优化爬虫，增加缓存，线程池，保存浏览器用户信息

2026-06-20 15:49:40 +08:00
parent e95f4a227c
commit 09bfa9ae01
26 changed files with 1584 additions and 454 deletions
@@ -0,0 +1,84 @@
+<?php
+
+declare(strict_types=1);
+
+namespace app\common\library\scrm;
+
+use think\Config;
+
+/**
+ * 云控蜘蛛 antiBot 配置构建器
+ *
+ * 约定：sessionKey = "{ticketType}:{host}"；profile=real 当 ticket_type 在 site 配置列表中
+ */
+class AntiBotConfigBuilder
+{
+    /**
+     * 为指定工单类型与落地页 URL 构建 Node antiBot 配置；未启用类型返回 null
+     *
+     * @return array<string, mixed>|null
+     */
+    public static function build(string $ticketType, string $pageUrl): ?array
+    {
+        $enabledTypes = self::getEnabledTypes();
+        if (!in_array($ticketType, $enabledTypes, true)) {
+            return null;
+        }
+
+        $host = (string) parse_url($pageUrl, PHP_URL_HOST);
+        if ($host === '') {
+            return null;
+        }
+
+        $sessionKey = $ticketType . ':' . $host;
+        $ttlMinutes = self::getSessionTtlMinutes();
+
+        return [
+            'enabled'            => true,
+            'profile'            => 'real',
+            'turnstile'          => true,
+            'solverFallback'     => true,
+            'sessionKey'         => $sessionKey,
+            'challengeTimeoutMs' => 60000,
+            'sessionTtlMinutes'  => $ttlMinutes,
+        ];
+    }
+
+    /**
+     * 解析 sessionKey（供 cron 分组调度使用）
+     */
+    public static function resolveSessionKey(string $ticketType, string $pageUrl): string
+    {
+        $host = (string) parse_url($pageUrl, PHP_URL_HOST);
+        return $ticketType . ':' . ($host !== '' ? $host : 'unknown');
+    }
+
+    /**
+     * 是否对该工单类型启用 antiBot
+     */
+    public static function isEnabledForType(string $ticketType): bool
+    {
+        return in_array($ticketType, self::getEnabledTypes(), true);
+    }
+
+    /**
+     * @return list<string>
+     */
+    public static function getEnabledTypes(): array
+    {
+        $raw = (string) Config::get('site.split_scrm_antibot_types');
+        if ($raw === '') {
+            return ['whatshub', 'chatknow'];
+        }
+        $parts = array_map('trim', explode(',', $raw));
+        return array_values(array_filter($parts, static function (string $v): bool {
+            return $v !== '';
+        }));
+    }
+
+    public static function getSessionTtlMinutes(): int
+    {
+        $minutes = (int) Config::get('site.split_scrm_session_ttl_minutes');
+        return $minutes > 0 ? $minutes : 120;
+    }
+}