md-pdf-studio 2.2.0 → 2.4.0

package/src/core/pagemarks.js

@@ -0,0 +1,142 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
+/**
+ * Page numbers and a running header/footer, printed in the page margins with CSS `@page`
+ * margin boxes. Chrome and Edge draw these both for `--print-to-pdf` and in their print
+ * dialog, so the CLI and the extension share this file: it needs nothing from Node and is
+ * shipped in the extension as extension/lib/pagemarks.js (a test keeps the two identical).
+ *
+ * Header and footer are text with placeholders: {page}, {pages}, {title}, {author}, {date}.
+ */
+(function (scope) {
+  const NUMBER_STYLES = {
+    full: 'Page {page} of {pages}',
+    simple: '{page}'
+  };
+  const DEFAULT_HEADER = '{title}';
+  const PLACEHOLDER = /\{(page|pages|title|author|date)\}/g;
+
+  /**
+   * The text for a header or footer option: a string as it is, `true` for the default.
+   * @param {*} value
+   * @param {string} fallback Text used for `true`
+   * @returns {string} '' when off
+   */
+  function textOf(value, fallback) {
+    if (value === true) return fallback;
+    if (typeof value === 'number') return String(value);
+    return typeof value === 'string' ? value.trim() : '';
+  }
+
+  /**
+   * The page number text for the `pageNumbers` option.
+   * @param {*} value true, 'full', 'simple', or false/empty for none
+   * @returns {string} '' when off
+   */
+  function numbersOf(value) {
+    if (value === undefined || value === null || value === false || value === '') return '';
+    if (value === true) return NUMBER_STYLES.full;
+    const style = String(value).trim().toLowerCase();
+    if (style === 'true') return NUMBER_STYLES.full;
+    if (style === 'false' || style === 'none' || style === 'off') return '';
+    if (!NUMBER_STYLES[style]) {
+      throw new Error(`Unknown page number style "${value}". Use one of: ${Object.keys(NUMBER_STYLES).join(', ')}`);
+    }
+    return NUMBER_STYLES[style];
+  }
+
+  /**
+   * A CSS string literal that cannot end early or close the <style> element.
+   * @param {string} text
+   * @returns {string}
+   */
+  function cssString(text) {
+    const safe = text
+      .replace(/\\/g, '\\\\')
+      .replace(/"/g, '\\"')
+      .replace(/[\r\n]+/g, ' ')
+      .replace(/</g, '\\3C ');
+    return `"${safe}"`;
+  }
+
+  /**
+   * Turns header/footer text into the value of a CSS `content` property.
+   * @param {string} text
+   * @param {{ title?: string, author?: string, date?: string }} meta
+   * @returns {string}
+   */
+  function contentValue(text, meta = {}) {
+    const parts = [];
+    let literal = '';
+    let last = 0;
+    for (const match of text.matchAll(PLACEHOLDER)) {
+      literal += text.slice(last, match.index);
+      last = match.index + match[0].length;
+      const name = match[1];
+      if (name === 'page' || name === 'pages') {
+        if (literal) parts.push(cssString(literal));
+        literal = '';
+        parts.push(`counter(${name})`);
+      } else {
+        literal += meta[name] ? String(meta[name]) : '';
+      }
+    }
+    literal += text.slice(last);
+    if (literal) parts.push(cssString(literal));
+    return parts.length ? parts.join(' ') : '""';
+  }
+
+  /**
+   * Where each text goes. The header sits top centre. The footer sits bottom centre, unless
+   * page numbers are also on: then the footer text goes left and the numbers right.
+   * @param {{ header?: *, footer?: *, pageNumbers?: * }} options
+   * @returns {Object<string, string>} margin box name -> text
+   */
+  function pageMarkBoxes(options = {}) {
+    const header = textOf(options.header, DEFAULT_HEADER);
+    const numbers = numbersOf(options.pageNumbers);
+    const footer = textOf(options.footer, NUMBER_STYLES.full);
+    const boxes = {};
+    if (header) boxes['top-center'] = header;
+    if (footer && numbers && footer !== numbers) {
+      boxes['bottom-left'] = footer;
+      boxes['bottom-right'] = numbers;
+    } else if (footer || numbers) {
+      boxes['bottom-center'] = footer || numbers;
+    }
+    return boxes;
+  }
+
+  /**
+   * CSS for the page numbers and running header/footer, or '' when none is on.
+   * A cover page (.cover-page) is printed without them.
+   * @param {{ header?: *, footer?: *, pageNumbers?: *, font?: string }} options
+   * @param {{ title?: string, author?: string, date?: string }} [meta]
+   * @returns {string}
+   */
+  function pageMarksCss(options = {}, meta = {}) {
+    const boxes = pageMarkBoxes(options);
+    const names = Object.keys(boxes);
+    if (!names.length) return '';
+
+    const font = options.font ? `${cssString(String(options.font))}, sans-serif` : 'sans-serif';
+    const style = `font-family: ${font}; font-size: 9px; color: #64748b;`;
+    const rules = names.map(
+      (name) => `  @${name} { content: ${contentValue(boxes[name], meta)}; ${style} }`
+    );
+    const blank = names.map((name) => `  @${name} { content: none; }`);
+    return [
+      '/* Page numbers and running header/footer */',
+      '@page {',
+      ...rules,
+      '}',
+      '.cover-page { page: cover; }',
+      '@page cover {',
+      ...blank,
+      '}'
+    ].join('\n');
+  }
+
+  const api = { pageMarksCss, pageMarkBoxes, contentValue };
+  if (typeof module !== 'undefined' && module.exports) module.exports = api;
+  else scope.MdPdfPageMarks = api;
+})(typeof globalThis !== 'undefined' ? globalThis : this);

package/src/core/parser.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 const { marked } = require('marked');
 const hljs = require('highlight.js/lib/common');
 const { 
@@ -11,6 +12,7 @@ const {
 } = require('./utils');
 const { getHtmlTemplate } = require('./templates');
 const { applyFootnotes } = require('./footnotes');
+const { pageMarksCss } = require('./pagemarks');
 
 /**
  * Builds marked custom renderer for diagrams, UI mockups, and headings
@@ -82,6 +84,8 @@ function parseMarkdownToHtml(rawMarkdownContent, options = {}) {
   const subtitle = frontMatter.subtitle || options.subtitle || '';
   const author = frontMatter.author || options.author || '';
   const date = frontMatter.date || options.date || '';
+  // Document language, as for EPUB: sets the cover labels and the page's lang attribute
+  const language = String(frontMatter.lang || frontMatter.language || options.language || 'en');
 
   // 2. Merge options
   const mergedOptions = {
@@ -92,8 +96,10 @@ function parseMarkdownToHtml(rawMarkdownContent, options = {}) {
     margin: options.margin || frontMatter.margin || '14mm 12mm 16mm 12mm',
     toc: options.toc !== undefined ? options.toc : (frontMatter.toc || false),
     cover: options.cover !== undefined ? options.cover : (frontMatter.cover || false),
-    header: options.header !== undefined ? options.header : (frontMatter.header || true),
-    footer: options.footer !== undefined ? options.footer : (frontMatter.footer || true),
+    // Running header/footer and page numbers: off unless asked for (see pagemarks.js)
+    header: options.header !== undefined ? options.header : frontMatter.header,
+    footer: options.footer !== undefined ? options.footer : frontMatter.footer,
+    pageNumbers: options.pageNumbers !== undefined ? options.pageNumbers : frontMatter.pageNumbers,
     mermaid: options.mermaid !== undefined ? options.mermaid : (frontMatter.mermaid !== false),
     katex: options.katex !== undefined ? options.katex : (frontMatter.katex !== false),
     highlight: options.highlight !== undefined ? options.highlight : (frontMatter.highlight !== false),
@@ -118,7 +124,8 @@ function parseMarkdownToHtml(rawMarkdownContent, options = {}) {
     subtitle,
     author,
     date,
-    cover: mergedOptions.cover
+    cover: mergedOptions.cover,
+    language
   });
 
   // 5. Read custom CSS file if specified
@@ -127,6 +134,15 @@ function parseMarkdownToHtml(rawMarkdownContent, options = {}) {
     customCss += '\n' + readCustomCss(mergedOptions.cssFile);
   }
 
+  // Page numbers and running header/footer, printed in the page margins
+  const pageCss = pageMarksCss(
+    { header: mergedOptions.header, footer: mergedOptions.footer, pageNumbers: mergedOptions.pageNumbers, font: mergedOptions.font },
+    { title, author, date }
+  );
+  if (pageCss) {
+    customCss = pageCss + '\n' + customCss;
+  }
+
   // 6. Parse Markdown using custom renderer
   marked.setOptions({ gfm: true, breaks: true });
   const renderer = buildMarkedRenderer(headingsMap, mergedOptions.highlight);
@@ -152,6 +168,7 @@ function parseMarkdownToHtml(rawMarkdownContent, options = {}) {
   return getHtmlTemplate({
     title,
     author,
+    language,
     bodyHtml,
     options: {
       ...mergedOptions,

package/src/core/sanitizer.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * HTML Sanitizer & Escaping Utilities
  */

package/src/core/templates.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 const { getThemeStyles } = require('./themes');
 
 /**
@@ -15,11 +16,12 @@ function escapeXml(str) {
  * @param {Object} params
  * @param {string} params.title Document title
  * @param {string} [params.author] Document author
+ * @param {string} [params.language] Document language tag (default en)
  * @param {string} params.bodyHtml Document body HTML
  * @param {Object} [params.options] Template options
  * @returns {string} Full HTML string
  */
-function getHtmlTemplate({ title, author = '', bodyHtml, options = {} }) {
+function getHtmlTemplate({ title, author = '', language = 'en', bodyHtml, options = {} }) {
   const theme = options.theme || 'modern';
   const enableMermaid = options.mermaid !== false;
   const enableKatex = options.katex !== false;
@@ -31,7 +33,7 @@ function getHtmlTemplate({ title, author = '', bodyHtml, options = {} }) {
   const themeStyles = getThemeStyles(theme, options.font);
 
   return `<!DOCTYPE html>
-<html lang="vi">
+<html lang="${escapeXml(language)}">
 <head>
   <meta charset="UTF-8">
   <title>${escapeXml(title)}</title>

package/src/core/themes/academic.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * Academic Theme Styles
  */

package/src/core/themes/cover.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * Cover Page CSS Styles
  */

package/src/core/themes/cyberpunk.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * Cyberpunk Theme Styles
  */

package/src/core/themes/dark.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * Dark Theme Styles
  */

package/src/core/themes/ebook.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * E-Book Theme Styles
  */

package/src/core/themes/github.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * GitHub Theme Styles
  */

package/src/core/themes/highlight.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * Syntax Highlighting Token Colors
  * Two palettes, picked by the background each theme gives its code blocks.

package/src/core/themes/index.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 const coverCss = require('./cover');
 const modernTheme = require('./modern');
 const darkTheme = require('./dark');

package/src/core/themes/minimal.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * Minimal Theme Styles
  */

package/src/core/themes/modern.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * Modern Theme Styles
  */

package/src/core/toc.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 /**
  * Table of Contents (TOC) Generator Utility
  */

package/src/core/utils.js

@@ -1,3 +1,4 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 const fs = require('fs');
 const path = require('path');
 const { parseFrontMatter } = require('./frontmatter');
@@ -18,11 +19,11 @@ function preprocessMarkdown(mdContent = '') {
 
   return mdContent
     // GitHub Callouts / Alerts
-    .replace(/^>\s*\[!NOTE\]\s*\r?\n/gm, '> **ℹ️ NOTE:** ')
-    .replace(/^>\s*\[!TIP\]\s*\r?\n/gm, '> **💡 TIP:** ')
-    .replace(/^>\s*\[!WARNING\]\s*\r?\n/gm, '> **⚠️ WARNING:** ')
-    .replace(/^>\s*\[!IMPORTANT\]\s*\r?\n/gm, '> **❗ IMPORTANT:** ')
-    .replace(/^>\s*\[!CAUTION\]\s*\r?\n/gm, '> **🚫 CAUTION:** ')
+    .replace(/^>[ \t]*\[!NOTE\][ \t]*(?:\r?\n>[ \t]?)?/gm, '> **ℹ️ NOTE:** ')
+    .replace(/^>[ \t]*\[!TIP\][ \t]*(?:\r?\n>[ \t]?)?/gm, '> **💡 TIP:** ')
+    .replace(/^>[ \t]*\[!WARNING\][ \t]*(?:\r?\n>[ \t]?)?/gm, '> **⚠️ WARNING:** ')
+    .replace(/^>[ \t]*\[!IMPORTANT\][ \t]*(?:\r?\n>[ \t]?)?/gm, '> **❗ IMPORTANT:** ')
+    .replace(/^>[ \t]*\[!CAUTION\][ \t]*(?:\r?\n>[ \t]?)?/gm, '> **🚫 CAUTION:** ')
     // Pagebreaks
     .replace(/<!--\s*page-?break\s*-->/gi, '<div class="page-break"></div>')
     .replace(/\\pagebreak/gi, '<div class="page-break"></div>')

package/src/core/zip.js

@@ -0,0 +1,78 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
+const zlib = require('zlib');
+
+/**
+ * A ZIP archive writer, enough for the two container formats built on ZIP
+ * that this tool exports: DOCX and EPUB.
+ *
+ * Entries are written in the order given, which EPUB relies on: its
+ * `mimetype` entry has to come first and uncompressed. Timestamps are fixed,
+ * so the same input gives the same bytes.
+ */
+
+// 1980-01-01 00:00:00, the earliest date the format can express.
+const DOS_TIME = 0;
+const DOS_DATE = (0 << 9) | (1 << 5) | 1;
+
+/**
+ * @param {{ name: string, data: Buffer | string, store?: boolean }[]} entries
+ * @returns {Buffer}
+ */
+function createZip(entries) {
+  const localParts = [];
+  const centralParts = [];
+  let offset = 0;
+
+  for (const entry of entries) {
+    const name = Buffer.from(entry.name, 'utf8');
+    const data = Buffer.isBuffer(entry.data) ? entry.data : Buffer.from(entry.data, 'utf8');
+    const compressed = entry.store ? data : zlib.deflateRawSync(data, { level: 9 });
+    const method = entry.store ? 0 : 8;
+    const crc = zlib.crc32(data);
+
+    const local = Buffer.alloc(30);
+    local.writeUInt32LE(0x04034b50, 0);
+    local.writeUInt16LE(20, 4); // version needed
+    local.writeUInt16LE(0x0800, 6); // names are UTF-8
+    local.writeUInt16LE(method, 8);
+    local.writeUInt16LE(DOS_TIME, 10);
+    local.writeUInt16LE(DOS_DATE, 12);
+    local.writeUInt32LE(crc, 14);
+    local.writeUInt32LE(compressed.length, 18);
+    local.writeUInt32LE(data.length, 22);
+    local.writeUInt16LE(name.length, 26);
+    local.writeUInt16LE(0, 28);
+
+    const central = Buffer.alloc(46);
+    central.writeUInt32LE(0x02014b50, 0);
+    central.writeUInt16LE(20, 4); // version made by
+    central.writeUInt16LE(20, 6); // version needed
+    central.writeUInt16LE(0x0800, 8);
+    central.writeUInt16LE(method, 10);
+    central.writeUInt16LE(DOS_TIME, 12);
+    central.writeUInt16LE(DOS_DATE, 14);
+    central.writeUInt32LE(crc, 16);
+    central.writeUInt32LE(compressed.length, 20);
+    central.writeUInt32LE(data.length, 24);
+    central.writeUInt16LE(name.length, 28);
+    central.writeUInt32LE(offset, 42);
+
+    localParts.push(local, name, compressed);
+    centralParts.push(central, name);
+    offset += local.length + name.length + compressed.length;
+  }
+
+  const centralSize = centralParts.reduce((size, part) => size + part.length, 0);
+  const end = Buffer.alloc(22);
+  end.writeUInt32LE(0x06054b50, 0);
+  end.writeUInt16LE(entries.length, 8);
+  end.writeUInt16LE(entries.length, 10);
+  end.writeUInt32LE(centralSize, 12);
+  end.writeUInt32LE(offset, 16);
+
+  return Buffer.concat([...localParts, ...centralParts, end]);
+}
+
+module.exports = {
+  createZip
+};

package/src/index.js

@@ -1,2 +1,3 @@
+/*! SPDX-License-Identifier: MIT | Copyright (c) 2026 Van Nguyen | github.com/vannt-dev/convert-md-to-pdf */
 // Re-export core API module
 module.exports = require('./core/index');