完整版V1 加入爬虫功能

2026-06-09 03:36:30 +08:00
parent 34d76cce74
commit a68b83fcbd
69 changed files with 5058 additions and 56 deletions
@@ -0,0 +1,165 @@
+<?php
+
+declare(strict_types=1);
+
+namespace app\common\library\scrm;
+
+use Exception;
+
+/**
+ * 云控蜘蛛抽象基类（Node Headless 拦截 + 翻页 + 清洗）
+ */
+abstract class AbstractScrmSpider implements ScrmSpiderInterface
+{
+    public const MODE_FETCH = 'fetch';
+
+    public const MODE_UI = 'ui_click';
+
+    protected string $nodeHost;
+
+    public function __construct(string $nodeHost = 'http://127.0.0.1:3001')
+    {
+        $this->nodeHost = rtrim($nodeHost, '/');
+    }
+
+    /** @return array<string, mixed> */
+    abstract protected function getSpiderConfig(): array;
+
+    /**
+     * @param array<string, mixed>|null $countData
+     */
+    abstract protected function extractListTotalPages($listFirstPageData, $countData = null);
+
+    /**
+     * @return array<string, mixed>
+     */
+    abstract protected function buildListPageParams(int $page): array;
+
+    /** @return array<string, mixed> */
+    abstract protected function getUiPaginationConfig(): array;
+
+    /**
+     * @param mixed $detailData
+     * @param array<int, mixed> $allListPagesData
+     */
+    abstract protected function parseToUnifiedData($detailData, array $allListPagesData): UnifiedScrmData;
+
+    public function run(): UnifiedScrmData
+    {
+        $config = $this->getSpiderConfig();
+
+        $listApi = (string) ($config['listApi'] ?? '');
+        $detailApi = $config['detailApi'] ?? null;
+        $countApi = $config['countApi'] ?? null;
+
+        $apiUrlsToIntercept = [$listApi];
+        if ($detailApi) {
+            $apiUrlsToIntercept[] = $detailApi;
+        }
+        if ($countApi) {
+            $apiUrlsToIntercept[] = $countApi;
+        }
+
+        $initResult = $this->requestNode('/api/auth-and-intercept', [
+            'pageUrl'     => $config['pageUrl'],
+            'apiUrls'     => $apiUrlsToIntercept,
+            'authActions' => $config['authActions'] ?? [],
+        ]);
+
+        if (empty($initResult['success'])) {
+            throw new Exception('初始化失败: ' . ($initResult['error'] ?? '未知'));
+        }
+
+        $interceptedApis = $initResult['interceptedApis'];
+        $cookies = $initResult['cookies'];
+
+        if (!isset($interceptedApis[$listApi])) {
+            throw new Exception("致命错误：未能拦截到必须的列表接口 [{$listApi}]");
+        }
+
+        $detailData = $detailApi && isset($interceptedApis[$detailApi])
+            ? $interceptedApis[$detailApi]['data'] : null;
+        $countData = $countApi && isset($interceptedApis[$countApi])
+            ? $interceptedApis[$countApi]['data'] : null;
+
+        $listApiNode = $interceptedApis[$listApi];
+        $allListPagesData = [$listApiNode['data']];
+
+        $totalPages = $this->extractListTotalPages($listApiNode['data'], $countData);
+        $mode = $config['paginationMode'] ?? self::MODE_FETCH;
+
+        if ($totalPages > 1 || $totalPages === null) {
+            if ($mode === self::MODE_FETCH && $totalPages !== null) {
+                $paramList = [];
+                for ($page = 2; $page <= $totalPages; $page++) {
+                    $paramList[] = $this->buildListPageParams($page);
+                }
+
+                $fetchResult = $this->requestNode('/api/batch-fetch', [
+                    'tasks' => [[
+                        'apiPath'   => $listApi,
+                        'fullUrl'   => $listApiNode['url'],
+                        'headers'   => $listApiNode['headers'] ?? '',
+                        'paramList' => $paramList,
+                        'method'    => $config['listMethod'] ?? 'GET',
+                    ]],
+                    'cookies' => $cookies,
+                ], 120);
+
+                if (!empty($fetchResult['success'])) {
+                    foreach ($fetchResult['results'][$listApi] as $pResult) {
+                        if (!empty($pResult['success'])) {
+                            $allListPagesData[] = $pResult['data'];
+                        }
+                    }
+                }
+            } elseif ($mode === self::MODE_UI) {
+                $uiConfig = $this->getUiPaginationConfig();
+                $firstPageData = $listApiNode['data'];
+                $clicksToPerform = ($totalPages === null) ? 9999 : ($totalPages - 1);
+
+                $uiResult = $this->requestNode('/api/ui-pagination', [
+                    'apiUrl'          => $listApi,
+                    'pageUrl'         => $config['pageUrl'],
+                    'nextBtnSelector' => $uiConfig['nextBtnSelector'] ?? '',
+                    'waitMs'          => $uiConfig['waitMs'] ?? 2000,
+                    'clicksToPerform' => $clicksToPerform,
+                    'cookies'         => $cookies,
+                    'firstPageData'   => $firstPageData,
+                    'authActions'     => $config['authActions'] ?? [],
+                ], 1200);
+
+                if (!empty($uiResult['success']) && !empty($uiResult['data'])) {
+                    foreach ($uiResult['data'] as $pageData) {
+                        $allListPagesData[] = $pageData;
+                    }
+                }
+            }
+        }
+
+        return $this->parseToUnifiedData($detailData, $allListPagesData);
+    }
+
+    /**
+     * @param array<string, mixed> $payload
+     * @return array<string, mixed>
+     */
+    protected function requestNode(string $endpoint, array $payload, int $timeout = 60): array
+    {
+        $ch = curl_init($this->nodeHost . $endpoint);
+        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+        curl_setopt($ch, CURLOPT_POST, true);
+        curl_setopt($ch, CURLOPT_POSTFIELDS, json_encode($payload, JSON_UNESCAPED_UNICODE));
+        curl_setopt($ch, CURLOPT_HTTPHEADER, ['Content-Type: application/json']);
+        curl_setopt($ch, CURLOPT_TIMEOUT, $timeout);
+        $response = curl_exec($ch);
+        if (curl_errno($ch)) {
+            $err = curl_error($ch);
+            curl_close($ch);
+            throw new Exception($err);
+        }
+        curl_close($ch);
+        $decoded = json_decode((string) $response, true);
+        return is_array($decoded) ? $decoded : [];
+    }
+}