forked from EduCraft/curriculum-project-hub
4e01c18cac
Teachers need ad-hoc Markdown → PDF. Ship an MCP tool powered by md-to-pdf (Marked + headless Chrome) so remote images/CSS work, with workspace-scoped basedir, front-matter stripped so untrusted markdown cannot override dest/basedir/launch options, and MathJax for $/$ math. Also include WebFetch and WebSearch in the unrestricted role tool surface by default. Deploy skips Puppeteer's browser download and expects a host Chrome/Chromium (PUPPETEER_EXECUTABLE_PATH / CHROME_PATH).
181 lines
5.7 KiB
TypeScript
181 lines
5.7 KiB
TypeScript
import { access, constants } from "node:fs";
|
|
import { mkdtemp, mkdir, readFile, rm, writeFile } from "node:fs/promises";
|
|
import { tmpdir } from "node:os";
|
|
import { join } from "node:path";
|
|
import { promisify } from "node:util";
|
|
import { afterEach, describe, expect, it } from "vitest";
|
|
import {
|
|
defaultMarkdownToPdfOutputPath,
|
|
MarkdownToPdfError,
|
|
renderMarkdownToPdf,
|
|
} from "../../src/agent/markdownToPdf.js";
|
|
import { claudeSdkToolConfigForRole, cphHubMcpToolsForRole } from "../../src/agent/roleTools.js";
|
|
|
|
const accessAsync = promisify(access);
|
|
const roots: string[] = [];
|
|
|
|
const CHROME_CANDIDATES = [
|
|
process.env["PUPPETEER_EXECUTABLE_PATH"],
|
|
process.env["CHROME_PATH"],
|
|
process.env["GOOGLE_CHROME_BIN"],
|
|
"/Applications/Google Chrome.app/Contents/MacOS/Google Chrome",
|
|
"/usr/bin/google-chrome-stable",
|
|
"/usr/bin/chromium",
|
|
"/usr/bin/chromium-browser",
|
|
].filter((v): v is string => typeof v === "string" && v.trim() !== "");
|
|
|
|
afterEach(async () => {
|
|
while (roots.length > 0) {
|
|
const root = roots.pop();
|
|
if (root !== undefined) await rm(root, { recursive: true, force: true });
|
|
}
|
|
});
|
|
|
|
describe("role tool defaults", () => {
|
|
it("includes WebFetch, WebSearch, and markdown_to_pdf in the unrestricted surface", () => {
|
|
const config = claudeSdkToolConfigForRole(undefined);
|
|
expect(config.tools).toEqual([
|
|
"Read",
|
|
"Write",
|
|
"Bash",
|
|
"Glob",
|
|
"Grep",
|
|
"WebFetch",
|
|
"WebSearch",
|
|
]);
|
|
expect(config.allowedTools).toContain("WebFetch");
|
|
expect(config.allowedTools).toContain("WebSearch");
|
|
expect(config.allowedTools).toContain("mcp__cph_hub__markdown_to_pdf");
|
|
expect(cphHubMcpToolsForRole(undefined)).toContain("markdown_to_pdf");
|
|
});
|
|
|
|
it("maps explicit markdown_to_pdf and web tools", () => {
|
|
expect(claudeSdkToolConfigForRole(["web_fetch", "web_search", "markdown_to_pdf"])).toEqual({
|
|
tools: ["WebFetch", "WebSearch"],
|
|
allowedTools: ["WebFetch", "WebSearch", "mcp__cph_hub__markdown_to_pdf"],
|
|
});
|
|
expect(cphHubMcpToolsForRole(["markdown_to_pdf"])).toEqual(["markdown_to_pdf"]);
|
|
});
|
|
});
|
|
|
|
describe("renderMarkdownToPdf", () => {
|
|
it("defaults the output next to build/<stem>.pdf", () => {
|
|
expect(defaultMarkdownToPdfOutputPath("notes/handout.md")).toBe(join("build", "handout.pdf"));
|
|
expect(defaultMarkdownToPdfOutputPath("README")).toBe(join("build", "README.pdf"));
|
|
});
|
|
|
|
it("renders markdown through md-to-pdf with a system Chrome", async () => {
|
|
let chromePath: string | undefined;
|
|
for (const candidate of CHROME_CANDIDATES) {
|
|
try {
|
|
await accessAsync(candidate, constants.X_OK);
|
|
chromePath = candidate;
|
|
break;
|
|
} catch {
|
|
// try next
|
|
}
|
|
}
|
|
if (chromePath === undefined) {
|
|
// No browser on this host — mapping tests above still cover the contract.
|
|
return;
|
|
}
|
|
|
|
const root = await mkdtemp(join(tmpdir(), "cph-md2pdf-test-"));
|
|
roots.push(root);
|
|
const workspace = join(root, "project");
|
|
await mkdir(workspace, { recursive: true });
|
|
await writeFile(
|
|
join(workspace, "lesson.md"),
|
|
[
|
|
"# 测试讲义",
|
|
"",
|
|
"段落里有行内数学 $E=mc^2$。",
|
|
"",
|
|
"$$",
|
|
"\\int_0^1 x\\,dx = \\frac{1}{2}",
|
|
"$$",
|
|
"",
|
|
"- item a",
|
|
"- item b",
|
|
"",
|
|
"",
|
|
"",
|
|
].join("\n"),
|
|
"utf8",
|
|
);
|
|
|
|
const result = await renderMarkdownToPdf({
|
|
workspaceRoot: root,
|
|
workspaceDir: workspace,
|
|
markdownPath: "lesson.md",
|
|
outputPath: "build/lesson.pdf",
|
|
chromePath,
|
|
});
|
|
|
|
expect(result.outputPath).toBe("build/lesson.pdf");
|
|
expect(result.bytes).toBeGreaterThan(500);
|
|
const pdf = await readFile(join(workspace, "build", "lesson.pdf"));
|
|
expect(pdf.subarray(0, 4).toString("utf8")).toBe("%PDF");
|
|
}, 120_000);
|
|
|
|
it("refuses output outside the workspace", async () => {
|
|
const root = await mkdtemp(join(tmpdir(), "cph-md2pdf-boundary-"));
|
|
roots.push(root);
|
|
const workspace = join(root, "project");
|
|
await mkdir(workspace, { recursive: true });
|
|
await writeFile(join(workspace, "a.md"), "# a\n", "utf8");
|
|
|
|
await expect(
|
|
renderMarkdownToPdf({
|
|
workspaceRoot: root,
|
|
workspaceDir: workspace,
|
|
markdownPath: "a.md",
|
|
outputPath: "../escape.pdf",
|
|
}),
|
|
).rejects.toMatchObject({ reason: "boundary" } satisfies Partial<MarkdownToPdfError>);
|
|
});
|
|
|
|
it("ignores front-matter that would override basedir/dest", async () => {
|
|
let chromePath: string | undefined;
|
|
for (const candidate of CHROME_CANDIDATES) {
|
|
try {
|
|
await accessAsync(candidate, constants.X_OK);
|
|
chromePath = candidate;
|
|
break;
|
|
} catch {
|
|
// try next
|
|
}
|
|
}
|
|
if (chromePath === undefined) return;
|
|
|
|
const root = await mkdtemp(join(tmpdir(), "cph-md2pdf-fm-"));
|
|
roots.push(root);
|
|
const workspace = join(root, "project");
|
|
await mkdir(workspace, { recursive: true });
|
|
await writeFile(
|
|
join(workspace, "evil.md"),
|
|
[
|
|
"---",
|
|
"basedir: /",
|
|
"dest: /tmp/should-not-write.pdf",
|
|
"---",
|
|
"",
|
|
"# safe body",
|
|
"",
|
|
].join("\n"),
|
|
"utf8",
|
|
);
|
|
|
|
const result = await renderMarkdownToPdf({
|
|
workspaceRoot: root,
|
|
workspaceDir: workspace,
|
|
markdownPath: "evil.md",
|
|
outputPath: "build/safe.pdf",
|
|
chromePath,
|
|
});
|
|
expect(result.outputPath).toBe("build/safe.pdf");
|
|
const pdf = await readFile(join(workspace, "build", "safe.pdf"));
|
|
expect(pdf.subarray(0, 4).toString("utf8")).toBe("%PDF");
|
|
}, 120_000);
|
|
});
|