YGGverse/YGGo
1
0
Fork 0
mirror of https://github.com/YGGverse/YGGo.git synced 2025-01-24 21:44:59 +00:00
Code Issues Projects Releases Wiki Activity

fix meta variables overwrite

Browse Source
This commit is contained in:
ghost 2023-06-14 02:53:14 +03:00
parent 0949d7f871
commit d2469e9adc
1 changed files with 36 additions and 41 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

77
crontab/crawler.php
View File

@ -483,16 +483,14 @@ foreach ($db->getHostPageCrawlQueue(CRAWL_PAGE_LIMIT, time() - CRAWL_PAGE_SECOND
continue;
}
// Define variables
$title = null;
$description = null;
$keywords = null;
$robots = null;
$yggoManifest = null;
// Is HTML document
if (false !== stripos(Filter::mime($contentType), 'text/html')) {
// Define variables
$metaDescription = null;
$metaKeywords = null;
$metaYggoManifest = null;
// Parse content
$dom = new DomDocument();
@ -511,26 +509,26 @@ foreach ($db->getHostPageCrawlQueue(CRAWL_PAGE_LIMIT, time() - CRAWL_PAGE_SECOND
} else {
$title = $title->item(0)->nodeValue;
$metaTitle = $title->item(0)->nodeValue;
}
// Get optional page meta data
foreach (@$dom->getElementsByTagName('meta') as $meta) {
if (@$meta->getAttribute('name') == 'description') {
$description = @$meta->getAttribute('content');
$metaDescription = @$meta->getAttribute('content');
}
if (@$meta->getAttribute('name') == 'keywords') {
$keywords = @$meta->getAttribute('content');
$metaKeywords = @$meta->getAttribute('content');
}
if (@$meta->getAttribute('name') == 'robots') {
$robots = @$meta->getAttribute('content');
$metaRobots = @$meta->getAttribute('content');
// Ban page with meta robots:noindex value
if (false !== stripos($robots, 'noindex')) {
if (false !== stripos($metaRobots, 'noindex')) {
$hostPagesBanned += $db->updateHostPageTimeBanned($queueHostPage->hostPageId, time());
@ -546,7 +544,30 @@ foreach ($db->getHostPageCrawlQueue(CRAWL_PAGE_LIMIT, time() - CRAWL_PAGE_SECOND
// Grab meta yggo:manifest link when available
if (@$meta->getAttribute('name') == 'yggo:manifest') {
$yggoManifest = Filter::url(@$meta->getAttribute('content'));
$metaYggoManifest = Filter::url(@$meta->getAttribute('content'));
}
}
// Add queued page description if not exists
$db->addHostPageDescription($queueHostPage->hostPageId,
$metaTitle,
$metaDescription ? Filter::pageDescription($metaDescription) : null,
$metaKeywords ? Filter::pageKeywords($metaKeywords) : null,
$content ? ($queueHostPage->crawlMetaOnly ? null : base64_encode($content)) : null,
time());
// Update manifest registry
if (CRAWL_MANIFEST && !empty($metaYggoManifest) && filter_var($metaYggoManifest, FILTER_VALIDATE_URL) && preg_match(CRAWL_URL_REGEXP, $metaYggoManifest)) {
$metaYggoManifestCRC32 = crc32($metaYggoManifest);
if (!$db->getManifest($metaYggoManifestCRC32)) {
$db->addManifest($metaYggoManifestCRC32,
$metaYggoManifest,
(string) CRAWL_MANIFEST_DEFAULT_STATUS,
time());
$manifestsAdded++;
}
}
@ -569,7 +590,7 @@ foreach ($db->getHostPageCrawlQueue(CRAWL_PAGE_LIMIT, time() - CRAWL_PAGE_SECOND
}
if (!$title = @$img->getAttribute('title')) {
$title = null;
$title = null;
}
// Skip encoded content
@ -692,7 +713,7 @@ foreach ($db->getHostPageCrawlQueue(CRAWL_PAGE_LIMIT, time() - CRAWL_PAGE_SECOND
// Get title attribute if available
if (!$title = @$a->getAttribute('title')) {
$title = null;
$title = null;
}
// Skip anchor links
@ -851,32 +872,6 @@ foreach ($db->getHostPageCrawlQueue(CRAWL_PAGE_LIMIT, time() - CRAWL_PAGE_SECOND
}
}
// Add queued page description if not exists
if ($title || $description || $keywords) {
$db->addHostPageDescription($queueHostPage->hostPageId,
$title ? Filter::pageTitle($title) : null,
$description ? Filter::pageDescription($description) : null,
$keywords ? Filter::pageKeywords($keywords) : null,
$content ? ($queueHostPage->crawlMetaOnly ? null : base64_encode($content)) : null,
time());
}
// Update manifest registry
if (CRAWL_MANIFEST && !empty($yggoManifest) && filter_var($yggoManifest, FILTER_VALIDATE_URL) && preg_match(CRAWL_URL_REGEXP, $yggoManifest)) {
$yggoManifestCRC32 = crc32($yggoManifest);
if (!$db->getManifest($yggoManifestCRC32)) {
$db->addManifest($yggoManifestCRC32,
$yggoManifest,
(string) CRAWL_MANIFEST_DEFAULT_STATUS,
time());
$manifestsAdded++;
}
}
// Begin snaps
$snapLocal = false;
$snapMega = false;