puppet-show.js

/**
 * Parallel Browser Bot Runner
 * -------------------------
 * 
 * This script runs multiple browser bot instances in parallel using Node.js worker threads.
 * Each bot instance will:
 * - Use random viewport sizes and user agents
 * - Have 50% chance of scrolling
 * - Have 30% chance of clicking the FAQ link
 * 
 * Usage:
 * 1. Basic run (10 instances, localhost:4321):
 *    node parallel.js
 * 
 * 2. With custom URL:
 *    node parallel.js https://example.com
 *    - or -
 *    TARGET_URL=https://example.com node parallel.js
 *
 * 
 * Notes:
 * - Each instance runs independently with its own random behaviors
 * - The script will report when all instances complete
 * - Sometimes a worker might fail trying to fetch the request, this is tolerable
 *   Since the intent is to study bot blocking capabilities of Plausible
 */

const {
  Worker,
  isMainThread,
  parentPort,
  workerData,
} = require("worker_threads");
const path = require("path");

if (isMainThread) {
  // Parse command line arguments for URL
  const args = process.argv.slice(2);
  if (args.length > 0) {
    process.env.TARGET_URL = args[0];
  }

  // Main thread logic
  const numWorkers = 10;
  const workers = new Set();
  console.log(`Starting ${numWorkers} browser instances...`);

  // Create workers
  for (let i = 0; i < numWorkers; i++) {
    const worker = new Worker(__filename, {
      workerData: { workerId: i },
    });

    workers.add(worker);

    worker.on("error", (error) => {
      console.error(`Worker ${i} error:`, error);
    });

    worker.on("exit", (code) => {
      workers.delete(worker);
      console.log(`Worker ${i} exited with code ${code}`);

      if (workers.size === 0) {
        console.log("All workers completed!");
      }
    });

    worker.on("message", (message) => {
      console.log(`Worker ${i}:`, message);
    });
  }
} else {
  // Worker thread logic - import and run the main bot script
  const { workerId } = workerData;
  const visitPage = require("./index.js").visitPage;

  // Get URL from environment variable or use default
  const targetUrl = process.env.TARGET_URL || "http://localhost:4321";

  async function runWorker() {
    try {
      console.log(`Worker ${workerId} starting...`);
      await visitPage(targetUrl);
      console.log(`Worker ${workerId} finished successfully`);
      process.exit(0);
    } catch (error) {
      console.error(`Worker ${workerId} failed:`, error);
      process.exit(1);
    }
  }

  runWorker();
}