View GitHub Repository Wikibot

Schedule Inspection
Subscribe
Passed
Push — master ( 2f16d6...f55b85 )
by Dispositif
02:32
created

RobotsRulesTrait::isRobotNoIndex()   A

↳ Parent: RobotsRulesTrait

Complexity

Conditions 4
Paths 2

Size

Total Lines 16
Code Lines 9

Duplication

Lines 0
Ratio 0 %

Importance

Changes 1
Bugs 0 Features 0
Metric Value
cc 4
eloc 9
c 1
b 0
f 0
nc 2
nop 2
dl 0
loc 16
rs 9.9666
1 
<?php
2 
/*
3 
 * This file is part of dispositif/wikibot application (@github)
4 
 * 2019-2023 © Philippe M./Irønie  <[email protected]>
5 
 * For the full copyright and MIT license information, view the license file.
6 
 */
7 

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    8
                

                                    
                                                    
                
                
                
declare(strict_types=1);

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    9
                

                                    
                                                    
                
                
                

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    10
                

                                    
                                                    
                
                
                
namespace App\Domain\ExternLink;

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    11
                

                                    
                                                    
                
                
                

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    12
                

                                    
                                                    
                
                
                
trait RobotsRulesTrait

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    13
                

                                    
                                                    
                
                
                
{

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    14
                

                                    
                                                    
                
                
                
    public $noindexWhitelist = ['legifrance.gouv.fr'];

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    15
                

                                    
                                                    
                
                
                

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    16
                

                                    
                                                    
                
                
                
    /**

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    17
                

                                    
                                                    
                
                
                
     * Detect if robots noindex

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    18
                

                                    
                                                    
                
                
                
     * https://developers.google.com/search/docs/crawling-indexing/robots-meta-tag?hl=fr

            


            
                                                                                                            
                                                                
            
                                    
            
            

                
                    19
                

                                    
                                                    
                
                
                
     */

            


            
                                                                        
                            
            
                                    
            
            

                
                    20
                

                                    
                                                    
                
                
                
    protected function isRobotNoIndex(array $pageData, string $url): bool

            


            
                                                                        
                            
            
                                    
            
            

                
                    21
                

                                    
                                                    
                
                
                
    {

            


            
                                                                        
                            
            
                                    
            
            

                
                    22
                

                                    
                                                    
                
                
                
        $robots = $pageData['meta']['robots'] ?? null;

            


            
                                                                        
                            
            
                                    
            
            

                
                    23
                

                                    
                                                    
                
                
                
        if (

            


            
                                                                        
                            
            
                                    
            
            

                
                    24
                

                                    
                                                    
                
                
                
            !empty($robots)

            


            
                                                                        
                            
            
                                    
            
            

                
                    25
                

                                    
                                                    
                
                
                
            && (

            


            
                                                                        
                            
            
                                    
            
            

                
                    26
                

                                    
                                                    
                
                
                
                stripos($robots, 'noindex') !== false

            


            
                                                                        
                            
            
                                    
            
            

                
                    27
                

                                    
                                                    
                
                
                
                || stripos($robots, 'none') !== false

            


            
                                                                        
                            
            
                                    
            
            

                
                    28
                

                                    
                                                    
                
                
                
            )

            


            
                                                                        
                            
            
                                    
            
            

                
                    29
                

                                    
                                                    
                
                
                
        ) {

            


            
                                                                        
                            
            
                                    
            
            

                
                    30
                

                                    
                                                    
                
                
                
            $this->log->notice('robots NOINDEX : ' . $url);

            


            
                                                                        
                            
            
                                    
            
            

                
                    31
                

                                    
                                                    
                
                
                

            


            
                                                                        
                            
            
                                    
            
            

                
                    32
                

                                    
                                                    
                
                
                
            return !$this->isNoIndexDomainWhitelisted($pageData['meta']['prettyDomainName']);

            


            
                                                                        
                            
            
                                    
            
            

                
                    33
                

                                    
                                                    
                
                
                
        }

            


            
                                                                        
                            
            
                                    
            
            

                
                    34
                

                                    
                                                    
                
                
                

            


            
                                                                        
                            
            
                                    
            
            

                
                    35
                

                                    
                                                    
                
                
                
        return false;

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    36
                

                                    
                                                    
                
                
                
    }

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    37
                

                                    
                                                    
                
                
                

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    38
                

                                    
                                                    
                
                
                
    protected function isNoIndexDomainWhitelisted(?string $prettyDomain): bool

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    39
                

                                    
                                                    
                
                
                
    {

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    40
                

                                    
                                                    
                
                
                
        if (in_array($prettyDomain ?? '', $this->noindexWhitelist)) {

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    41
                

                                    
                                                    
                
                
                
            $this->log->notice('ROBOT_NOINDEX_WHITELIST ' . $prettyDomain);

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    42
                

                                    
                                                    
                
                
                

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    43
                

                                    
                                                    
                
                
                
            return true;

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    44
                

                                    
                                                    
                
                
                
        }

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    45
                

                                    
                                                    
                
                
                

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    46
                

                                    
                                                    
                
                
                
        return false;

            


            
                                                                                                            
                                                                
            
                                    
            
            

                
                    47
                

                                    
                                                    
                
                
                
    }

            


            
                                                        
            
                                    
            
            

                
                    48
                

                                    
                                                    
                
                
                
}