feat: 实现 Pandoc DOCX 转换服务

2026-07-30 15:05:50 +08:00
parent fa159d916f
commit bf43433d38
32 changed files with 2908 additions and 61 deletions
@@ -0,0 +1,184 @@
+import {
+  copyFile,
+  mkdtemp,
+  rm,
+  writeFile
+} from "node:fs/promises";
+import { existsSync } from "node:fs";
+import os from "node:os";
+import path from "node:path";
+import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
+import {
+  defaultExportConfig,
+  type PreparedDocxMedia,
+  type ThemeManifest
+} from "@md-to-pdf/core";
+import {
+  PandocDocxConversionError,
+  PandocDocxConverter,
+  type PandocProcessRunner,
+  type PandocRuntimeProvider
+} from "../src/index.js";
+import { createTestBaselineReference } from "./reference-test-fixture.js";
+
+function theme(): 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: { preset: "technical" },
+    bundled: true
+  };
+}
+
+const emptyMedia: PreparedDocxMedia = {
+  resources: [],
+  echartsErrors: [],
+  mermaidErrors: [],
+  warnings: [],
+  totalBytes: 0
+};
+
+function input() {
+  return {
+    markdown: "# 测试",
+    fileName: "测试.md",
+    language: "zh-CN",
+    exportConfig: defaultExportConfig,
+    theme: theme(),
+    metadata: {
+      title: "测试",
+      author: "测试人",
+      subject: "",
+      keywords: [],
+      language: "zh-CN"
+    },
+    media: emptyMedia
+  };
+}
+
+function argumentAfter(arguments_: readonly string[], name: string) {
+  const index = arguments_.indexOf(name);
+  if (index < 0 || !arguments_[index + 1]) {
+    throw new Error(`缺少参数 ${name}`);
+  }
+  return arguments_[index + 1]!;
+}
+
+describe("Pandoc DOCX 转换器", () => {
+  let temporaryRoot: string;
+  const baseline = createTestBaselineReference();
+  const runtime: PandocRuntimeProvider = {
+    async getExecutablePath() {
+      return "pandoc";
+    },
+    async getDefaultReferenceDocx() {
+      return baseline;
+    }
+  };
+
+  beforeEach(async () => {
+    temporaryRoot = await mkdtemp(
+      path.join(os.tmpdir(), "docx-converter-test-")
+    );
+  });
+
+  afterEach(async () => {
+    await rm(temporaryRoot, { recursive: true, force: true });
+  });
+
+  it("使用固定参数转换、校验并清理请求级临时目录", async () => {
+    let requestDirectory = "";
+    const runner = vi.fn<PandocProcessRunner>(
+      async (_executable, arguments_, options) => {
+        requestDirectory = options.cwd!;
+        expect(arguments_).toContain("--lua-filter");
+        expect(arguments_).toContain("--data-dir");
+        expect(arguments_).toContain("--resource-path");
+        expect(
+          options.env?.MD_TO_PDF_DOCX_MEDIA_MAP
+        ).toContain("media-map.json");
+        await copyFile(
+          argumentAfter(arguments_, "--reference-doc"),
+          argumentAfter(arguments_, "--output")
+        );
+        return {
+          outcome: "completed",
+          exitCode: 0,
+          stdout: new Uint8Array(),
+          stderr: ""
+        };
+      }
+    );
+    const converter = new PandocDocxConverter({
+      runtime,
+      runner,
+      temporaryRoot
+    });
+    const result = await converter.convert(input());
+
+    expect(result.docx.byteLength).toBeGreaterThan(0);
+    expect(result.validation.partCount).toBeGreaterThan(8);
+    expect(runner).toHaveBeenCalledTimes(1);
+    expect(existsSync(requestDirectory)).toBe(false);
+  });
+
+  it("超时和无效输出均返回稳定错误并清理目录", async () => {
+    let timeoutDirectory = "";
+    const timeoutRunner = vi.fn<PandocProcessRunner>(
+      async (_executable, _arguments, options) => {
+        timeoutDirectory = options.cwd!;
+        return {
+          outcome: "timeout",
+          exitCode: null,
+          stdout: new Uint8Array(),
+          stderr: ""
+        };
+      }
+    );
+    await expect(
+      new PandocDocxConverter({
+        runtime,
+        runner: timeoutRunner,
+        temporaryRoot
+      }).convert(input())
+    ).rejects.toMatchObject({
+      code: "DOCX_RENDER_TIMEOUT"
+    } satisfies Partial<PandocDocxConversionError>);
+    expect(existsSync(timeoutDirectory)).toBe(false);
+
+    const invalidRunner = vi.fn<PandocProcessRunner>(
+      async (_executable, arguments_) => {
+        await writeFile(
+          argumentAfter(arguments_, "--output"),
+          "not a docx"
+        );
+        return {
+          outcome: "completed",
+          exitCode: 0,
+          stdout: new Uint8Array(),
+          stderr: ""
+        };
+      }
+    );
+    await expect(
+      new PandocDocxConverter({
+        runtime,
+        runner: invalidRunner,
+        temporaryRoot
+      }).convert(input())
+    ).rejects.toMatchObject({
+      code: "DOCX_OUTPUT_INVALID"
+    } satisfies Partial<PandocDocxConversionError>);
+  });
+});