lolcat
/
4get
6
13
Fork 4
Code Issues 13 Pull Requests Packages Projects Releases Wiki Activity

google hotfix

This commit is contained in:
lolcat 2024-03-14 22:33:01 -04:00
parent d03d251091
commit c4c008c192
1 changed files with 40 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

61
scraper/google.php
View File

@ -3,7 +3,6 @@
// todo: // todo:
// aliexpress tracking links // aliexpress tracking links
// enhanced msx notice // enhanced msx notice
// detect "sorry" page
class google{ class google{
@ -654,6 +653,7 @@ class google{
throw new Exception("Failed to get HTML"); throw new Exception("Failed to get HTML");
} }
//$html = file_get_contents("scraper/google.html");
} }
return $this->parsepage($html, "web", $search, $ip); return $this->parsepage($html, "web", $search, $ip);
@ -2322,11 +2322,11 @@ class google{
->getElementsByClassName( ->getElementsByClassName(
$this->findstyles( $this->findstyles(
[ [
"font-weight" => "bold",
"font-size" => "16px",
"color" => "#000", "color" => "#000",
"font-size" => "16px",
"font-weight" => "bold",
"margin" => "0", "margin" => "0",
"padding" => "12px 16px 0 16px" "padding" => "12px 16px 0px 16px"
], ],
self::is_class self::is_class
), ),
@ -2529,6 +2529,19 @@ class google{
$description = []; $description = [];
$pcitems =
$this->fuckhtml
->getElementsByClassName(
"pcitem",
"div"
);
if(count($pcitems) !== 0){
// ignore elements with carousels in them
continue;
}
foreach($as as $a){ foreach($as as $a){
// //
@ -3075,27 +3088,38 @@ class google{
private function findstyles($rules, $is){ private function findstyles($rules, $is){
ksort($rules); $c = count($rules);
foreach($this->computedstyle as $stylename => $styles){ foreach($this->computedstyle as $classname => $styles){
if($styles == $rules){ if($classname[0] != $is){
preg_match( // not a class, skip
'/\\' . $is . '([^ .]+)/', continue;
$stylename, }
$out
); $i = 0;
foreach($styles as $stylename => $stylevalue){
if(count($out) === 2){ if(
isset($rules[$stylename]) &&
$rules[$stylename] == $stylevalue
){
return $out[1]; $i++;
}else{
continue 2;
} }
}
if($c === $i){
return false; return ltrim($classname, $is);
} }
} }
// fail, did not find classname.
return false; return false;
} }
@ -3103,7 +3127,7 @@ class google{
// get style tags // get style tags
preg_match_all( preg_match_all(
'/([^{]+){([^}]+)}/', '/([^{]+){([^}]*)}/',
$style, $style,
$tags_regex $tags_regex
); );
@ -3139,11 +3163,6 @@ class google{
} }
} }
foreach($tags as &$value){
ksort($value);
}
return $tags; return $tags;
} }