wip: skyslope scraper — fix name extraction via body text parsing, preview+download flow ready

teressa-copeland-homes/scripts/debug-dom.ts

@@ -0,0 +1,43 @@
+import { chromium } from 'playwright';
+import { config } from 'dotenv';
+import path from 'path';
+config({ path: path.resolve(process.cwd(), '.env.local') });
+
+(async () => {
+  const savedState = await import('node:fs/promises').then(fs =>
+    fs.readFile(path.resolve(process.cwd(), 'scripts/.ure-session.json'), 'utf8').then(JSON.parse).catch(() => null)
+  );
+  const browser = await chromium.launch({ headless: true });
+  const context = await browser.newContext({ storageState: savedState ?? undefined });
+  const page = await browser.newPage();
+  await page.goto('https://www.utahrealestate.com/sso/connect/client/skyslope', { waitUntil: 'domcontentloaded' });
+  const [newPage] = await Promise.all([
+    context.waitForEvent('page', { timeout: 10_000 }).catch(() => null),
+    Promise.resolve(),
+  ]);
+  const activePage = newPage ?? page;
+  await activePage.waitForLoadState('domcontentloaded');
+  await activePage.waitForTimeout(3000);
+
+  // Navigate to browse-libraries
+  await activePage.goto('https://forms.skyslope.com/browse-libraries', { waitUntil: 'domcontentloaded' });
+  await activePage.waitForTimeout(4000);
+
+  // Inspect the first Add button's ancestors
+  const result = await activePage.evaluate(() => {
+    const addBtns = Array.from(document.querySelectorAll('button')).filter(b => b.textContent?.trim() === 'Add');
+    if (addBtns.length === 0) return 'No Add buttons found';
+    const btn = addBtns[0];
+    // Walk up the tree and print each ancestor's tag, class, and text
+    const info: string[] = [];
+    let el: Element | null = btn;
+    for (let i = 0; i < 6; i++) {
+      if (!el) break;
+      info.push(`[${i}] <${el.tagName.toLowerCase()} class="${el.className}"> text="${el.textContent?.replace(/\s+/g,' ').trim().slice(0,80)}"`);
+      el = el.parentElement;
+    }
+    return info.join('\n');
+  });
+  console.log('Add button ancestry:\n', result);
+  await browser.close();
+})();

teressa-copeland-homes/scripts/debug-dom2.ts

@@ -0,0 +1,47 @@
+import { chromium } from 'playwright';
+import { config } from 'dotenv';
+import path from 'path';
+config({ path: path.resolve(process.cwd(), '.env.local') });
+
+(async () => {
+  const savedState = await import('node:fs/promises').then(fs =>
+    fs.readFile(path.resolve(process.cwd(), 'scripts/.ure-session.json'), 'utf8').then(JSON.parse).catch(() => null)
+  );
+  const browser = await chromium.launch({ headless: false });
+  const context = await browser.newContext({ storageState: savedState ?? undefined, acceptDownloads: true });
+  const page = await context.newPage();
+  await page.goto('https://www.utahrealestate.com/sso/connect/client/skyslope', { waitUntil: 'domcontentloaded' });
+  const newPage = await context.waitForEvent('page', { timeout: 10_000 }).catch(() => null);
+  const activePage = newPage ?? page;
+  await activePage.waitForLoadState('domcontentloaded');
+  await activePage.waitForTimeout(3000);
+  await activePage.goto('https://forms.skyslope.com/browse-libraries', { waitUntil: 'domcontentloaded' });
+  await activePage.waitForTimeout(5000);
+
+  // Count all buttons and get text of first 5
+  const result = await activePage.evaluate(() => {
+    const allBtns = Array.from(document.querySelectorAll('button'));
+    const btnTexts = allBtns.slice(0, 10).map(b => `"${b.textContent?.trim()}"`);
+    // Also get first row-like elements
+    const rows = Array.from(document.querySelectorAll('tr, [class*="row"], [class*="item"]'))
+      .slice(0, 3)
+      .map(el => `<${el.tagName} class="${el.className.toString().slice(0,50)}"> "${el.textContent?.replace(/\s+/g,' ').trim().slice(0,60)}"`);
+    return { totalBtns: allBtns.length, btnTexts, rows };
+  });
+  console.log(JSON.stringify(result, null, 2));
+
+  // Try clicking first form name and screenshot
+  const firstFormName = activePage.locator('text="Acknowledgement of Third Party Approval Addendum to REPC - UAR"').first();
+  if (await firstFormName.isVisible({ timeout: 3000 }).catch(() => false)) {
+    console.log('Clicking first form name...');
+    await firstFormName.click();
+    await activePage.waitForTimeout(3000);
+    await activePage.screenshot({ path: 'scripts/debug-after-click.png' });
+    console.log('Screenshot saved: debug-after-click.png');
+    // Get all buttons visible now
+    const btns = await activePage.locator('button').allTextContents();
+    console.log('Buttons after click:', btns.slice(0, 15));
+  }
+
+  await browser.close();
+})();

teressa-copeland-homes/scripts/debug-inspect.ts

@@ -0,0 +1,30 @@
+import { chromium } from 'playwright';
+import { config } from 'dotenv';
+import path from 'path';
+config({ path: path.resolve(process.cwd(), '.env.local') });
+
+(async () => {
+  const browser = await chromium.launch({ headless: true });
+  const page = await browser.newPage();
+  await page.goto('https://skyslope.com/forms-login/');
+  await page.waitForLoadState('networkidle');
+
+  const inputs = await page.locator('input').all();
+  console.log('Input count:', inputs.length);
+  for (const input of inputs) {
+    const attrs = await input.evaluate((el: HTMLInputElement) => ({
+      type: el.type, name: el.name, id: el.id, placeholder: el.placeholder
+    }));
+    console.log('Input:', attrs);
+  }
+
+  const links = await page.locator('a[href*="login"], a[href*="signin"], a[href*="forms.skyslope"]').all();
+  for (const link of links) {
+    const text = await link.textContent();
+    const href = await link.getAttribute('href');
+    console.log('Link:', text?.trim().slice(0, 60), '→', href);
+  }
+
+  await page.screenshot({ path: 'scripts/debug-inspect.png', fullPage: false });
+  await browser.close();
+})();

teressa-copeland-homes/scripts/debug-inspect2.ts

@@ -0,0 +1,28 @@
+import { chromium } from 'playwright';
+import { config } from 'dotenv';
+import path from 'path';
+config({ path: path.resolve(process.cwd(), '.env.local') });
+
+(async () => {
+  const browser = await chromium.launch({ headless: true });
+  const page = await browser.newPage();
+
+  // Try the direct forms app URL first
+  await page.goto('https://forms.skyslope.com', { waitUntil: 'domcontentloaded', timeout: 30_000 });
+  await page.waitForTimeout(3000);
+
+  console.log('URL after goto forms.skyslope.com:', page.url());
+  await page.screenshot({ path: 'scripts/debug-forms-home.png' });
+
+  // Look for any login/sign-in link
+  const loginLinks = await page.locator('a, button').all();
+  for (const el of loginLinks) {
+    const text = await el.textContent().catch(() => '');
+    const href = await el.getAttribute('href').catch(() => '');
+    if (/login|sign.?in|log.?in|get started|access/i.test(text + href)) {
+      console.log('Found login element:', text?.trim().slice(0,60), href?.slice(0,80));
+    }
+  }
+
+  await browser.close();
+})();

teressa-copeland-homes/scripts/debug-inspect3.ts

@@ -0,0 +1,38 @@
+import { chromium } from 'playwright';
+import { config } from 'dotenv';
+import path from 'path';
+config({ path: path.resolve(process.cwd(), '.env.local') });
+
+(async () => {
+  const browser = await chromium.launch({ headless: false });
+  const page = await browser.newPage();
+
+  await page.goto('https://forms.skyslope.com/?tab=all', { waitUntil: 'domcontentloaded', timeout: 30_000 });
+  // Wait for JS to render
+  await page.waitForTimeout(5000);
+
+  console.log('URL:', page.url());
+  await page.screenshot({ path: 'scripts/debug-forms-rendered.png', fullPage: false });
+  console.log('Screenshot saved');
+
+  // Look for form items
+  const bodyText = await page.locator('body').innerText().catch(() => '');
+  console.log('Page text (first 500 chars):', bodyText.slice(0, 500));
+
+  // Look for any list items or cards
+  const items = await page.locator('[class*="form"], [class*="Form"], [class*="item"], [class*="card"]').count();
+  console.log('Form-like elements:', items);
+
+  // Look for the NRDS/authorization prompt
+  const hasNRDS = bodyText.toLowerCase().includes('nrds') || bodyText.toLowerCase().includes('authorization');
+  console.log('Has NRDS/auth prompt:', hasNRDS);
+
+  // Check if we're on a tab with forms
+  const tabs = await page.locator('[role="tab"]').all();
+  for (const tab of tabs) {
+    console.log('Tab:', await tab.textContent());
+  }
+
+  await page.waitForTimeout(2000);
+  await browser.close();
+})();

teressa-copeland-homes/scripts/scrape-skyslope-forms.ts

@@ -1,230 +1,385 @@
 /**
- * SkySlope Forms Scraper
+ * SkySlope Forms Scraper — via Utah Real Estate SSO
  *
- * Downloads all PDFs from your SkySlope form libraries into seeds/forms/.
- * Run: DOTENV_CONFIG_PATH=.env.local npx tsx scripts/scrape-skyslope-forms.ts
+ * Flow: utahrealestate.com login → Forms → MLS Forms → SkySlope → download all library PDFs
  *
- * Credentials are read from env vars:
- *   SKYSLOPE_LAST_NAME   — last name on NAR records (e.g. Copeland)
- *   SKYSLOPE_NRDS_ID     — 9-digit NRDS ID (e.g. 837075029)
+ * Run: npm run scrape:forms
+ *
+ * Credentials read from .env.local:
+ *   URE_USERNAME    — Utah Real Estate username
+ *   URE_PASSWORD    — Utah Real Estate password
  */
 
 import { chromium } from 'playwright';
 import * as fs from 'node:fs/promises';
 import * as path from 'node:path';
+import { config } from 'dotenv';
 
-const LAST_NAME = process.env.SKYSLOPE_LAST_NAME;
-const NRDS_ID = process.env.SKYSLOPE_NRDS_ID;
+config({ path: path.resolve(process.cwd(), '.env.local') });
+
+const URE_USERNAME = process.env.URE_USERNAME;
+const URE_PASSWORD = process.env.URE_PASSWORD;
 const OUTPUT_DIR = path.resolve(process.cwd(), 'seeds/forms');
-const BASE_URL = 'https://forms.skyslope.com';
 
-if (!LAST_NAME || !NRDS_ID) {
-  console.error('Missing required env vars: SKYSLOPE_LAST_NAME, SKYSLOPE_NRDS_ID');
+if (!URE_USERNAME || !URE_PASSWORD) {
+  console.error('Missing required env vars: URE_USERNAME, URE_PASSWORD');
   process.exit(1);
 }
 
 async function main() {
   await fs.mkdir(OUTPUT_DIR, { recursive: true });
 
-  const browser = await chromium.launch({ headless: false }); // visible so you can watch/intervene
-  const context = await browser.newContext({ acceptDownloads: true });
+  const browser = await chromium.launch({ headless: false });
+  const savedState = await fs.readFile(path.resolve(process.cwd(), 'scripts/.ure-session.json'), 'utf8').then(JSON.parse).catch(() => null);
+  const context = await browser.newContext({
+    acceptDownloads: true,
+    storageState: savedState ?? undefined,
+  });
   const page = await context.newPage();
 
-  console.log('Navigating to SkySlope Forms...');
-  await page.goto(`${BASE_URL}/welcome/authorization`);
+  // ── Step 1: Login to Utah Real Estate ─────────────────────────────────────
+  console.log('Navigating to utahrealestate.com...');
+  await page.goto('https://www.utahrealestate.com/auth/login', { waitUntil: 'domcontentloaded', timeout: 30_000 });
+  await page.waitForTimeout(2000);
 
-  // Fill NRDS authorization form
-  console.log('Filling NRDS credentials...');
-  await page.getByLabel(/last name/i).fill(LAST_NAME!);
-  await page.getByLabel(/nrds/i).fill(NRDS_ID!);
-  await page.getByRole('button', { name: /next/i }).click();
+  console.log(`Login page URL: ${page.url()}`);
+  await page.screenshot({ path: 'scripts/debug-ure-login.png' });
 
-  // Wait for the main forms library page to load
-  console.log('Waiting for forms library...');
-  await page.waitForURL(/forms\.skyslope\.com\/(?!welcome)/, { timeout: 30_000 });
-  await page.waitForLoadState('networkidle');
+  const COOKIE_FILE = path.resolve(process.cwd(), 'scripts/.ure-session.json');
 
-  // Find all library tabs/sections
-  const libraries = await page.$$eval(
-    '[class*="library"], [class*="Library"], [data-testid*="library"]',
-    els => els.map(el => ({ text: el.textContent?.trim(), id: el.id }))
-  );
-  console.log(`Found ${libraries.length} library elements`);
+  // Only fill login form if we're actually on the login page
+  if (page.url().includes('/auth/login') || page.url().includes('/login')) {
+    const usernameInput = page.locator('input[name="username"], input[name="user"], input[id*="user"], input[placeholder*="user" i]').first();
+    const passwordInput = page.locator('input[type="password"]').first();
 
-  // Collect all form links across all tabs
+    await usernameInput.waitFor({ timeout: 10_000 });
+    await usernameInput.fill(URE_USERNAME!);
+    await passwordInput.fill(URE_PASSWORD!);
+
+    console.log('Credentials filled, submitting...');
+    await page.locator('button[type="submit"], input[type="submit"], button:has-text("Login"), button:has-text("Sign In"), button:has-text("Log In")').first().click();
+    await page.waitForLoadState('domcontentloaded');
+    await page.waitForTimeout(3000);
+  } else {
+    console.log('Already logged in (session restored).');
+  }
+
+  console.log(`After login URL: ${page.url()}`);
+  await page.screenshot({ path: 'scripts/debug-ure-after-login.png' });
+
+  // ── Step 1b: Handle 2FA if present ────────────────────────────────────────
+  const pageText = await page.locator('body').innerText().catch(() => '');
+  if (pageText.includes('verification code') || pageText.includes('one-time')) {
+    console.log('\n⚡ 2FA detected — please complete it in the browser window.');
+    console.log('   (Select Text/Email, enter the code, and click Submit)');
+    console.log('   Waiting up to 2 minutes for you to finish...\n');
+
+    // Poll every 2s until 2FA page is gone (up to 2 minutes)
+    const deadline = Date.now() + 120_000;
+    while (Date.now() < deadline) {
+      await page.waitForTimeout(2000);
+      const text = await page.locator('body').innerText().catch(() => '');
+      if (!text.includes('verification code') && !text.includes('one-time')) break;
+      process.stdout.write('.');
+    }
+    console.log();
+    await page.waitForLoadState('domcontentloaded');
+    await page.waitForTimeout(2000);
+    console.log(`After 2FA URL: ${page.url()}`);
+
+    // Save session so we skip 2FA next time
+    await context.storageState({ path: COOKIE_FILE });
+    console.log('Session saved — 2FA will be skipped on next run.');
+  }
+
+  // ── Step 2: Navigate directly to SkySlope SSO URL ────────────────────────
+  console.log('Navigating to SkySlope via SSO...');
+
+  const [newPage] = await Promise.all([
+    context.waitForEvent('page', { timeout: 15_000 }).catch(() => null),
+    page.goto('https://www.utahrealestate.com/sso/connect/client/skyslope', { waitUntil: 'domcontentloaded' }),
+  ]);
+
+  // The SSO link opens a new tab
+  const activePage = newPage ?? page;
+  await activePage.waitForLoadState('domcontentloaded');
+  await activePage.waitForTimeout(5000);
+  console.log(`SkySlope URL: ${activePage.url()}`);
+  await activePage.screenshot({ path: 'scripts/debug-skyslope-landing.png' });
+
+  await downloadAllForms(activePage, context, OUTPUT_DIR);
+  if (newPage) await newPage.close();
+
+  await browser.close();
+  console.log('\nDone.');
+}
+
+async function downloadAllForms(
+  page: import('playwright').Page,
+  context: import('playwright').BrowserContext,
+  outputDir: string
+) {
   const downloaded: string[] = [];
   const failed: string[] = [];
 
-  // Navigate to the forms list — look for a tab or section with all forms
-  // Try clicking "All" tab if it exists
-  const allTab = page.getByRole('tab', { name: /all/i }).first();
-  if (await allTab.isVisible().catch(() => false)) {
-    await allTab.click();
-    await page.waitForLoadState('networkidle');
+  // Handle NRDS auth if it appears
+  await handleNRDSAuth(page);
+
+  // Wait for forms library to load
+  await page.waitForTimeout(4000);
+  console.log(`Forms library URL: ${page.url()}`);
+
+  // Navigate to Browse Libraries
+  console.log('Clicking Browse Libraries...');
+  const browseLink = page.locator('a:has-text("Browse Libraries"), a[href*="libraries"], nav a:has-text("Libraries")').first();
+  if (await browseLink.isVisible({ timeout: 5_000 }).catch(() => false)) {
+    await browseLink.click();
+    await page.waitForLoadState('domcontentloaded');
+    await page.waitForTimeout(3000);
+  } else {
+    await page.goto('https://forms.skyslope.com/libraries', { waitUntil: 'domcontentloaded' });
+    await page.waitForTimeout(3000);
   }
 
-  // Scrape all visible form items across all library sections
-  await downloadFormsOnPage(page, OUTPUT_DIR, downloaded, failed);
+  console.log(`Libraries URL: ${page.url()}`);
+  await page.screenshot({ path: 'scripts/debug-libraries-page.png' });
 
-  // Also try each library tab
-  const tabs = await page.getByRole('tab').all();
-  for (const tab of tabs) {
-    const tabName = await tab.textContent();
-    if (!tabName) continue;
-    console.log(`\nChecking library tab: ${tabName.trim()}`);
-    await tab.click();
-    await page.waitForLoadState('networkidle');
-    await downloadFormsOnPage(page, OUTPUT_DIR, downloaded, failed);
+  const bodyText2 = await page.locator('body').innerText().catch(() => '');
+  console.log('Libraries page text (first 600):', bodyText2.slice(0, 600));
+
+  // Find all library cards/links
+  const libraryLinks = await page.locator(
+    'a[href*="/library/"], a[href*="/libraries/"], [class*="library-card"], [class*="libraryCard"]'
+  ).all();
+  console.log(`Found ${libraryLinks.length} library links`);
+
+  if (libraryLinks.length > 0) {
+    const libraryHrefs: string[] = [];
+    for (const link of libraryLinks) {
+      const href = await link.getAttribute('href').catch(() => '');
+      const name = await link.textContent().catch(() => '');
+      if (href) {
+        libraryHrefs.push(href);
+        console.log(`  Library: ${name?.trim().slice(0, 50)} → ${href}`);
+      }
     }
 
-  console.log(`\n✓ Downloaded ${downloaded.length} forms to ${OUTPUT_DIR}`);
-  if (failed.length > 0) {
-    console.log(`✗ Failed: ${failed.join(', ')}`);
+    for (const href of libraryHrefs) {
+      const url = href.startsWith('http') ? href : `https://forms.skyslope.com${href}`;
+      console.log(`\n── Opening library: ${url} ──`);
+      await page.goto(url, { waitUntil: 'domcontentloaded' });
+      await page.waitForTimeout(4000);
+      await page.screenshot({ path: `scripts/debug-library-${Date.now()}.png` });
+      await downloadFormsInView(page, context, outputDir, downloaded, failed);
+    }
+  } else {
+    // Fallback: libraries might be listed as clickable items
+    const bodyText = await page.locator('body').innerText().catch(() => '');
+    const libraryNames = ['Data Forms - URE', 'Utah Association of Realtors', 'Utah CCIM'];
+
+    for (const libName of libraryNames) {
+      const libLink = page.locator(`a:has-text("${libName}"), button:has-text("${libName}")`).first();
+      if (await libLink.isVisible({ timeout: 3_000 }).catch(() => false)) {
+        console.log(`\n── Library: ${libName} ──`);
+        await libLink.click();
+        await page.waitForLoadState('domcontentloaded');
+        await page.waitForTimeout(4000);
+        await downloadFormsInView(page, context, outputDir, downloaded, failed);
+        await page.goBack();
+        await page.waitForTimeout(2000);
+      }
     }
 
-  await browser.close();
+    if (libraryLinks.length === 0) {
+      // We're already on the all-forms page — download directly
+      console.log('All forms visible on current page — downloading...');
+      await downloadFormsInView(page, context, outputDir, downloaded, failed);
+    }
+  }
+
+  console.log(`\n✓ Downloaded ${downloaded.length} forms`);
+  if (failed.length > 0) console.log(`✗ Failed: ${failed.length} — ${failed.join(', ')}`);
 }
 
-async function downloadFormsOnPage(
+async function handleNRDSAuth(page: import('playwright').Page) {
+  const LAST_NAME = process.env.SKYSLOPE_LAST_NAME || 'Copeland';
+  const NRDS_ID = process.env.SKYSLOPE_NRDS_ID || '837075029';
+
+  // Check if NRDS auth page appears
+  const isNRDS = await page.locator('input[placeholder*="last" i], input[placeholder*="nrds" i]').first().isVisible({ timeout: 5_000 }).catch(() => false);
+  if (!isNRDS) return;
+
+  console.log('NRDS authorization required — filling...');
+  const inputs = await page.locator('input').all();
+  if (inputs.length >= 2) {
+    await inputs[0].fill(LAST_NAME);
+    await inputs[1].fill(NRDS_ID);
+    await page.locator('button:has-text("Next"), button[type="submit"]').first().click();
+    await page.waitForLoadState('domcontentloaded');
+    await page.waitForTimeout(3000);
+    console.log(`After NRDS URL: ${page.url()}`);
+  }
+}
+
+async function downloadFormsInView(
   page: import('playwright').Page,
+  context: import('playwright').BrowserContext,
   outputDir: string,
   downloaded: string[],
   failed: string[]
 ) {
-  // Look for form items — SkySlope renders them as list rows with a download or view button
-  // Strategy: intercept PDF responses by clicking each form's download/view button
-  const formRows = await page.$$('[class*="form-item"], [class*="formItem"], [class*="form-row"], [role="row"], [class*="list-item"]');
+  // Flow: click form name → preview opens → click Download button → save file
 
-  if (formRows.length === 0) {
-    // Fallback: look for any link or button that mentions PDF or has a form name
-    console.log('  No form rows found with primary selectors, trying fallback...');
-    await downloadViaLinks(page, outputDir, downloaded, failed);
+  // Extract form names from the page body text — the list renders as "Name\nAdd\nName\nAdd..."
+  const bodyText = await page.locator('body').innerText().catch(() => '');
+  const lines = bodyText.split('\n').map(l => l.trim()).filter(l => l.length > 3);
+  const formNames: string[] = [];
+  for (let i = 0; i < lines.length; i++) {
+    if (lines[i] === 'Add' && i > 0 && lines[i - 1] !== 'Add' && lines[i - 1].length > 3) {
+      formNames.push(lines[i - 1]);
+    }
+  }
+  const names = [...new Set(formNames)];
+
+  console.log(`  Found ${names.length} forms to download`);
+  if (names.length === 0) {
+    await page.screenshot({ path: `scripts/debug-no-forms-${Date.now()}.png` });
     return;
   }
 
-  console.log(`  Found ${formRows.length} form rows`);
-
-  for (const row of formRows) {
-    const name = await row.$eval(
-      '[class*="name"], [class*="title"], span, td',
-      el => el.textContent?.trim()
-    ).catch(() => null);
-
-    if (!name) continue;
-
-    // Skip already downloaded
-    const sanitized = name.replace(/[^a-z0-9 ._-]/gi, ' ').replace(/\s+/g, ' ').trim();
+  for (const formName of names) {
+    const sanitized = formName.replace(/[^a-z0-9 ._-]/gi, ' ').replace(/\s+/g, ' ').trim().slice(0, 100);
     const destPath = path.join(outputDir, `${sanitized}.pdf`);
 
-    try {
-      await fs.access(destPath);
-      console.log(`  ⊙ Skipping (already exists): ${sanitized}.pdf`);
+    // Skip already downloaded
+    try { await fs.access(destPath); process.stdout.write(`  ⊙ skip: ${sanitized}\n`); continue; } catch { /* proceed */ }
+
+    // Click the form name to open preview
+    const nameEl = page.locator(`text="${formName}"`).first();
+    if (!await nameEl.isVisible({ timeout: 3_000 }).catch(() => false)) {
+      process.stdout.write(`  ⚠ not found: ${sanitized}\n`);
+      failed.push(sanitized);
       continue;
-    } catch {
-      // file doesn't exist, proceed
     }
 
-    // Find the download/view button within this row
-    const downloadBtn = await row.$('button[aria-label*="download" i], button[title*="download" i], [class*="download"]');
-    const viewBtn = await row.$('button[aria-label*="view" i], button[title*="view" i], [class*="view"]');
-    const btn = downloadBtn ?? viewBtn;
+    await nameEl.click().catch(() => {});
+    // Wait for preview/modal to appear (up to 5s)
+    await page.waitForTimeout(2000);
 
-    if (!btn) {
-      console.log(`  ⚠ No download button found for: ${name}`);
+    // Click Download button in the preview
+    const downloadBtn = page.locator(
+      'button:has-text("Download"), a:has-text("Download"), [aria-label*="download" i], button[title*="download" i]'
+    ).first();
+
+    if (!await downloadBtn.isVisible({ timeout: 5_000 }).catch(() => false)) {
+      process.stdout.write(`  ⚠ no Download button found for: ${sanitized}\n`);
+      await page.screenshot({ path: `scripts/debug-no-download-btn-${Date.now()}.png` });
+      await page.keyboard.press('Escape').catch(() => {});
+      await page.waitForTimeout(500);
+      failed.push(sanitized);
       continue;
     }
 
     try {
-      // Intercept the PDF download
       const [download] = await Promise.all([
-        page.waitForEvent('download', { timeout: 15_000 }),
-        btn.click(),
-      ]).catch(async () => {
-        // If no download event, try waiting for a new page/tab with PDF
-        const newPagePromise = page.context().waitForEvent('page', { timeout: 10_000 });
-        await btn.click();
-        const newPage = await newPagePromise;
-        await newPage.waitForLoadState();
-        const url = newPage.url();
-        if (url.endsWith('.pdf') || url.includes('/pdf')) {
-          const pdfBuffer = await newPage.evaluate(async (pdfUrl) => {
-            const res = await fetch(pdfUrl);
-            const buf = await res.arrayBuffer();
-            return Array.from(new Uint8Array(buf));
-          }, url);
-          await fs.writeFile(destPath, Buffer.from(pdfBuffer));
-          await newPage.close();
-          return [null];
-        }
-        await newPage.close();
-        return [null];
-      });
-
-      if (download && typeof download === 'object' && 'saveAs' in download) {
-        await (download as import('playwright').Download).saveAs(destPath);
-        console.log(`  ✓ ${sanitized}.pdf`);
+        page.waitForEvent('download', { timeout: 20_000 }),
+        downloadBtn.click(),
+      ]);
+      await download.saveAs(destPath);
+      process.stdout.write(`  ✓ ${sanitized}.pdf\n`);
       downloaded.push(sanitized);
-      } else {
-        console.log(`  ✓ ${sanitized}.pdf (via page)`);
-        downloaded.push(sanitized);
-      }
     } catch (err) {
-      console.log(`  ✗ Failed: ${name} — ${(err as Error).message}`);
-      failed.push(name);
+      process.stdout.write(`  ✗ download failed: ${sanitized} — ${(err as Error).message.slice(0, 60)}\n`);
+      failed.push(sanitized);
     }
+
+    // Close preview and return to list
+    await page.keyboard.press('Escape').catch(() => {});
+    await page.waitForTimeout(800);
   }
 }
 
-async function downloadViaLinks(
+async function downloadViaTextRows(
   page: import('playwright').Page,
   outputDir: string,
   downloaded: string[],
   failed: string[]
 ) {
-  // Intercept all PDF network requests triggered by clicking form items
-  const pdfResponses: { url: string; name: string }[] = [];
-
-  page.on('response', async (response) => {
-    const contentType = response.headers()['content-type'] ?? '';
-    if (contentType.includes('pdf') && response.status() === 200) {
-      const url = response.url();
-      const name = path.basename(new URL(url).pathname, '.pdf') || `form-${Date.now()}`;
-      pdfResponses.push({ url, name });
-    }
+  // Legacy fallback — kept for safety but downloadFormsInView handles all cases now
+  console.log('  (downloadViaTextRows called — should not reach here normally)');
+  const rows = await page.evaluate(() => {
+    const candidates = Array.from(document.querySelectorAll('tr, li, [class*="row"], [class*="item"]'));
+    return candidates
+      .map(el => ({
+        text: el.textContent?.replace(/\s+/g, ' ').trim() ?? '',
+        hasAdd: el.textContent?.includes('Add') ?? false,
+      }))
+      .filter(r => r.hasAdd && r.text.length > 10)
+      .map(r => r.text.replace(/\s*Add\s*$/, '').trim());
   });
 
-  // Click each item that looks like a form
-  const items = await page.$$('li, [role="listitem"], [class*="item"]');
-  for (const item of items.slice(0, 100)) { // cap at 100 to avoid infinite loops
-    const text = await item.textContent().catch(() => '');
-    if (!text?.trim()) continue;
+  console.log(`  Found ${rows.length} form names via text extraction`);
 
-    const before = pdfResponses.length;
-    await item.click({ timeout: 3_000 }).catch(() => {});
-    await page.waitForTimeout(1_000);
-
-    if (pdfResponses.length > before) {
-      const { url, name } = pdfResponses[pdfResponses.length - 1];
-      const sanitized = text.replace(/[^a-z0-9 ._-]/gi, ' ').replace(/\s+/g, ' ').trim().slice(0, 80);
+  for (const formName of rows) {
+    if (!formName || formName.length < 3) continue;
+    const sanitized = formName.replace(/[^a-z0-9 ._-]/gi, ' ').replace(/\s+/g, ' ').trim().slice(0, 100);
     const destPath = path.join(outputDir, `${sanitized}.pdf`);
+
+    try { await fs.access(destPath); process.stdout.write(`  ⊙ skip: ${sanitized}\n`); continue; } catch { /* proceed */ }
+
+    const el = page.locator(`text="${formName}"`).first();
+    if (!await el.isVisible({ timeout: 2_000 }).catch(() => false)) {
+      process.stdout.write(`  ⚠ not visible: ${sanitized}\n`);
+      continue;
+    }
+
+    const pdfUrl = await interceptPdfOnClick(page, el);
+    if (pdfUrl) {
       try {
-        const res = await page.evaluate(async (pdfUrl) => {
-          const r = await fetch(pdfUrl);
-          const buf = await r.arrayBuffer();
-          return Array.from(new Uint8Array(buf));
-        }, url);
-        await fs.writeFile(destPath, Buffer.from(res));
-        console.log(`  ✓ ${sanitized}.pdf`);
+        const buf = await page.evaluate(async (url) => {
+          const r = await fetch(url, { credentials: 'include' });
+          const ab = await r.arrayBuffer();
+          return Array.from(new Uint8Array(ab));
+        }, pdfUrl);
+        await fs.writeFile(destPath, Buffer.from(buf));
+        process.stdout.write(`  ✓ ${sanitized}.pdf\n`);
         downloaded.push(sanitized);
-      } catch (err) {
-        console.log(`  ✗ Failed saving ${name}`);
-        failed.push(name);
+      } catch {
+        process.stdout.write(`  ✗ ${sanitized}\n`);
+        failed.push(sanitized);
       }
+    } else {
+      process.stdout.write(`  ⚠ no PDF: ${sanitized}\n`);
+      failed.push(sanitized);
     }
+
+    await page.keyboard.press('Escape').catch(() => {});
+    await page.waitForTimeout(500);
   }
 }
 
+async function interceptPdfOnClick(
+  page: import('playwright').Page,
+  row: import('playwright').Locator
+): Promise<string | null> {
+  return new Promise(async (resolve) => {
+    let resolved = false;
+    const handler = (response: import('playwright').Response) => {
+      const ct = response.headers()['content-type'] ?? '';
+      if (!resolved && (ct.includes('pdf') || response.url().endsWith('.pdf'))) {
+        resolved = true;
+        resolve(response.url());
+      }
+    };
+    page.on('response', handler);
+    await row.click({ timeout: 3_000 }).catch(() => {});
+    await page.waitForTimeout(2000);
+    page.off('response', handler);
+    if (!resolved) resolve(null);
+  });
+}
+
 main().catch(err => {
-  console.error('Fatal:', err);
+  console.error('Fatal:', err.message);
   process.exit(1);
 });