Function main

apps/sim/scripts/process-docs.ts:185–252

* Main entry point with CLI argument parsing

()

 * Main entry point with CLI argument parsing
 */
async function main() {
  const args = process.argv.slice(2)

  const options: ProcessingOptions = {
    clearExisting: args.includes('--clear'),
    dryRun: args.includes('--dry-run'),
    verbose: args.includes('--verbose'),
  }

  // Parse custom path if provided
  const pathIndex = args.indexOf('--path')
  if (pathIndex !== -1 && args[pathIndex + 1]) {
    options.docsPath = args[pathIndex + 1]
  }

  // Parse custom base URL if provided
  const urlIndex = args.indexOf('--url')
  if (urlIndex !== -1 && args[urlIndex + 1]) {
    options.baseUrl = args[urlIndex + 1]
  }

  // Parse chunk size if provided
  const chunkSizeIndex = args.indexOf('--chunk-size')
  if (chunkSizeIndex !== -1 && args[chunkSizeIndex + 1]) {
    options.chunkSize = Number.parseInt(args[chunkSizeIndex + 1], 10)
  }

  // Show help if requested
  if (args.includes('--help') || args.includes('-h')) {
    console.log(`
📚 Process Documentation Script

Usage: bun run process-docs.ts [options]

By default, processes English (en) documentation only.
Note: Use --clear flag when changing language scope to remove old embeddings.

Options:
  --clear          Clear existing embeddings before processing
  --dry-run        Process and display results without saving to DB
  --verbose        Show detailed output including text previews
  --path <path>    Custom path to docs directory (default: docs/en)
  --url <url>      Custom base URL for links
  --chunk-size <n> Custom chunk size in tokens (default: 1024)
  --help, -h       Show this help message

Examples:
  # Dry run to test chunking (English docs)
  bun run process-docs.ts --dry-run

  # Process and save to database (English docs)
  bun run process-docs.ts

  # Clear existing and reprocess (English docs)
  bun run process-docs.ts --clear

  # Process a different language
  bun run process-docs.ts --path ../../apps/docs/content/docs/es