18366940435

Committed 09 Oct 2025 05:55AM UTC coverage: 69.666% (+0.4%) from 69.284%

Build # 18366940435

Build Type

push

github

Committed by

web-flow

Commit Message

[TASK] Add RequestHelper to cleanup the BackendControllers (#1198)

Run Details

23 of 46 new or added lines in 4 files covered. (50.0%)

1920 of 2756 relevant lines covered (69.67%)

3.24 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

0.0

/Classes/Controller/Backend/BackendModuleStartCrawlingController.php

<?php

declare(strict_types=1);

namespace AOE\Crawler\Controller\Backend;

/*
 * (c) 2022-     Tomas Norre Mikkelsen <tomasnorre@gmail.com>
 *
 * This file is part of the TYPO3 Crawler Extension.
 *
 * It is free software; you can redistribute it and/or modify it under
 * the terms of the GNU General Public License, either version 2
 * of the License, or any later version.
 *
 * For the full copyright and license information, please read the
 * LICENSE.txt file that was distributed with this source code.
 *
 * The TYPO3 project - inspiring people to share!
 */

use AOE\Crawler\Controller\Backend\Helper\RequestHelper;
use AOE\Crawler\Controller\Backend\Helper\UrlBuilder;
use AOE\Crawler\Controller\CrawlerController;
use AOE\Crawler\Domain\Model\Reason;
use AOE\Crawler\Event\InvokeQueueChangeEvent;
use AOE\Crawler\Utility\MessageUtility;
use Psr\Http\Message\ResponseInterface;
use Psr\Http\Message\ServerRequestInterface;
use Psr\Http\Message\UriInterface;
use Symfony\Contracts\Service\Attribute\Required;
use TYPO3\CMS\Backend\Routing\Exception\RouteNotFoundException;
use TYPO3\CMS\Backend\Routing\UriBuilder;
use TYPO3\CMS\Backend\Template\ModuleTemplate;
use TYPO3\CMS\Core\EventDispatcher\EventDispatcher;
use TYPO3\CMS\Core\Utility\GeneralUtility;

/**
 * @internal since v12.0.0
 */
class BackendModuleStartCrawlingController extends AbstractBackendModuleController implements BackendModuleControllerInterface
{
    private const BACKEND_MODULE = 'web_site_crawler_start';
    private const LINE_FEED = 10;
    private const CARRIAGE_RETURN = 13;
    private int $reqMinute = 1000;
    private EventDispatcher $eventDispatcher;

    /**
     * @var array holds the selection of configuration from the configuration selector box
     */
    private $incomingConfigurationSelection = [];

    public function __construct(
        private readonly CrawlerController $crawlerController,
        private readonly UriBuilder $backendUriBuilder,
    ) {
    }

    #[Required]
    public function setEventDispatcher(): void
    {
        $this->eventDispatcher = GeneralUtility::makeInstance(EventDispatcher::class);
    }

    public function handleRequest(ServerRequestInterface $request): ResponseInterface
    {
        $this->makeCrawlerProcessableChecks($this->extensionSettings);
        $this->pageUid = (int) ($request->getQueryParams()['id'] ?? -1);
        $this->moduleTemplate = $this->setupView($request, $this->pageUid);
        $this->moduleTemplate->makeDocHeaderModuleMenu([
            'id' => $request->getQueryParams()['id'] ?? -1,
        ]);

        $this->assignValues($request);
        return $this->moduleTemplate->renderResponse('Backend/ShowCrawlerInformation');
    }

    private function assignValues(ServerRequestInterface $request): ModuleTemplate
    {
        $logUrl = $this->backendUriBuilder->buildUriFromRoute('web_site_crawler_log', [
            'id' => $this->pageUid,
        ]);

        $crawlingDepth = RequestHelper::getIntFromRequest($request, 'crawlingDepth');
        $submitCrawlUrls = RequestHelper::getBoolFromRequest($request, '_crawl');
        $downloadCrawlUrls = RequestHelper::getBoolFromRequest($request, '_download');
        $scheduledTime = $this->getScheduledTime(RequestHelper::getStringFromRequest($request, 'tstamp', 'now'));

        $this->incomingConfigurationSelection = RequestHelper::getArrayFromRequest($request, 'configurationSelection');

        //$this->crawlerController = $this->getCrawlerController();
        $this->crawlerController->setID = GeneralUtility::md5int(microtime());

        $queueRows = '';
        $noConfigurationSelected = $this->isNoConfigurationSelected();
        if ($noConfigurationSelected) {
            MessageUtility::addWarningMessage(
                $this->getLanguageService()->sL(
                    'LLL:EXT:crawler/Resources/Private/Language/locallang.xlf:labels.noConfigSelected'
                )
            );
        } else {
            if ($submitCrawlUrls) {
                $reason = new Reason();
                $reason->setReason(Reason::REASON_GUI_SUBMIT);
                $reason->setDetailText(
                    'The user ' . $GLOBALS['BE_USER']->user['username'] . ' added pages to the crawler queue manually'
                );
                $this->eventDispatcher->dispatch(new InvokeQueueChangeEvent($reason));
            }

            $queueRows = $this->crawlerController->getPageTreeAndUrls(
                $this->pageUid,
                $crawlingDepth,
                $scheduledTime,
                $this->reqMinute,
                $submitCrawlUrls,
                $downloadCrawlUrls,
                // Do not filter any processing instructions
                [],
                $this->incomingConfigurationSelection
            );
        }

        // Download Urls to crawl:
        $downloadUrls = $this->crawlerController->downloadUrls;
        if ($downloadCrawlUrls) {
            // Creating output header:
            header('Content-Type: application/octet-stream');
            header('Content-Disposition: attachment; filename=CrawlerUrls.txt');

            // Printing the content of the CSV lines:
            echo implode(chr(self::CARRIAGE_RETURN) . chr(self::LINE_FEED), $downloadUrls);
            exit;
        }

        return $this->moduleTemplate->assignMultiple([
            'currentPageId' => $this->pageUid,
            'noConfigurationSelected' => $noConfigurationSelected,
            'submitCrawlUrls' => $submitCrawlUrls,
            'amountOfUrls' => count($this->crawlerController->duplicateTrack ?? []),
            'selectors' => $this->generateConfigurationSelectors($this->pageUid, (string) $crawlingDepth, $request),
            'queueRows' => $queueRows,
            'displayActions' => 0,
            'actionUrl' => $this->getActionUrl(),
            'logUrl' => $logUrl,
        ]);
    }

    /**
     * Generates the configuration selectors for compiling URLs:
     */
    private function generateConfigurationSelectors(
        int $pageId,
        string $crawlingDepth,
        ServerRequestInterface $request
    ): array {
        $selectors = [];
        $selectors['depth'] = $this->selectorBox(
            [
                0 => $this->getLanguageService()->sL(
                    'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_0'
                ),
                1 => $this->getLanguageService()->sL(
                    'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_1'
                ),
                2 => $this->getLanguageService()->sL(
                    'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_2'
                ),
                3 => $this->getLanguageService()->sL(
                    'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_3'
                ),
                4 => $this->getLanguageService()->sL(
                    'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_4'
                ),
                99 => $this->getLanguageService()->sL(
                    'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_infi'
                ),
            ],
            'crawlingDepth',
            $crawlingDepth,
            false
        );

        // Configurations
        $availableConfigurations = $this->crawlerController->getConfigurationsForBranch(
            $pageId,
            (int) $crawlingDepth,
        );
        $selectors['configurations'] = $this->selectorBox(
            empty($availableConfigurations) ? [] : array_combine($availableConfigurations, $availableConfigurations),
            'configurationSelection',
            $this->incomingConfigurationSelection,
            true
        );

        // Scheduled time:
        $selectors['scheduled'] = $this->selectorBox(
            [
                'now' => $this->getLanguageService()->sL(
                    'LLL:EXT:crawler/Resources/Private/Language/locallang.xlf:labels.time.now'
                ),
                'midnight' => $this->getLanguageService()->sL(
                    'LLL:EXT:crawler/Resources/Private/Language/locallang.xlf:labels.time.midnight'
                ),
                '04:00' => $this->getLanguageService()->sL(
                    'LLL:EXT:crawler/Resources/Private/Language/locallang.xlf:labels.time.4am'
                ),
            ],
            'tstamp',
            RequestHelper::getStringFromRequest($request, 'tstamp', 'now'),
            false
        );

        return $selectors;
    }

    /**
     * Create selector box
     *
     * @param array $optArray Options key(value) => label pairs
     * @param string $name Selector box name
     * @param string|array|null $value Selector box value (array for multiple...)
     * @param boolean $multiple If set, will draw multiple box.
     *
     * @return string HTML select element
     */
    private function selectorBox($optArray, $name, string|array|null $value, bool $multiple): string
    {
        if (!is_string($value) && !is_array($value)) {
            $value = '';
        }

        $options = [];
        foreach ($optArray as $key => $val) {
            $selected = (
                !$multiple && is_string($value) && !strcmp($value, (string) $key)
            ) || (
                $multiple && in_array($key, (array) $value, true)
            );
            $options[] = '
                <option value="' . $key . '" ' . ($selected ? ' selected="selected"' : '') . '>' . htmlspecialchars(
                (string) $val,
                ENT_QUOTES | ENT_HTML5
            ) . '</option>';
        }

        return '<select class="form-select" name="' . htmlspecialchars(
            $name . ($multiple ? '[]' : ''),
            ENT_QUOTES | ENT_HTML5
        ) . '"' . ($multiple ? ' multiple' : '') . '>' . implode('', $options) . '</select>';
    }

    private function getScheduledTime(string $time): int
    {
        $scheduledTime = match ($time) {
            'midnight' => mktime(0, 0, 0),
            '04:00' => mktime(0, 0, 0) + 4 * 3600,
            default => time(),
        };

        if (!$scheduledTime) {
            return time();
        }

        return $scheduledTime;
    }

    /**
     * @throws RouteNotFoundException
     */
    private function getActionUrl(): UriInterface
    {
        return GeneralUtility::makeInstance(UrlBuilder::class)->getBackendModuleUrl([], self::BACKEND_MODULE);
    }

    private function isNoConfigurationSelected(): bool
    {
        return empty($this->incomingConfigurationSelection)
            || (count($this->incomingConfigurationSelection) === 1 && empty($this->incomingConfigurationSelection[0]));
    }
}

1	<?php
2
3	declare(strict_types=1);
4
5	namespace AOE\Crawler\Controller\Backend;
6
7	/*
8	* (c) 2022- Tomas Norre Mikkelsen <tomasnorre@gmail.com>
9	*
10	* This file is part of the TYPO3 Crawler Extension.
11	*
12	* It is free software; you can redistribute it and/or modify it under
13	* the terms of the GNU General Public License, either version 2
14	* of the License, or any later version.
15	*
16	* For the full copyright and license information, please read the
17	* LICENSE.txt file that was distributed with this source code.
18	*
19	* The TYPO3 project - inspiring people to share!
20	*/
21
22	use AOE\Crawler\Controller\Backend\Helper\RequestHelper;
23	use AOE\Crawler\Controller\Backend\Helper\UrlBuilder;
24	use AOE\Crawler\Controller\CrawlerController;
25	use AOE\Crawler\Domain\Model\Reason;
26	use AOE\Crawler\Event\InvokeQueueChangeEvent;
27	use AOE\Crawler\Utility\MessageUtility;
28	use Psr\Http\Message\ResponseInterface;
29	use Psr\Http\Message\ServerRequestInterface;
30	use Psr\Http\Message\UriInterface;
31	use Symfony\Contracts\Service\Attribute\Required;
32	use TYPO3\CMS\Backend\Routing\Exception\RouteNotFoundException;
33	use TYPO3\CMS\Backend\Routing\UriBuilder;
34	use TYPO3\CMS\Backend\Template\ModuleTemplate;
35	use TYPO3\CMS\Core\EventDispatcher\EventDispatcher;
36	use TYPO3\CMS\Core\Utility\GeneralUtility;
37
38	/**
39	* @internal since v12.0.0
40	*/
41	class BackendModuleStartCrawlingController extends AbstractBackendModuleController implements BackendModuleControllerInterface
42	{
43	private const BACKEND_MODULE = 'web_site_crawler_start';
44	private const LINE_FEED = 10;
45	private const CARRIAGE_RETURN = 13;
46	private int $reqMinute = 1000;
47	private EventDispatcher $eventDispatcher;
48
49	/**
50	* @var array holds the selection of configuration from the configuration selector box
51	*/
52	private $incomingConfigurationSelection = [];
53
54	public function __construct(
55	private readonly CrawlerController $crawlerController,
56	private readonly UriBuilder $backendUriBuilder,
57	) {
58	}	×
59
60	#[Required]
61	public function setEventDispatcher(): void
62	{
63	$this->eventDispatcher = GeneralUtility::makeInstance(EventDispatcher::class);	×
64	}
65
66	public function handleRequest(ServerRequestInterface $request): ResponseInterface
67	{
68	$this->makeCrawlerProcessableChecks($this->extensionSettings);	×
69	$this->pageUid = (int) ($request->getQueryParams()['id'] ?? -1);	×
70	$this->moduleTemplate = $this->setupView($request, $this->pageUid);	×
71	$this->moduleTemplate->makeDocHeaderModuleMenu([	×
72	'id' => $request->getQueryParams()['id'] ?? -1,	×
73	]);	×
74
75	$this->assignValues($request);	×
76	return $this->moduleTemplate->renderResponse('Backend/ShowCrawlerInformation');	×
77	}
78
79	private function assignValues(ServerRequestInterface $request): ModuleTemplate
80	{
81	$logUrl = $this->backendUriBuilder->buildUriFromRoute('web_site_crawler_log', [	×
82	'id' => $this->pageUid,	×
83	]);	×
84
NEW 85	$crawlingDepth = RequestHelper::getIntFromRequest($request, 'crawlingDepth');	×
NEW 86	$submitCrawlUrls = RequestHelper::getBoolFromRequest($request, '_crawl');	×
NEW 87	$downloadCrawlUrls = RequestHelper::getBoolFromRequest($request, '_download');	×
NEW 88	$scheduledTime = $this->getScheduledTime(RequestHelper::getStringFromRequest($request, 'tstamp', 'now'));	×
89
NEW 90	$this->incomingConfigurationSelection = RequestHelper::getArrayFromRequest($request, 'configurationSelection');	×
91
92	//$this->crawlerController = $this->getCrawlerController();
93	$this->crawlerController->setID = GeneralUtility::md5int(microtime());	×
94
95	$queueRows = '';	×
NEW 96	$noConfigurationSelected = $this->isNoConfigurationSelected();	×
97	if ($noConfigurationSelected) {	×
98	MessageUtility::addWarningMessage(	×
99	$this->getLanguageService()->sL(	×
100	'LLL:EXT:crawler/Resources/Private/Language/locallang.xlf:labels.noConfigSelected'	×
101	)	×
102	);	×
103	} else {
104	if ($submitCrawlUrls) {	×
105	$reason = new Reason();	×
106	$reason->setReason(Reason::REASON_GUI_SUBMIT);	×
107	$reason->setDetailText(	×
108	'The user ' . $GLOBALS['BE_USER']->user['username'] . ' added pages to the crawler queue manually'	×
109	);	×
110	$this->eventDispatcher->dispatch(new InvokeQueueChangeEvent($reason));	×
111	}
112
113	$queueRows = $this->crawlerController->getPageTreeAndUrls(	×
114	$this->pageUid,	×
115	$crawlingDepth,	×
116	$scheduledTime,	×
117	$this->reqMinute,	×
118	$submitCrawlUrls,	×
119	$downloadCrawlUrls,	×
120	// Do not filter any processing instructions
121	[],	×
122	$this->incomingConfigurationSelection	×
123	);	×
124	}
125
126	// Download Urls to crawl:
127	$downloadUrls = $this->crawlerController->downloadUrls;	×
128	if ($downloadCrawlUrls) {	×
129	// Creating output header:
130	header('Content-Type: application/octet-stream');	×
131	header('Content-Disposition: attachment; filename=CrawlerUrls.txt');	×
132
133	// Printing the content of the CSV lines:
134	echo implode(chr(self::CARRIAGE_RETURN) . chr(self::LINE_FEED), $downloadUrls);	×
135	exit;	×
136	}
137
138	return $this->moduleTemplate->assignMultiple([	×
139	'currentPageId' => $this->pageUid,	×
140	'noConfigurationSelected' => $noConfigurationSelected,	×
141	'submitCrawlUrls' => $submitCrawlUrls,	×
142	'amountOfUrls' => count($this->crawlerController->duplicateTrack ?? []),	×
NEW 143	'selectors' => $this->generateConfigurationSelectors($this->pageUid, (string) $crawlingDepth, $request),	×
144	'queueRows' => $queueRows,	×
145	'displayActions' => 0,	×
146	'actionUrl' => $this->getActionUrl(),	×
147	'logUrl' => $logUrl,	×
148	]);	×
149	}
150
151	/**
152	* Generates the configuration selectors for compiling URLs:
153	*/
154	private function generateConfigurationSelectors(
155	int $pageId,
156	string $crawlingDepth,
157	ServerRequestInterface $request
158	): array {
159	$selectors = [];	×
160	$selectors['depth'] = $this->selectorBox(	×
161	[	×
162	0 => $this->getLanguageService()->sL(	×
163	'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_0'	×
164	),	×
165	1 => $this->getLanguageService()->sL(	×
166	'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_1'	×
167	),	×
168	2 => $this->getLanguageService()->sL(	×
169	'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_2'	×
170	),	×
171	3 => $this->getLanguageService()->sL(	×
172	'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_3'	×
173	),	×
174	4 => $this->getLanguageService()->sL(	×
175	'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_4'	×
176	),	×
177	99 => $this->getLanguageService()->sL(	×
178	'LLL:EXT:core/Resources/Private/Language/locallang_core.xlf:labels.depth_infi'	×
179	),	×
180	],	×
181	'crawlingDepth',	×
182	$crawlingDepth,	×
183	false	×
184	);	×
185
186	// Configurations
187	$availableConfigurations = $this->crawlerController->getConfigurationsForBranch(	×
188	$pageId,	×
189	(int) $crawlingDepth,	×
190	);	×
191	$selectors['configurations'] = $this->selectorBox(	×
192	empty($availableConfigurations) ? [] : array_combine($availableConfigurations, $availableConfigurations),	×
193	'configurationSelection',	×
194	$this->incomingConfigurationSelection,	×
195	true	×
196	);	×
197
198	// Scheduled time:
199	$selectors['scheduled'] = $this->selectorBox(	×
200	[	×
201	'now' => $this->getLanguageService()->sL(	×
202	'LLL:EXT:crawler/Resources/Private/Language/locallang.xlf:labels.time.now'	×
203	),	×
204	'midnight' => $this->getLanguageService()->sL(	×
205	'LLL:EXT:crawler/Resources/Private/Language/locallang.xlf:labels.time.midnight'	×
206	),	×
207	'04:00' => $this->getLanguageService()->sL(	×
208	'LLL:EXT:crawler/Resources/Private/Language/locallang.xlf:labels.time.4am'	×
209	),	×
210	],	×
211	'tstamp',	×
NEW 212	RequestHelper::getStringFromRequest($request, 'tstamp', 'now'),	×
213	false	×
214	);	×
215
216	return $selectors;	×
217	}
218
219	/**
220	* Create selector box
221	*
222	* @param array $optArray Options key(value) => label pairs
223	* @param string $name Selector box name
224	* @param string\|array\|null $value Selector box value (array for multiple...)
225	* @param boolean $multiple If set, will draw multiple box.
226	*
227	* @return string HTML select element
228	*/
229	private function selectorBox($optArray, $name, string\|array\|null $value, bool $multiple): string
230	{
231	if (!is_string($value) && !is_array($value)) {	×
232	$value = '';	×
233	}
234
235	$options = [];	×
236	foreach ($optArray as $key => $val) {	×
237	$selected = (	×
238	!$multiple && is_string($value) && !strcmp($value, (string) $key)	×
239	) \|\| (	×
240	$multiple && in_array($key, (array) $value, true)	×
241	);	×
242	$options[] = '	×
243	<option value="' . $key . '" ' . ($selected ? ' selected="selected"' : '') . '>' . htmlspecialchars(	×
244	(string) $val,	×
245	ENT_QUOTES \| ENT_HTML5	×
246	) . '</option>';	×
247	}
248
249	return '<select class="form-select" name="' . htmlspecialchars(	×
250	$name . ($multiple ? '[]' : ''),	×
251	ENT_QUOTES \| ENT_HTML5	×
252	) . '"' . ($multiple ? ' multiple' : '') . '>' . implode('', $options) . '</select>';	×
253	}
254
255	private function getScheduledTime(string $time): int
256	{
257	$scheduledTime = match ($time) {	×
258	'midnight' => mktime(0, 0, 0),	×
259	'04:00' => mktime(0, 0, 0) + 4 * 3600,	×
260	default => time(),	×
261	};	×
262
263	if (!$scheduledTime) {	×
264	return time();	×
265	}
266
267	return $scheduledTime;	×
268	}
269
270	/**
271	* @throws RouteNotFoundException
272	*/
273	private function getActionUrl(): UriInterface
274	{
275	return GeneralUtility::makeInstance(UrlBuilder::class)->getBackendModuleUrl([], self::BACKEND_MODULE);	×
276	}
277
278	private function isNoConfigurationSelected(): bool
279	{
NEW 280	return empty($this->incomingConfigurationSelection)	×
NEW 281	\|\| (count($this->incomingConfigurationSelection) === 1 && empty($this->incomingConfigurationSelection[0]));	×
282	}
283	}

tomasnorre / crawler / 18366940435

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous