vikarti.anatra/main
1
0
Fork 0
mirror of https://github.com/manualdousuario/marreta.git synced 2025-09-01 18:20:22 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

removido logs e liberado selenium como ultimo fallback

Browse source
This commit is contained in:
Renan Bernordi 2024-12-24 13:35:20 -03:00
parent 775b6bdd3f
commit c9e59dfce2
3 changed files with 17 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
TESTED_URLS.md
View file

@ -39,6 +39,8 @@ https://oantagonista.com.br/brasil/lewandowski-insiste-na-pec-da-seguranca/
https://jornaldebrasilia.com.br/noticias/politica-e-poder/lula-aguarda-pt-para-troca-em-pastas-chefiadas-por-petistas-em-reforma-ministerial/
https://opopular.com.br/cidades/ex-secretario-de-saude-de-goiania-deixa-hospital-e-volta-para-a-cadeia-1.3207162
https://www.cartacapital.com.br/politica/surpresa-natalina/
https://seucreditodigital.com.br/123milhas-devera-apresentar-plano-de-recuperacao-ainda-este-mes/
https://www.matinaljornalismo.com.br/matinal/reportagem-matinal/vazao-guaiba-porto-alegre/
## Internacional
https://www.nytimes.com/2024/11/20/us/politics/matt-gaetz-venmo-payments-sex.html

1
app/data/blocked_domains.php
View file

@ -16,6 +16,7 @@ return [
'utppublishing.com',
'chronicle.com',
'nexojornal.com',
'nexojornal.com.br',
'lesoir.be',
'weeklytimesnow.com.au',
'barrons.com',

17
app/inc/URLAnalyzer.php
View file

@ -123,7 +123,6 @@ class URLAnalyzer
$host = preg_replace('/^www\./', '', $host);
if (in_array($host, BLOCKED_DOMAINS)) {
Logger::getInstance()->log($cleanUrl, 'BLOCKED_DOMAIN');
throw new Exception('Este domínio está bloqueado para extração.');
}
@ -153,7 +152,7 @@ class URLAnalyzer
return $processedContent;
}
} catch (Exception $e) {
Logger::getInstance()->log($cleanUrl, 'DIRECT_FETCH_ERROR', $e->getMessage());
error_log("DIRECT_FETCH_ERROR: " . $e->getMessage());
}
// 6. Tenta buscar do Wayback Machine como fallback
@ -165,7 +164,19 @@ class URLAnalyzer
return $processedContent;
}
} catch (Exception $e) {
Logger::getInstance()->log($cleanUrl, 'WAYBACK_FETCH_ERROR', $e->getMessage());
error_log("WAYBACK_FETCH_ERROR: " . $e->getMessage());
}
// 7. Tenta buscar com Selenium como fallback
try {
$content = $this->fetchFromSelenium($cleanUrl, 'firefox');
if (!empty($content)) {
$processedContent = $this->processContent($content, $host, $cleanUrl);
$this->cache->set($cleanUrl, $processedContent);
return $processedContent;
}
} catch (Exception $e) {
error_log("SELENIUM_ERROR: " . $e->getMessage());
}
Logger::getInstance()->log($cleanUrl, 'GENERAL_FETCH_ERROR');