Начало обработки по ссылкам

This commit is contained in:
2024-09-04 16:12:03 +03:00
parent 48b4d6cccf
commit 04374fef40
10 changed files with 633 additions and 103 deletions

View File

@ -2,6 +2,7 @@
namespace ContingentParser\Http;
use ContingentParser\Logger\HttpLogger;
use ContingentParser\Printer;
use CurlHandle;
/**
* Summary of CurlHelper
@ -38,7 +39,7 @@ final class CurlHelper
curl_setopt($this->curl, CURLOPT_CONNECTTIMEOUT, 90);
}
/**
* Прекратить сессии
* Прекратить сессию
*/
public function __destruct()
{
@ -55,7 +56,6 @@ final class CurlHelper
if ($this->checkLocation($this->url, $html)) {
$html = $this->getContent();
}
$this->reportError();
return $html;
}
/**
@ -77,14 +77,14 @@ final class CurlHelper
* Сообщить об ошибке
* @return void
*/
private function reportError() : void
public function reportError() : void
{
$httpLogger = new HttpLogger('log/http-curl.log');
$httpCode = curl_getinfo($this->curl, CURLINFO_HTTP_CODE);
if ($httpCode != 200 && $httpCode != 0) {
print("\033[91mHTTP-code: $httpCode\033[0m\n");
Printer::println("HTTP-code: $httpCode", 'red');
$message = implode(' ', $this->site) . ' HTTP-code(' . $httpCode.')';
$httpLogger->log($message, $httpCode);
} else if ($httpCode == 0) {
@ -93,7 +93,7 @@ final class CurlHelper
$message .= " cURL error ({$errno}): ".curl_strerror($errno);
$httpLogger->log($message);
} else {
print("\033[94mHTTP-code: $httpCode\033[0m\n");
Printer::println("HTTP-code: $httpCode", 'blue');
}
}
}

View File

@ -1,12 +1,8 @@
<?php
namespace ContingentParser\Http;
use ContingentParser\Printer;
use GuzzleHttp\Client;
use GuzzleHttp\Exception\ClientException;
use GuzzleHttp\Exception\ConnectException;
use GuzzleHttp\Exception\RequestException;
use GuzzleHttp\Exception\ServerException;
use GuzzleHttp\Psr7\Exception\MalformedUriException;
use GuzzleHttp\TransferStats;
final class HttpClientFacade
@ -17,10 +13,8 @@ final class HttpClientFacade
public function __construct() {}
/**
* Обработка численности обучающихся
* @param string $url
* URL сайта
* @param array $site
* Идентификатор организации, и базовый URL
* @param string $url URL сайта
* @param array $site Идентификатор организации, и базовый URL
* @return string
*/
public function processEducationContingentSites(
@ -35,24 +29,19 @@ final class HttpClientFacade
$redirectUrl = $stats->getEffectiveUri();
}
]);
print("Redirect $url -> $redirectUrl" . PHP_EOL);
Printer::println("Redirect $url -> $redirectUrl");
$url .= substr($url, -1) == '/' ? '':'/';
$url .= "sveden/education/";
print("Parsing for $url" . PHP_EOL);
$url .= "sveden/education/study";
Printer::println("Parsing for $url");
$response = $client->get($url);
$httpCode = $response->getStatusCode();
print("\033[94mHTTP-code: $httpCode\033[0m\n");
Printer::println("HTTP-code: $httpCode", 'blue');
$html = $response->getBody()->getContents();
} catch (ClientException
| RequestException
| ConnectException
| ServerException
| MalformedUriException $e
} catch (\Exception $e
) {
print("\033[91mHTTP-code: ". $e->getCode(). "\033[0m\n");
Printer::println("HTTP-code: ".$e->getCode(), 'red');
$html = $this->handleException($url, $site);
} finally {
return $html;
@ -60,23 +49,24 @@ final class HttpClientFacade
}
/**
* Обработка исключения
* Повторная попытка спомощью CurlHelper
* @param string $url
* URL сайта
* Повторная попытка с помощью CurlHelper
* @param string $url URL сайта
* @param array $site
* @return string
*/
private function handleException(string $url, array $site) : string
private function handleException(string $url, array $site): string
{
$curlHelper = new CurlHelper($url, $site);
return $curlHelper->getContent();
$html = $curlHelper->getContent();
$curlHelper->reportError();
return $html;
}
/**
* Создать клиента с базовым URL
* @param string $url
* @return \GuzzleHttp\Client
*/
private function createClient(string $url) : Client
private function createClient(string $url): Client
{
$this->config = $this->config() + ["base_uri" => $url];
return new Client($this->config);