|
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
|
|
|
use App\Application\Http\ExternHttpClient; |
|
13
|
|
|
use App\Domain\InfrastructurePorts\DeadlinkArchiverInterface; |
|
14
|
|
|
use App\Domain\InfrastructurePorts\InternetDomainParserInterface; |
|
15
|
|
|
use App\Domain\Models\Summary; |
|
16
|
|
|
use App\Domain\Models\WebarchiveDTO; |
|
17
|
|
|
use App\Domain\Publisher\ExternMapper; |
|
18
|
|
|
use App\Infrastructure\InternetDomainParser; |
|
19
|
|
|
use Psr\Log\LoggerInterface; |
|
20
|
|
|
use Psr\Log\NullLogger; |
|
21
|
|
|
|
|
22
|
|
|
/** |
|
23
|
|
|
* Transform dead link url in {lien brisé}. |
|
24
|
|
|
* TODO : check wikiwix and return {lien web|url=wikiwix… |
|
25
|
|
|
*/ |
|
26
|
|
|
class DeadLinkTransformer |
|
27
|
|
|
{ |
|
28
|
|
|
public function __construct( |
|
29
|
|
|
protected ?DeadlinkArchiverInterface $archiver = null, |
|
30
|
|
|
protected ?InternetDomainParserInterface $domainParser = null, |
|
31
|
|
|
protected ?ExternRefTransformerInterface $externRefTransformer = null, |
|
32
|
|
|
protected LoggerInterface $log = new NullLogger() |
|
33
|
|
|
) |
|
34
|
|
|
{ |
|
35
|
|
|
} |
|
36
|
|
|
|
|
37
|
|
|
public function formatFromUrl(string $url, \DateTimeInterface $now = new \DateTimeImmutable()): string |
|
38
|
|
|
{ |
|
39
|
|
|
if ($this->archiver instanceof DeadlinkArchiverInterface) { |
|
40
|
|
|
$webarchive = $this->archiver->searchWebarchive($url); |
|
41
|
|
|
if ($webarchive instanceof WebarchiveDTO) { |
|
42
|
|
|
$this->log->notice('🥝wikiwix found'); |
|
43
|
|
|
return $this->generateLienWebFromArchive($webarchive); |
|
44
|
|
|
} |
|
45
|
|
|
$this->log->notice('wikiwix not found'); |
|
46
|
|
|
} |
|
47
|
|
|
|
|
48
|
|
|
return $this->generateLienBrise($url, $now); |
|
49
|
|
|
} |
|
50
|
|
|
|
|
51
|
|
|
private function generateLienWebFromArchive(WebarchiveDTO $dto): string |
|
52
|
|
|
{ |
|
53
|
|
|
sleep(1); |
|
54
|
|
|
|
|
55
|
|
|
return $this->externRefProcessOnArchive($dto); |
|
56
|
|
|
|
|
57
|
|
|
// OLD SOLUTION without a second GET request to wikiwix (todo make an switch option to the current class?) |
|
58
|
|
|
// return sprintf( |
|
59
|
|
|
// '{{Lien web |url= %s |titre=%s |site= %s |consulté le=%s |archive-date=%s}}', |
|
60
|
|
|
// $dto->getArchiveUrl(), |
|
61
|
|
|
// 'Archive '. $this->generateTitleFromURLText($dto->getOriginalUrl()), |
|
62
|
|
|
// $dto->getArchiver(), |
|
63
|
|
|
// date('d-m-Y'), |
|
64
|
|
|
// $dto->getArchiveDate() ? $dto->getArchiveDate()->format('d-m-Y') : '' |
|
65
|
|
|
// ); |
|
66
|
|
|
} |
|
67
|
|
|
|
|
68
|
|
|
/** |
|
69
|
|
|
* To extract the title+author+lang+… from the webarchive page. |
|
70
|
|
|
*/ |
|
71
|
|
|
private function externRefProcessOnArchive(WebarchiveDTO $dto): string |
|
72
|
|
|
{ |
|
73
|
|
|
$summary = new Summary('test'); |
|
74
|
|
|
if (!$this->externRefTransformer instanceof ExternRefTransformerInterface) { |
|
75
|
|
|
$this->externRefTransformer = new ExternRefTransformer( |
|
76
|
|
|
new ExternMapper($this->log), |
|
77
|
|
|
new ExternHttpClient($this->log), |
|
78
|
|
|
new InternetDomainParser(), |
|
79
|
|
|
$this->log, |
|
80
|
|
|
null |
|
81
|
|
|
); // todo inverse dependency |
|
82
|
|
|
} |
|
83
|
|
|
|
|
84
|
|
|
$options = $this->domainParser |
|
85
|
|
|
? ['originalRegistrableDomain' => $this->domainParser->getRegistrableDomainFromURL($dto->getOriginalUrl())] |
|
86
|
|
|
: []; |
|
87
|
|
|
|
|
88
|
|
|
return $this->externRefTransformer->process($dto->getArchiveUrl(), $summary, $options); |
|
|
|
|
|
|
89
|
|
|
} |
|
90
|
|
|
|
|
91
|
|
|
protected function generateLienBrise(string $url, \DateTimeInterface $now): string |
|
92
|
|
|
{ |
|
93
|
|
|
return sprintf( |
|
94
|
|
|
'{{Lien brisé |url= %s |titre=%s |brisé le=%s}}', |
|
95
|
|
|
$url, |
|
96
|
|
|
$this->generateTitleFromURLText($url), |
|
97
|
|
|
$now->format('d-m-Y') |
|
98
|
|
|
); |
|
99
|
|
|
} |
|
100
|
|
|
|
|
101
|
|
|
protected function generateTitleFromURLText(string $url): string |
|
102
|
|
|
{ |
|
103
|
|
|
$text = str_replace(['https://', 'http://', 'www.'], '', $url); |
|
104
|
|
|
if (strlen($text) > 30) { |
|
105
|
|
|
$text = substr($text, 0, 30) . '…'; |
|
106
|
|
|
} |
|
107
|
|
|
|
|
108
|
|
|
return $text; |
|
109
|
|
|
} |
|
110
|
|
|
} |
Dispositif /
Wikibot
This check looks for calls to methods that do not seem to exist on a given type. It looks for the method on the type itself as well as in inherited classes or implemented interfaces.
This is most likely a typographical error or the method has been renamed.