3721152729

Build Type

github

Committed by GitHub

Commit Message

Merge 64c242322 into 4735d49b5

Pull Request Pull Request #988: [WIP][FEATURE] Setup new Backend Module

Run Details

434 of 434 new or added lines in 12 files covered. (100.0%)

1595 of 2533 relevant lines covered (62.97%)

3.21 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\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 Symfony\Contracts\Service\Attribute\Required;
use TYPO3\CMS\Backend\Routing\Exception\RouteNotFoundException;
use TYPO3\CMS\Backend\Template\ModuleTemplate;
use TYPO3\CMS\Core\EventDispatcher\EventDispatcher;
use TYPO3\CMS\Core\Http\Uri;
use TYPO3\CMS\Core\Utility\GeneralUtility;

/**
 * @internal since v12.0.0
 */
final class BackendModuleStartCrawlingController extends AbstractBackendModuleController implements BackendModuleControllerInterface
{
    private const BACKEND_MODULE = 'web_site_crawler_start';
    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 CrawlerController $crawlerController
    ) {
    }

    #[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();

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

    private function assignValues(): ModuleTemplate
    {
        $backendUriBuilder = GeneralUtility::makeInstance(\TYPO3\CMS\Backend\Routing\UriBuilder::class);
        $logUrl = $backendUriBuilder->buildUriFromRoute('web_site_crawler_log', ['id' => $this->pageUid]);

        $crawlingDepth = GeneralUtility::_GP('crawlingDepth') ?? '0';
        $crawlParameter = GeneralUtility::_GP('_crawl');
        $downloadParameter = GeneralUtility::_GP('_download');
        $scheduledTime = $this->getScheduledTime((string) GeneralUtility::_GP('tstamp'));
        $submitCrawlUrls = isset($crawlParameter);
        $downloadCrawlUrls = isset($downloadParameter);

        $this->incomingConfigurationSelection = GeneralUtility::_GP('configurationSelection');
        $this->incomingConfigurationSelection = is_array(
            $this->incomingConfigurationSelection
        ) ? $this->incomingConfigurationSelection : [];

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

        $queueRows = '';
        $noConfigurationSelected = empty($this->incomingConfigurationSelection)
            || (count($this->incomingConfigurationSelection) === 1 && empty($this->incomingConfigurationSelection[0]));
        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(13) . chr(10), $downloadUrls);
            exit;
        }

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

    /**
     * Generates the configuration selectors for compiling URLs:
     */
    private function generateConfigurationSelectors(int $pageId, string $crawlingDepth): 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',
            GeneralUtility::_POST('tstamp'),
            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 && !strcmp($value, (string) $key)) || ($multiple && in_array(
                $key,
                (array) $value,
                true
            ));
            $options[] = '
                <option value="' . $key . '" ' . ($selected ? ' selected="selected"' : '') . '>' . htmlspecialchars(
                $val,
                ENT_QUOTES | ENT_HTML5
            ) . '</option>';
        }

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

    private function getScheduledTime(string $time): float|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(): Uri
    {
        return GeneralUtility::makeInstance(UrlBuilder::class)->getBackendModuleUrl([], self::BACKEND_MODULE);
    }
}

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

tomasnorre / crawler / 3721152729

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