}
export const DOCX_SLOT_WORD_STYLE_BINDINGS: Readonly<
Partial<
Record<
DocxStyleSlotName,
readonly DocxWordStyleBinding[]
>
>
> = {
"document-title": [paragraph("Title")],
"document-author": [paragraph("Author")],
paragraph: [
paragraph("Normal", ""),
paragraph("BodyText"),
paragraph("FirstParagraph"),
paragraph("Definition"),
paragraph("Figure")
],
strong: [character("Strong")],
emphasis: [character("Emphasis")],
strikethrough: [character("Strikeout")],
"inline-code": [character("VerbatimChar", "BodyTextChar")],
"heading-1": [
paragraph("Heading1"),
character("Heading1Char")
],
"heading-2": [
paragraph("Heading2"),
character("Heading2Char")
],
"heading-3": [
paragraph("Heading3"),
character("Heading3Char")
],
"heading-4": [
paragraph("Heading4"),
character("Heading4Char")
],
"heading-5": [
paragraph("Heading5"),
character("Heading5Char")
],
"heading-6": [
paragraph("Heading6"),
character("Heading6Char")
],
"unordered-list": [paragraph("MdUnorderedList", "BodyText")],
"ordered-list": [paragraph("MdOrderedList", "BodyText")],
"list-item": [
paragraph("Compact", "BodyText"),
paragraph("ListParagraph", "BodyText")
],
"block-quote": [paragraph("BlockText", "BodyText")],
"code-block": [paragraph("SourceCode")],
hyperlink: [character("Hyperlink", "BodyTextChar")],
figure: [paragraph("Figure")],
caption: [
paragraph("Caption"),
paragraph("TableCaption", "Caption"),
paragraph("ImageCaption", "Caption")
],
footnotes: [paragraph("FootnoteText")],
"official-masthead": [paragraph("MdOfficialMasthead")],
"official-classification": [
paragraph("MdOfficialClassification")
],
"official-issuer": [paragraph("MdOfficialIssuer")],
"official-issue-row": [paragraph("MdOfficialIssueRow")],
"official-number": [paragraph("MdOfficialNumber")],
"official-signatory": [paragraph("MdOfficialSignatory")],
"official-title": [paragraph("MdOfficialTitle")],
"official-signature": [paragraph("MdOfficialSignature")],
"official-edition": [paragraph("MdOfficialEdition")],
"briefing-masthead": [paragraph("MdBriefingMasthead")],
"briefing-meta": [paragraph("MdBriefingMeta")],
"briefing-title": [paragraph("MdBriefingTitle")],
"briefing-contact": [paragraph("MdBriefingContact")],
"project-report-cover": [paragraph("MdProjectReportCover")],
"project-report-project-name": [
paragraph("MdProjectReportProjectName")
],
"project-report-title": [paragraph("MdProjectReportTitle")],
"project-report-version": [
paragraph("MdProjectReportVersion")
],
"project-report-owner": [paragraph("MdProjectReportOwner")],
"project-report-prepared-by": [
paragraph("MdProjectReportPreparedBy")
],
"project-report-date": [paragraph("MdProjectReportDate")],
"tender-cover": [paragraph("MdTenderCover")],
"tender-copy-mark": [paragraph("MdTenderCopyMark")],
"tender-project-name": [paragraph("MdTenderProjectName")],
"tender-project-number": [
paragraph("MdTenderProjectNumber")
],
"tender-title": [paragraph("MdTenderTitle")],
"tender-volume": [paragraph("MdTenderVolume")],
"tender-bidder": [paragraph("MdTenderBidder")],
"tender-representative": [
paragraph("MdTenderRepresentative")
],
"tender-date": [paragraph("MdTenderDate")]
};
const eastAsiaFontPattern =
/(?:yahei|simsun|simhei|fangsong|kaiti|dengxian|cjk|source\s+han|pingfang|heiti|songti|||仿||线)/iu;
const genericFontNames = new Set([
"serif",
"sans-serif",
"monospace",
"cursive",
"fantasy",
"system-ui"
]);
export function resolveTokenFonts(
token: DocxSlotStyleToken | undefined,
fallback: DocxFontFamily
): DocxFontFamily {
const candidates = (token?.fontCandidates ?? []).filter(
(candidate) =>
!genericFontNames.has(candidate.trim().toLowerCase())
);
if (!candidates.length) {
return fallback;
}
const eastAsia =
candidates.find((candidate) =>
eastAsiaFontPattern.test(candidate)
) ?? candidates[0]!;
const latin =
candidates.find(
(candidate) => !eastAsiaFontPattern.test(candidate)
) ?? candidates[0]!;
return {
latin,
eastAsia,
complexScript: latin
};
}
export function createDocxTokenSlotMap(
input: DocxThemeTokenSet
): ReadonlyMap<DocxStyleSlotName, DocxResolvedStyleSlot> {
const tokens = docxThemeTokenSetSchema.parse(input);
return new Map(
tokens.slots.map((entry) => [entry.slot, entry])
);
}
export function collectDocxTokenFonts(
input: DocxThemeTokenSet
): Set<string> {
const fonts = new Set<string>();
for (const { style } of input.slots) {
for (const candidate of style.fontCandidates) {
if (!genericFontNames.has(candidate.trim().toLowerCase())) {
fonts.add(candidate);
}
}
}
return fonts;
}
@@ -13,6 +13,7 @@ import {
type PreparedDocxMedia,
type ThemeManifest
} from "@md-to-pdf/core";
import type { DocxThemeTokenSet } from "@md-to-pdf/docx-theme-engine";
import {
PandocDocxConversionError,
PandocDocxConverter,
@@ -49,6 +50,16 @@ const emptyMedia: PreparedDocxMedia = {
totalBytes: 0
};
const themeTokens: DocxThemeTokenSet = {
schemaVersion: 1,
themeId: "test-theme",
themeFingerprint: "c".repeat(64),
mode: "auto-with-overrides",
basePreset: "technical",
slots: [],
diagnostics: []
};
function input() {
return {
markdown: "# 测试",
@@ -56,6 +67,7 @@ function input() {
language: "zh-CN",
exportConfig: defaultExportConfig,
theme: theme(),
themeTokens,
metadata: {
title: "测试",
author: "测试人",
@@ -6,6 +6,10 @@ import {
type ExportConfig,
type ThemeManifest
} from "@md-to-pdf/core";
import type {
DocxResolvedStyleSlot,
DocxThemeTokenSet
} from "@md-to-pdf/docx-theme-engine";
import {
createDynamicReferenceDocx,
validateGeneratedDocx,
@@ -129,6 +133,21 @@ function theme(
};
}
function themeTokens(
fingerprint = "c".repeat(64),
slots: DocxResolvedStyleSlot[] = []
): DocxThemeTokenSet {
return {
schemaVersion: 1,
themeId: "test-theme",
themeFingerprint: fingerprint,
mode: "auto-with-overrides",
basePreset: "technical",
slots,
diagnostics: []
};
}
function createOptions(exportConfig: ExportConfig) {
return {
exportConfig,
@@ -137,7 +156,8 @@ function createOptions(exportConfig: ExportConfig) {
metadata: {
title: "年度 <报告>",
author: "测试人"
}
},
themeTokens: themeTokens()
};
}
@@ -159,6 +179,10 @@ describe("动态 reference.docx", () => {
author: "测试人"
}
});
const changedTokens = createDynamicReferenceDocx(baseline, {
...createOptions(defaultExportConfig),
themeTokens: themeTokens("d".repeat(64))
});
const entries = unzipSync(first.content);
const documentXml = decoder.decode(entries["word/document.xml"]);
const stylesXml = decoder.decode(entries["word/styles.xml"]);
@@ -169,6 +193,7 @@ describe("动态 reference.docx", () => {
);
expect(first.cacheKey).toBe(second.cacheKey);
expect(first.cacheKey).not.toBe(changedMetadata.cacheKey);
expect(first.cacheKey).not.toBe(changedTokens.cacheKey);
expect(documentXml).toContain('w:w="11906"');
expect(documentXml).toContain('w:h="16838"');
expect(documentXml).toContain('w:top="907"');
@@ -186,6 +211,106 @@ describe("动态 reference.docx", () => {
});
});
it("将通用主题令牌写入标准样式、结构样式和字体部件", () => {
const result = createDynamicReferenceDocx(
createBaselineReference(),
{
...createOptions(defaultExportConfig),
themeTokens: themeTokens("e".repeat(64), [
{
slot: "paragraph",
source: "computed-css",
confidence: "exact",
style: {
fontCandidates: [
"Source Han Serif SC",
"Times New Roman"
],
fontSizePt: 14,
color: "#112233",
lineSpacing: 1.8,
firstLineIndentPt: 28
}
},
{
slot: "heading-1",
source: "computed-css",
confidence: "exact",
style: {
fontCandidates: ["SimHei", "Arial"],
fontSizePt: 22,
bold: true,
color: "#aa0000",
alignment: "center",
keepWithNext: true
}
},
{
slot: "table",
source: "computed-css",
confidence: "exact",
style: {
fontCandidates: [],
widthPercent: 100,
paddingPt: {
top: 3,
right: 4,
bottom: 3,
left: 4
},
borders: {
top: {
widthPt: 1,
color: "#445566",
style: "single"
}
}
}
},
{
slot: "official-title",
source: "computed-css",
confidence: "exact",
style: {
fontCandidates: ["SimSun"],
fontSizePt: 22,
color: "#ff0000",
alignment: "center",
keepWithNext: true
}
}
])
}
);
const entries = unzipSync(result.content);
const stylesXml = decoder.decode(entries["word/styles.xml"]);
const fontTableXml = decoder.decode(
entries["word/fontTable.xml"]
);
const themeXml = decoder.decode(
entries["word/theme/theme1.xml"]
);
expect(stylesXml).toContain('w:styleId="MdOfficialTitle"');
expect(stylesXml).toContain('w:eastAsia="Source Han Serif SC"');
expect(stylesXml).toContain('w:ascii="Times New Roman"');
expect(stylesXml).toContain('w:val="112233"');
expect(stylesXml).toContain('w:styleId="Heading1"');
expect(stylesXml).toContain('w:val="AA0000"');
expect(stylesXml).toContain('w:w="5000" w:type="pct"');
expect(stylesXml).toContain('w:color="445566"');
expect(fontTableXml).toContain(
'w:name="Source Han Serif SC"'
);
expect(fontTableXml).toContain('w:name="SimHei"');
expect(themeXml).toContain(
'<a:latin typeface="Arial"'
);
expect(themeXml).toContain(
'<a:ea typeface="SimHei"'
);
});
it("生成横向、自定义页边距和奇偶首页页眉页脚", () => {
const exportConfig: ExportConfig = {
...defaultExportConfig,
@@ -0,0 +1,76 @@
import { describe, expect, it } from "vitest";
import type { DocxThemeTokenSet } from "@md-to-pdf/docx-theme-engine";
import {
DOCX_SLOT_WORD_STYLE_BINDINGS,
collectDocxTokenFonts,
createDocxTokenSlotMap,
resolveTokenFonts
} from "../src/index.js";
const tokens: DocxThemeTokenSet = {
schemaVersion: 1,
themeId: "test-theme",
themeFingerprint: "a".repeat(64),
mode: "auto",
basePreset: "general",
slots: [
{
slot: "paragraph",
source: "computed-css",
confidence: "exact",
style: {
fontCandidates: [
"Source Han Serif SC",
"Times New Roman",
"serif"
]
}
}
],
diagnostics: []
};
describe("DOCX 令牌样式映射", () => {
it("以稳定样式 ID 覆盖标准 Markdown 和结构槽位", () => {
expect(
DOCX_SLOT_WORD_STYLE_BINDINGS["heading-1"]
).toEqual([
{
styleId: "Heading1",
type: "paragraph",
basedOn: "Normal"
},
{
styleId: "Heading1Char",
type: "character",
basedOn: "DefaultParagraphFont"
}
]);
expect(
DOCX_SLOT_WORD_STYLE_BINDINGS["official-title"]?.[0]
?.styleId
).toBe("MdOfficialTitle");
expect(
DOCX_SLOT_WORD_STYLE_BINDINGS["tender-representative"]?.[0]
?.styleId
).toBe("MdTenderRepresentative");
});
it("按文字系统选择字体并过滤 CSS 通用族名", () => {
const slot = createDocxTokenSlotMap(tokens).get("paragraph");
expect(
resolveTokenFonts(slot?.style, {
latin: "Arial",
eastAsia: "SimSun"
})
).toEqual({
latin: "Times New Roman",
eastAsia: "Source Han Serif SC",
complexScript: "Times New Roman"
});
expect([...collectDocxTokenFonts(tokens)]).toEqual([
"Source Han Serif SC",
"Times New Roman"
]);
});
});
+3 -2
View File
@@ -29,6 +29,7 @@ DOCX 主题映射引擎负责在浏览器主题 CSS 与 Word 样式之间建立
Server Playwright 与 Desktop Electron 适配器负责在现有受限浏览器生命周期
中执行本包脚本;主题 CSS 和字体资源继续使用各平台已有的同源资源协议。
`@md-to-pdf/docx-engine` 后续消费本包令牌生成 `reference.docx` 和最终
OOXML。Flex、Grid 和超出 Word 上限的装饰边不会静默丢弃,而是保留近似
`@md-to-pdf/application` 已负责指纹、快照缓存和令牌准备,
`@md-to-pdf/docx-engine` 已消费本包令牌生成动态 `reference.docx`
Flex、Grid 和超出 Word 上限的装饰边不会静默丢弃,而是保留近似
令牌与诊断,供语义文档结构层选择 Word 表格、分节或段落布局实现。
+3 -1
View File
@@ -759,7 +759,9 @@ export function resolveDocxThemeTokens(
style: normalized.style
};
} else {
const missing = entry?.matched !== true;
const missing =
input.config.mode !== "explicit" &&
entry?.matched !== true;
diagnostics.push({
severity: missing ? "warning" : "info",
code:
@@ -25,6 +25,7 @@ export type DocxThemeStyleCaptureRequest = z.infer<
>;
export interface DocxThemeStyleCaptureAdapter {
readonly themeStyleBaseUrl: string;
captureThemeStyle(
request: DocxThemeStyleCaptureRequest,
signal?: AbortSignal