fetch-page.mjs

#!/usr/bin/env node
/**
 * Fetch a web page using Plasmate and print the semantic content.
 */
import { execSync } from "node:child_process";

const url = process.argv[2] || "https://news.ycombinator.com";

try {
  const output = execSync(`plasmate fetch "${url}"`, {
    encoding: "utf-8",
    stdio: ["pipe", "pipe", "pipe"],
  });

  const som = JSON.parse(output);

  console.log(`Title: ${som.title ?? "N/A"}`);
  console.log(`Language: ${som.lang ?? "N/A"}`);
  console.log(`Regions: ${(som.regions ?? []).length}`);
  console.log();

  for (const region of som.regions ?? []) {
    console.log(`[${region.role}] ${region.id}`);
    for (const el of (region.elements ?? []).slice(0, 5)) {
      const text = (el.text ?? "").slice(0, 80);
      console.log(`  ${el.role}: ${text}`);
    }
  }
} catch (err) {
  console.error(`Error: ${err.message}`);
  process.exit(1);
}