Replace mr-pdf with Puppeteer-based PDF crawler

Remove the mr-pdf dependency and rework the PDF generator to use a single Puppeteer browser instance. gen-pdf.mjs now: launches one browser, renders a cover to PDF, crawls documentation pages following the "next" pagination link while injecting CSS and header/footer HTML, prints each page to PDF buffers, and merges all buffers with pdf-lib into the final PDF. Also updates metadata producer, simplifies temp file handling, and ensures the browser is closed on success or error. package.json no longer includes mr-pdf.

Author: alegauss

package.json

@@ -43,7 +43,6 @@
     "@docusaurus/tsconfig": "^3.9.2",
     "@docusaurus/types": "^3.7.0",
     "@mermaid-js/mermaid-cli": "^11.12.0",
-    "mr-pdf": "^1.1.0",
     "pdf-lib": "^1.17.1",
     "typescript": "^5.9.3",
     "wait-on": "^9.0.4"

scripts/gen-pdf.mjs

@@ -2,18 +2,17 @@
 /**
  * Turing ES Documentation — PDF Generator
  *
- * Orchestrates three stages:
- *   1. Render the branded cover (HTML → PDF via Puppeteer)
- *   2. Generate documentation pages (mr-pdf)
- *   3. Merge cover + docs into a single PDF (pdf-lib)
+ * Uses a single Puppeteer browser instance to:
+ *   1. Render the branded cover (HTML → PDF)
+ *   2. Crawl every documentation page following "next" links, printing each to PDF
+ *   3. Merge everything into a single PDF with pdf-lib
  *
  * Prerequisites:
- *   - Docusaurus dev/serve running on localhost:3000
- *   - npm dependencies installed (puppeteer, mr-pdf, pdf-lib, wait-on)
+ *   - Docusaurus serve running on localhost:3000
+ *   - npm dependencies installed (puppeteer, pdf-lib)
  */
 
-import { execFileSync } from 'node:child_process';
-import { readFileSync, writeFileSync, unlinkSync } from 'node:fs';
+import { readFileSync, writeFileSync } from 'node:fs';
 import { fileURLToPath } from 'node:url';
 import { dirname, join } from 'node:path';
 import { PDFDocument } from 'pdf-lib';
@@ -22,133 +21,149 @@ const __filename = fileURLToPath(import.meta.url);
 const __dirname  = dirname(__filename);
 const ROOT       = join(__dirname, '..');
 
-const COVER_HTML  = join(__dirname, 'pdf-cover.html');
-const STYLE_CSS   = join(__dirname, 'pdf-style.css');
-const COVER_PDF   = join(ROOT, '.cover-tmp.pdf');
-const DOCS_PDF    = join(ROOT, '.docs-tmp.pdf');
-const OUTPUT_PDF  = join(ROOT, 'turing-es-2026.1-documentation.pdf');
-const BASE_URL    = process.env.PDF_BASE_URL || 'http://localhost:3000';
-const ENTRY_PATH  = '/turing/getting-started/intro';
+const COVER_HTML = join(__dirname, 'pdf-cover.html');
+const STYLE_CSS  = join(__dirname, 'pdf-style.css');
+const OUTPUT_PDF = join(ROOT, 'turing-es-2026.1-documentation.pdf');
+const BASE_URL   = process.env.PDF_BASE_URL || 'http://localhost:3000';
+const ENTRY_PATH = '/turing/getting-started/intro';
+
+const HEADER_HTML = [
+  '<div style="width:100%;padding:0 15mm;font-family:system-ui,sans-serif;',
+  'font-size:8px;display:flex;justify-content:space-between;align-items:center;',
+  'border-bottom:0.5px solid #FED7AA;padding-bottom:3px;margin-bottom:4px;">',
+  '<span style="color:#C2410C;font-weight:700;letter-spacing:0.08em;">TURING ES</span>',
+  '<span style="color:#94a3b8;">Documentation</span>',
+  '</div>',
+].join('');
+
+const FOOTER_HTML = [
+  '<div style="width:100%;padding:0 15mm;font-family:system-ui,sans-serif;',
+  'font-size:7px;display:flex;justify-content:space-between;align-items:center;',
+  'border-top:0.5px solid #e2e8f0;padding-top:3px;margin-top:4px;">',
+  '<span style="color:#94a3b8;">viglet.com</span>',
+  '<span style="color:#64748b;">',
+  '<span class="pageNumber"></span> / <span class="totalPages"></span>',
+  '</span></div>',
+].join('');
+
+let browser;
+
+async function launchBrowser() {
+  const puppeteer = await import('puppeteer');
+  browser = await puppeteer.launch({
+    headless: 'new',
+    args: ['--no-sandbox', '--disable-setuid-sandbox', '--disable-dev-shm-usage'],
+  });
+}
 
 /* ──────────────────────────────────────────────────────
-   Stage 1 — Cover (Puppeteer)
+   Stage 1 — Cover (HTML → PDF)
    ────────────────────────────────────────────────────── */
 async function generateCover() {
   console.log('  [1/3]  Rendering cover pages …');
 
-  // Dynamic import so the script still parses even if puppeteer
-  // is only available through the mermaid-cli transitive dep.
-  const puppeteer = await import('puppeteer');
-  const browser   = await puppeteer.launch({
-    headless: 'new',
-    args: ['--no-sandbox', '--disable-setuid-sandbox'],
-  });
-
   const page = await browser.newPage();
   const html = readFileSync(COVER_HTML, 'utf-8');
 
   await page.setContent(html, { waitUntil: 'networkidle0' });
-
-  // Wait for Google Fonts to load
   await page.evaluateHandle('document.fonts.ready');
 
-  await page.pdf({
-    path: COVER_PDF,
+  const pdfBuffer = await page.pdf({
     format: 'A4',
     printBackground: true,
     preferCSSPageSize: false,
   });
 
-  await browser.close();
+  await page.close();
   console.log('         Cover rendered ✓');
+  return pdfBuffer;
 }
 
 /* ──────────────────────────────────────────────────────
-   Stage 2 — Documentation pages (mr-pdf)
+   Stage 2 — Crawl doc pages & print each to PDF
    ────────────────────────────────────────────────────── */
-function generateDocs() {
+async function generateDocs() {
   console.log('  [2/3]  Generating documentation pages …');
 
   const cssContent = readFileSync(STYLE_CSS, 'utf-8');
-
-  const headerHtml = [
-    '<div style="width:100%;padding:0 15mm;font-family:system-ui,sans-serif;',
-    'font-size:8px;display:flex;justify-content:space-between;align-items:center;',
-    'border-bottom:0.5px solid #FED7AA;padding-bottom:3px;margin-bottom:4px;">',
-    '<span style="color:#C2410C;font-weight:700;letter-spacing:0.08em;">TURING ES</span>',
-    '<span style="color:#94a3b8;">Documentation</span>',
-    '</div>',
-  ].join('');
-
-  const footerHtml = [
-    '<div style="width:100%;padding:0 15mm;font-family:system-ui,sans-serif;',
-    'font-size:7px;display:flex;justify-content:space-between;align-items:center;',
-    'border-top:0.5px solid #e2e8f0;padding-top:3px;margin-top:4px;">',
-    '<span style="color:#94a3b8;">viglet.com</span>',
-    '<span style="color:#64748b;">',
-    '<span class="pageNumber"></span> / <span class="totalPages"></span>',
-    '</span></div>',
-  ].join('');
-
-  const args = [
-    'mr-pdf',
-    '--initialDocURLs',    `${BASE_URL}${ENTRY_PATH}`,
-    '--contentSelector',   'article',
-    '--paginationSelector','a.pagination-nav__link--next',
-    '--pdfFormat',         'A4',
-    '--pdfMargin',         '25,15,20,15',
-    '--cssStyle',          cssContent,
-    '--headerTemplate',    headerHtml,
-    '--footerTemplate',    footerHtml,
-    '--disableTOC',
-    '--outputPDFFilename', DOCS_PDF,
-  ];
-
-  execFileSync('npx', args, {
-    cwd: ROOT,
-    stdio: 'inherit',
-    timeout: 300_000, // 5 min max
-  });
-
-  console.log('         Documentation pages generated ✓');
+  const page = await browser.newPage();
+  const pagePDFs = [];
+  let url = `${BASE_URL}${ENTRY_PATH}`;
+  const visited = new Set();
+
+  while (url) {
+    // Normalise to avoid revisiting with trailing slashes etc.
+    const normalized = url.replace(/\/+$/, '');
+    if (visited.has(normalized)) break;
+    visited.add(normalized);
+
+    const pageNum = visited.size;
+    process.stdout.write(`         [${pageNum}] ${normalized.replace(BASE_URL, '')} … `);
+
+    await page.goto(url, { waitUntil: 'networkidle0', timeout: 30_000 });
+
+    // Inject our PDF stylesheet
+    await page.addStyleTag({ content: cssContent });
+
+    // Small delay for styles to apply
+    await new Promise((r) => setTimeout(r, 300));
+
+    // Print this page to PDF
+    const pdfBuf = await page.pdf({
+      format: 'A4',
+      printBackground: true,
+      displayHeaderFooter: true,
+      headerTemplate: HEADER_HTML,
+      footerTemplate: FOOTER_HTML,
+      margin: { top: '25mm', bottom: '20mm', left: '15mm', right: '15mm' },
+    });
+
+    pagePDFs.push(pdfBuf);
+    console.log('✓');
+
+    // Find the "next" pagination link
+    url = await page.evaluate(() => {
+      const next = document.querySelector('a.pagination-nav__link--next');
+      return next ? next.href : null;
+    });
+  }
+
+  await page.close();
+  console.log(`         ${pagePDFs.length} pages generated ✓`);
+  return pagePDFs;
 }
 
 /* ──────────────────────────────────────────────────────
-   Stage 3 — Merge (pdf-lib)
+   Stage 3 — Merge all PDFs (pdf-lib)
    ────────────────────────────────────────────────────── */
-async function mergePDFs() {
-  console.log('  [3/3]  Merging cover + documentation …');
-
-  const coverBytes = readFileSync(COVER_PDF);
-  const docsBytes  = readFileSync(DOCS_PDF);
+async function mergePDFs(coverBuf, docBuffers) {
+  console.log('  [3/3]  Merging PDFs …');
 
-  const merged  = await PDFDocument.create();
+  const merged = await PDFDocument.create();
 
-  // Copy cover pages (cover + inner title)
-  const coverDoc   = await PDFDocument.load(coverBytes);
+  // Cover pages
+  const coverDoc   = await PDFDocument.load(coverBuf);
   const coverPages = await merged.copyPages(coverDoc, coverDoc.getPageIndices());
   coverPages.forEach((p) => merged.addPage(p));
 
-  // Copy documentation pages
-  const docsDoc   = await PDFDocument.load(docsBytes);
-  const docsPages = await merged.copyPages(docsDoc, docsDoc.getPageIndices());
-  docsPages.forEach((p) => merged.addPage(p));
+  // Documentation pages
+  for (const buf of docBuffers) {
+    const doc   = await PDFDocument.load(buf);
+    const pages = await merged.copyPages(doc, doc.getPageIndices());
+    pages.forEach((p) => merged.addPage(p));
+  }
 
   // Metadata
   merged.setTitle('Turing ES Documentation');
   merged.setSubject('Enterprise Search Platform — v2026.1');
   merged.setAuthor('Viglet');
   merged.setCreator('Viglet PDF Generator');
-  merged.setProducer('pdf-lib + Puppeteer + mr-pdf');
+  merged.setProducer('pdf-lib + Puppeteer');
   merged.setCreationDate(new Date());
 
   const mergedBytes = await merged.save();
   writeFileSync(OUTPUT_PDF, mergedBytes);
 
-  // Clean up temp files
-  try { unlinkSync(COVER_PDF); } catch { /* ignore */ }
-  try { unlinkSync(DOCS_PDF);  } catch { /* ignore */ }
-
   const sizeMB = (mergedBytes.length / 1_048_576).toFixed(2);
   console.log(`         Merged PDF saved (${sizeMB} MB) ✓`);
 }
@@ -163,17 +178,22 @@ async function main() {
   console.log('  ╚══════════════════════════════════════════╝');
   console.log();
 
-  await generateCover();
-  generateDocs();
-  await mergePDFs();
+  await launchBrowser();
+
+  const coverBuf  = await generateCover();
+  const docBufs   = await generateDocs();
+  await mergePDFs(coverBuf, docBufs);
+
+  await browser.close();
 
   console.log();
   console.log(`  ✅  ${OUTPUT_PDF}`);
   console.log();
 }
 
-main().catch((err) => {
+main().catch(async (err) => {
   console.error('\n  ❌  PDF generation failed:\n');
   console.error(err);
+  if (browser) await browser.close().catch(() => {});
   process.exit(1);
 });