google scraper fix

This commit is contained in:
2026-08-01 20:04:04 -04:00
parent 8e7fe83774
commit db90a39c4f
5 changed files with 2115 additions and 60 deletions

View File

@@ -4,6 +4,9 @@ class backend{
public function __construct($scraper){
$this->scraper = $scraper;
$this->db = null;
$this->db_path_prefix = "data/sessions/";
$this->db_path = "{$this->db_path_prefix}{$scraper}.sqlite";
}
/*
@@ -208,4 +211,127 @@ class backend{
$payload[1] // proxy
];
}
//
// Sessions
//
public function session_init_db($force_init = false){
if(
$force_init ||
$this->db === null
){
$this->db = new SQLite3($this->db_path);
// create table if it doesnt exist
$this->db->exec(
"CREATE TABLE IF NOT EXISTS sessions ( " .
"id INTEGER PRIMARY KEY AUTOINCREMENT, " .
"created_at INTEGER NOT NULL, " .
"proxy TEXT NOT NULL, " .
"headers_json TEXT NOT NULL" .
")"
);
return true;
}
return false;
}
public function session_count(){
$this->session_init_db();
$stmt = $this->db->prepare(
"SELECT COUNT(*) as count FROM sessions"
);
$result = $stmt->execute();
$row = $result->fetchArray(SQLITE3_ASSOC);
return (int)($row["count"] ?? 0);
}
public function session_write($proxy, $headers){
$this->session_init_db();
$stmt = $this->db->prepare(
"INSERT INTO sessions ( " .
"created_at, " .
"proxy, " .
"headers_json " .
") VALUES ( " .
":created_at, " .
":proxy, " .
":headers_json" .
")"
);
$stmt->bindValue(':created_at', time(), SQLITE3_INTEGER);
$stmt->bindValue(':proxy', $proxy, SQLITE3_TEXT);
$stmt->bindValue(':headers_json', json_encode($headers), SQLITE3_TEXT);
$stmt->execute();
return $this->db->lastInsertRowID();
}
public function session_get(){
$count = $this->session_count();
if($count === 0){
return false;
}
$offset = apcu_inc("d." . $this->scraper) % $count;
$stmt = $this->db->prepare(
"SELECT * " .
"FROM sessions " .
"ORDER BY id ASC " .
"LIMIT 1 OFFSET :offset"
);
$stmt->bindValue(':offset', $offset, SQLITE3_INTEGER);
$result = $stmt->execute();
$row = $result->fetchArray(SQLITE3_ASSOC);
return [
"id" => $row["id"],
"created_at" => $row["created_at"],
"proxy" => $row["proxy"],
"headers" => json_decode($row["headers_json"], true)
];
}
public function session_destroy_list(){
if(
!preg_match(
'/^[A-Za-z0-9-]+$/',
$this->scraper
)
){
// path traversal
return false;
}
if(
file_exists("{$this->db_path_prefix}{$this->scraper}.sqlite") &&
is_file("{$this->db_path_prefix}{$this->scraper}.sqlite")
){
unlink("{$this->db_path_prefix}{$this->scraper}.sqlite");
return true;
}
return false;
}
}