qananasikq/dubizzle
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Stabilize worker sync flow

Browse Source
This commit is contained in:
qananasikq
2026-04-17 17:30:28 +03:00
parent 3c3aea8eb3
commit 05d1ffb5ac
9 changed files with 803 additions and 80 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
iaai_scraper/browser/listing.py
View File

@@ -630,18 +630,45 @@ class ListingCollector:
try:
return bool(page.evaluate(
"""
() => Array.from(document.querySelectorAll('a,button,[role="button"]')).some((el) => {
const text = (el.textContent || '').trim().toLowerCase();
const aria = (el.getAttribute('aria-label') || '').trim().toLowerCase();
const title = (el.getAttribute('title') || '').trim().toLowerCase();
const rel = (el.getAttribute('rel') || '').trim().toLowerCase();
const cls = (el.getAttribute('class') || '').trim().toLowerCase();
const disabled = el.hasAttribute('disabled') || el.getAttribute('aria-disabled') === 'true' || cls.includes('disabled');
const visible = !!(el.offsetWidth || el.offsetHeight || el.getClientRects().length);
const hasRightArrowIcon = !!el.querySelector('img[src*="icon-arrow-right"], img[src*="arrow-right"]');
const looksNumericNext = /^\d+$/.test(text);
return !disabled && visible && (rel === 'next' || aria.includes('next') || title.includes('next') || cls.includes('next') || hasRightArrowIcon || looksNumericNext || ['next', '›', '»', '>'].includes(text));
})
() => {
const visible = (el) => !!(el && (el.offsetWidth || el.offsetHeight || el.getClientRects().length));
const controls = Array.from(document.querySelectorAll('a,button,[role="button"]')).filter((el) => {
const cls = (el.getAttribute('class') || '').trim().toLowerCase();
const disabled = el.hasAttribute('disabled') || el.getAttribute('aria-disabled') === 'true' || cls.includes('disabled');
return !disabled && visible(el);
});
const hasExplicitNext = controls.some((el) => {
const text = (el.textContent || '').trim().toLowerCase();
const aria = (el.getAttribute('aria-label') || '').trim().toLowerCase();
const title = (el.getAttribute('title') || '').trim().toLowerCase();
const rel = (el.getAttribute('rel') || '').trim().toLowerCase();
const cls = (el.getAttribute('class') || '').trim().toLowerCase();
const hasRightArrowIcon = !!el.querySelector('img[src*="icon-arrow-right"], img[src*="arrow-right"]');
return rel === 'next' || aria.includes('next') || title.includes('next') || cls.includes('next') || hasRightArrowIcon || ['next', '›', '»', '>'].includes(text);
});
if (hasExplicitNext) {
return true;
}
const current = controls.find((el) => {
const text = (el.textContent || '').trim();
const cls = (el.getAttribute('class') || '').toLowerCase();
const ariaCurrent = (el.getAttribute('aria-current') || '').toLowerCase();
return /^\d+$/.test(text) && (ariaCurrent === 'page' || cls.includes('active') || cls.includes('current') || cls.includes('selected'));
});
if (!current) {
return false;
}
const currentPage = parseInt((current.textContent || '').trim(), 10);
return controls.some((el) => {
const text = (el.textContent || '').trim();
return /^\d+$/.test(text) && parseInt(text, 10) === currentPage + 1;
});
}
"""
))
except Exception: