Glassdoor 职位数据抓取(by URL)

<?php

// API URL
const API_URL = "https://openapi.coreclaw.com/api/v1/scraper/run";

// Your API KEY
const API_KEY = "<YOUR_API_KEY>";

// curl timeout (seconds)
const TIMEOUT = 30;

/**
 * Run scraper
 *
 * @param array $params Request parameters
 * @param string $apiKey API Key
 * @return array Return result ["success" => bool, "run_slug" => string|null, "error" => string|null]
 */
function runScraper(array $params, string $apiKey): array
{
    // Initialize cURL
    $ch = curl_init();

    // Set cURL options
    curl_setopt_array($ch, [
        CURLOPT_URL => API_URL,
        CURLOPT_RETURNTRANSFER => true,
        CURLOPT_ENCODING => "",
        CURLOPT_MAXREDIRS => 10,
        CURLOPT_TIMEOUT => TIMEOUT,
        CURLOPT_FOLLOWLOCATION => true,
        CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
        CURLOPT_CUSTOMREQUEST => "POST",
        CURLOPT_POSTFIELDS => json_encode($params),
        CURLOPT_HTTPHEADER => [
            "api-key: " . $apiKey,
            "Content-Type: application/json"
        ],
    ]);

    // Execute request
    $response = curl_exec($ch);
    $httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
    $error = curl_error($ch);

    // Close cURL
    curl_close($ch);

    // Check cURL error
    if ($error) {
        return [
            "success" => false,
            "run_slug" => null,
            "error" => "cURL error: " . $error
        ];
    }

    // Check HTTP status code
    if ($httpCode !== 200) {
        return [
            "success" => false,
            "run_slug" => null,
            "error" => "HTTP error: " . $httpCode . " - " . $response
        ];
    }

    // Parse response
    $result = json_decode($response, true);
    if (json_last_error() !== JSON_ERROR_NONE) {
        return [
            "success" => false,
            "run_slug" => null,
            "error" => "JSON decode error: " . json_last_error_msg()
        ];
    }

    // Check business error code
    if (isset($result["code"]) && $result["code"] !== 0) {
        return [
            "success" => false,
            "run_slug" => null,
            "error" => "Business error: " . (isset($result["message"]) ? $result["message"] : "Unknown error") . " (code: " . $result["code"] . ")"
        ];
    }

    // Return success result
    return [
        "success" => true,
        "run_slug" => isset($result["data"]["run_slug"]) ? $result["data"]["run_slug"] : null,
        "error" => null
    ];
}

/**
 * Main function
 */
function main()
{
    // Build request parameters
    $requestParams = [
        "scraper_slug" => "01KG2DV66JTCN65ZBTRX3M456E",
        "version" => "v1.0.8",
        "input" => [
            "parameters" => [
                "system" => [
                    "proxy_region" => "",
                    "cpus" => 0.125,
                    "memory" => 512,
                    "execute_limit_time_seconds" => 1800,
                    "max_total_charge" => 0,
                    "max_total_traffic" => 0
                ],
                "custom" => {
          'url': [
                    {
                              'url': 'https://www.glassdoor.com/job-listing/staff-product-analyst-intuit-JV_IC4508422_KO0,21_KE22,28.htm?jl=1008980804695&cs=1_d6b1b697&s=58&t=SR&pos=104&src=GD_JOB_AD&cpc=44CD5376B8534B8F&guid=0000018c0ae77315b0674c403cf266a5&jobListingId=1008980804695&ao=1110586&vt=w&jrtk=5-yul1-0-1hg5eesq3ir3v800-4dda97f9dab7b990---6NYlbfkN0BKYl3zWsktiTMfhnn8eMRsNrNhPXFqHgxANdB9sZgO2gg8MIfiMoKrpC4RrjRFuq5cN1FayxkUzXNpGKe4khcoosOdCsWKdjNUQlszKRvhnVCv-3GxQD0UURkmC2SRM5G4PDS-csRoploh14dWMXKmKSHFODefNWscELEdL9st3xF8QQKpVtPfaY0ycbd-ETjhmsqIBjkCxzKMwE-vSoMbWao2wahtIcpefclgD97rRwH69BkyKCUkw6_dAAZrs2ADvu3bHDeyOTLSIlwBZqRxVe05yri9f7rkilGsITPCgvv3Cfg9tReQUvDznUmY5TAQO1ykaDDCaaDZNjow2bO8henmAFyasHjvhLiubxbzO_JweUKeGcMqwwy80Wgmp5Xcz5rtTc-1vuQXmy3hK-lo-60GgtgLcZbRdKat1Z8tQww1fcL-tfB4nfNQiZmhXUNKNp-K-ChnjJ8FFwNL48I5GC9uunA74XRNacCIRa-mQAw991VyPZoA4xpzc-zfhk12BJvfbAKDgmBIOLa13A9Yw7poZBIKu1yvAsV1a2ejboFEtXzdxtOfmooRjrOTHx67cpRbppk-KmwyuMvjHWJrz_3XSepcuUm7zp9tGBliAyeRX01IPuVMoGsN7CBoTn3zS9kEwLVU6ejV0l1hF5437iDfn61jqnywAzCrmOCaKouMCFJBj2Yage5R4dIdFcm53lB9xSqWy9zGdJ31N5Xj&cb=1700989989992&ctt=1700990085479'
                    }
          ]
}
            ]
        ],
        "callback_url" => "https://your-domain.com/callback"
    ];

    // Send request
    echo "Sending request to API...
";
    $result = runScraper($requestParams, API_KEY);

    // Handle result
    if ($result["success"]) {
        echo "Worker run successful!
";
        echo "Run record ID: " . $result["run_slug"] . "
";
        echo "You can use this ID to query run status and results
";
    } else {
        echo "Request failed!
";
        echo "Error message: " . $result["error"] . "
";
    }
}

// Execute main function
main();