a horrible script to convert the entire PHP manual from HTML files to Markdown

convert.js

const fs = require('fs');
const path = require('path');
const TurndownService = require('turndown');

// Directories
const inputDir = path.join(__dirname, 'html');
const outputDir = path.join(__dirname, 'markdown');

// Initialize Turndown
const turndownService = new TurndownService({
  codeBlockStyle: 'fenced',
});

const replaceWithFencedCodeBlock = (language) => (content, node, options) => {
  // find <br/> elements in node and replace them with line breaks:
  var brs = node.getElementsByTagName('br')
  for (var i = 0; i < brs.length; i++) {
    brs[i].textContent = '\n'
  }
 
  var code = node.firstChild.textContent

  var fence = '```';

  return (
    '\n\n' + fence + language + '\n' +
    code.replace(/\n$/, '') +
    '\n' + fence + '\n\n'
  )
}

turndownService.addRule('phpCodeBlock', {
  filter: function (node, options) {
    return (
      options.codeBlockStyle === 'fenced' &&
      node.className === 'phpcode' &&
      node.firstChild &&
      node.firstChild.nodeName === 'PRE' &&
      node.firstChild.firstChild &&
      node.firstChild.firstChild.nodeName === 'CODE'
    )
  },

  replacement: replaceWithFencedCodeBlock('php')
});

turndownService.addRule('phpCodeExample', {
  filter: function (node, options) {
    return (
      options.codeBlockStyle === 'fenced' &&
      node.nodeName === 'PRE' &&
      node.className === 'examplescode'
    )
  },

  replacement: replaceWithFencedCodeBlock('')
});

// Ensure output directory exists
if (!fs.existsSync(outputDir)) {
    fs.mkdirSync(outputDir);
}

// Read all HTML files
const files = fs.readdirSync(inputDir).filter(file => file.endsWith('.html'));

console.log(`Total files to process: ${files.length}`);

// Process each file

files.forEach((file, index) => {
    const inputPath = path.join(inputDir, file);
    const outputPath = path.join(outputDir, file.replace(/\.html$/, '.md'));

    // Read HTML content
    var htmlContent = fs.readFileSync(inputPath, 'utf8');

    // replace `<code>` with `<pre><code>` in PHP code samples:

    htmlContent = htmlContent.replace(
      /<div class="phpcode"><code>(.*?)<\/code><\/div>/g,
      '<div class="phpcode"><pre><code>$1</code></pre></div>'
    );

    // Convert to Markdown
    const markdownContent = turndownService.turndown(htmlContent);

    // Write Markdown file
    fs.writeFileSync(outputPath, markdownContent);

    console.log(`[${index + 1} of ${files.length}] ${inputPath} -> ${outputPath}`);
});

console.log('Conversion complete.');

// serialize all of the markdown files into a single output file `manual.md`:

const markdownFiles = fs.readdirSync(outputDir).filter(file => file.endsWith('.md'));

const manualPath = path.join(__dirname, 'manual.md');

// open the output file for writing:

const manualStream = fs.openSync(manualPath, 'w');

// write the content of each markdown file to the output file:

markdownFiles.forEach((file, index) => {
  const filePath = path.join(outputDir, file);
  const markdownContent = fs.readFileSync(filePath, 'utf8');

  fs.appendFileSync(manualStream, markdownContent + "\n\n");

  console.log(`[${index + 1} of ${markdownFiles.length}] ${filePath} -> ${manualPath}`);
});

// close the output file:

fs.closeSync(manualStream);

console.log('Manual file created.');