Using observers

For more structured crawl handling, you can create observer classes instead of using closures. An observer must extend Spatie\Crawler\CrawlObservers\CrawlObserver:

namespace App;

use GuzzleHttp\Exception\RequestException;
use Spatie\Crawler\CrawlObservers\CrawlObserver;
use Spatie\Crawler\CrawlProgress;
use Spatie\Crawler\CrawlResponse;
use Spatie\Crawler\Enums\FinishReason;
use Spatie\Crawler\Enums\ResourceType;
use Spatie\Crawler\TransferStatistics;

class MyCrawlObserver extends CrawlObserver
{
    public function willCrawl(string $url, ?string $linkText, ?ResourceType $resourceType = null): void
    {
        // called before a URL is crawled
    }

    public function crawled(
        string $url,
        CrawlResponse $response,
        CrawlProgress $progress,
    ): void {
        // called when a URL has been successfully crawled
    }

    public function crawlFailed(
        string $url,
        RequestException $requestException,
        CrawlProgress $progress,
        ?string $foundOnUrl = null,
        ?string $linkText = null,
        ?ResourceType $resourceType = null,
        ?TransferStatistics $transferStats = null,
    ): void {
        // called when a URL could not be crawled
    }

    public function finishedCrawling(FinishReason $reason, CrawlProgress $progress): void
    {
        // called when the entire crawl is complete
    }
}

Pass the observer to the crawler:

Crawler::create('https://example.com')
    ->addObserver(new MyCrawlObserver())
    ->start();

The crawled() method receives a CrawlProgress object with live crawl statistics and a CrawlResponse that provides access to foundOnUrl(), linkText(), resourceType(), and other response data. See tracking progress for more on CrawlProgress.

The crawlFailed() method also includes $foundOnUrl, $linkText, and $resourceType parameters since there is no CrawlResponse available for failed requests. The $transferStats parameter provides transfer timing data (connection time, total transfer time, etc.) for the failed request. This is useful for detecting timeouts: if $transferStats->transferTimeInMs() exceeds your threshold, the request likely timed out rather than returning an error.

##Using multiple observers

You can add multiple observers. They will all be notified of every crawl event. The addObserver() method accepts multiple observers at once:

Crawler::create('https://example.com')
    ->addObserver(new LoggingObserver(), new MetricsObserver())
    ->start();

You can also chain separate addObserver() calls:

Crawler::create('https://example.com')
    ->addObserver(new LoggingObserver())
    ->addObserver(new MetricsObserver())
    ->start();

You can also combine observers with closure callbacks. Both will be called:

Crawler::create('https://example.com')
    ->addObserver(new MyObserver())
    ->onCrawled(function (string $url, CrawlResponse $response) {
        // this will also be called alongside the observer
    })
    ->start();