opencode-browser-tool-installer/scripts/google_es_puertas_rank.mjs

import path from "node:path"
import { BrowserManager } from "../dist/browser/manager.js"

const browser = new BrowserManager(path.resolve("./artifacts"))

const query = "puertas cortafuegos de madera"
const targetDomain = "puertastecnicasbcn.com"
const maxPages = 10

const sleep = (ms) => new Promise((r) => setTimeout(r, ms))

const organicScript = `() => {
  const root = document.querySelector("#search") || document
  const anchors = Array.from(root.querySelectorAll("a[href]"))
  const results = []
  for (const a of anchors) {
    const href = a.getAttribute("href") || ""
    if (!href.startsWith("http")) continue
    if (href.includes("google.") || href.includes("webcache.googleusercontent.com")) continue
    const h3 = a.querySelector("h3")
    if (!h3) continue
    const title = (h3.textContent || "").trim()
    if (!title) continue
    if (!results.some((x) => x.href === href)) {
      results.push({ href, title })
    }
  }
  return results
}`

const run = async () => {
  const out = { pages: [] }

  out.open = await browser.open({
    headless: false,
    width: 1366,
    height: 900,
    browserKind: "testing",
    persistentProfile: true,
    userDataDir: "/home/pancho/.chromium-perfil-google",
    startUrl: "https://www.google.es",
    recordVideo: false,
  })

  out.cookies = await browser.handleConsent("accept").catch((e) => ({ error: e.message }))
  await browser.type('textarea[name="q"]', query)
  await browser.press("Enter")
  await browser.waitFor({ for: "timeout", value: "1200" })

  out.afterSearch = await browser.observe().catch((e) => ({ error: e.message }))
  if ((out.afterSearch.url || "").includes("/sorry/index")) {
    out.humanAttempt = await browser.handleHumanCheck().catch((e) => ({ error: e.message }))
    await browser.waitFor({ for: "timeout", value: "3000" })
    out.afterChallenge = await browser.observe().catch((e) => ({ error: e.message }))
  }

  let found = null
  let globalRank = 0

  for (let pageNo = 1; pageNo <= maxPages; pageNo += 1) {
    await browser.waitFor({ for: "timeout", value: "800" })
    const currentUrl = browser.getState().currentUrl || ""
    if (!currentUrl.includes("/search")) {
      out.pages.push({ page: pageNo, status: "not_on_search", url: currentUrl })
      break
    }

    const extracted = await browser.evaluate(organicScript, {}).catch(() => ({ result: [] }))
    const organic = Array.isArray(extracted.result) ? extracted.result : []
    const idx = organic.findIndex((x) => {
      try {
        return new URL(String(x.href)).hostname.includes(targetDomain)
      } catch {
        return String(x.href).includes(targetDomain)
      }
    })

    out.pages.push({
      page: pageNo,
      url: currentUrl,
      organicCount: organic.length,
      found: idx >= 0,
      positionOnPage: idx >= 0 ? idx + 1 : null,
    })

    if (idx >= 0) {
      globalRank += idx + 1
      found = {
        page: pageNo,
        positionOnPage: idx + 1,
        globalRank,
        href: organic[idx].href,
        title: organic[idx].title,
      }
      break
    }

    globalRank += organic.length

    let moved = false
    for (const sel of ["#pnnext", "a#pnnext", "a[aria-label*='siguiente' i]", "a[aria-label*='next' i]"]) {
      const ex = await browser.query({ selector: sel, mode: "exists" }).catch(() => ({ result: false }))
      if (ex.result === true) {
        await browser.click(sel).catch(() => null)
        moved = true
        break
      }
    }
    if (!moved) {
      const tried = await browser
        .evaluate(
          `() => {
          const nodes = Array.from(document.querySelectorAll("a,button,[role=button]"))
          for (const n of nodes) {
            const t = (n.innerText || "").trim()
            if (!t) continue
            if (t === "Siguiente" || t.includes("Siguiente") || t === "Next" || t.includes("Next")) {
              n.click()
              return { clicked: true, text: t }
            }
          }
          return { clicked: false, text: null }
        }`,
          {},
        )
        .catch(() => ({ result: { clicked: false, text: null } }))
      out.pages[out.pages.length - 1].nextFallback = tried.result
      if (!tried.result?.clicked) {
        break
      }
    }
    await sleep(1400)
  }

  out.result = found || { found: false }

  if (found?.href) {
    await browser.navigate(found.href, "domcontentloaded")
    await browser.waitFor({ for: "timeout", value: "1000" })
    out.destinationObserve = await browser.observe().catch((e) => ({ error: e.message }))
    out.destinationSnapshot = await browser.snapshot({ label: "google-puertas-destination", fullPage: false })
  }

  out.finalObserve = await browser.observe().catch((e) => ({ error: e.message }))
  out.close = await browser.close()
  console.log(JSON.stringify(out, null, 2))
}

run().catch(async (err) => {
  console.error("GOOGLE_PUERTAS_ERROR", err?.message || err)
  try {
    const snap = await browser.snapshot({ label: "google-puertas-error", fullPage: true })
    console.error("ERROR_SNAPSHOT", snap.filePath)
  } catch {
    // ignore
  }
  try {
    await browser.close()
  } catch {
    // ignore
  }
  process.exit(1)
})