devol
/
wallabag
mirror of https://github.com/wallabag/wallabag.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #7837 from wallabag/use-domcrawler-component 

Use DomCrawler component
This commit is contained in:
Yassine Guedidi 2024-11-20 09:26:02 +01:00 committed by GitHub
parent b9796cce41 9f2fb0c6fa
commit 1321b2ce66
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 36 additions and 53 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
src/ExpressionLanguage/AuthenticatorProvider.php
View File

@ -3,6 +3,7 @@
namespace Wallabag\ExpressionLanguage;
use GuzzleHttp\ClientInterface;
use Symfony\Component\DomCrawler\Crawler;
use Symfony\Component\ExpressionLanguage\ExpressionFunction;
use Symfony\Component\ExpressionLanguage\ExpressionFunctionProviderInterface;
@ -69,27 +70,19 @@ class AuthenticatorProvider implements ExpressionFunctionProviderInterface
throw new \Exception('Not supported');
},
function (array $arguments, $xpathQuery, $html) {
$useInternalErrors = libxml_use_internal_errors(true);
try {
$crawler = new Crawler((string) $html);
$doc = new \DOMDocument();
$doc->loadHTML((string) $html, \LIBXML_NOCDATA | \LIBXML_NOWARNING | \LIBXML_NOERROR);
$xpath = new \DOMXPath($doc);
$domNodeList = $xpath->query($xpathQuery);
if (0 === $domNodeList->length) {
$crawler = $crawler->filterXPath($xpathQuery);
} catch (\Throwable $e) {
return '';
}
$domNode = $domNodeList->item(0);
libxml_use_internal_errors($useInternalErrors);
if (null === $domNode || null === $domNode->attributes) {
if (0 === $crawler->count()) {
return '';
}
return $domNode->attributes->getNamedItem('value')->nodeValue;
return (string) $crawler->first()->attr('value');
}
);
}

25
src/Import/HtmlImport.php
View File

@ -2,6 +2,7 @@
namespace Wallabag\Import;
use Symfony\Component\DomCrawler\Crawler;
use Wallabag\Entity\Entry;
use Wallabag\Event\EntrySavedEvent;
@ -29,27 +30,23 @@ abstract class HtmlImport extends AbstractImport
return false;
}
$html = new \DOMDocument();
$crawler = new Crawler(file_get_contents($this->filepath));
libxml_use_internal_errors(true);
$html->loadHTMLFile($this->filepath);
$hrefs = $html->getElementsByTagName('a');
libxml_use_internal_errors(false);
$hrefs = $crawler->filterXPath('//a');
if (0 === $hrefs->length) {
if (0 === $hrefs->count()) {
$this->logger->error('Wallabag HTML: no entries in imported file');
return false;
}
$entries = [];
foreach ($hrefs as $href) {
$entry = [];
$entry['url'] = $href->getAttribute('href');
$entry['tags'] = $href->getAttribute('tags');
$entry['created_at'] = $href->getAttribute('add_date');
$entries[] = $entry;
}
$entries = $hrefs->each(function (Crawler $node) {
return [
'url' => $node->attr('href'),
'tags' => $node->attr('tags'),
'created_at' => $node->attr('add_date'),
];
});
if ($this->producer) {
$this->parseEntriesForProducer($entries);

26
src/Import/PocketHtmlImport.php
View File

@ -2,6 +2,8 @@
namespace Wallabag\Import;
use Symfony\Component\DomCrawler\Crawler;
class PocketHtmlImport extends HtmlImport
{
protected $filepath;
@ -44,27 +46,23 @@ class PocketHtmlImport extends HtmlImport
return false;
}
$html = new \DOMDocument();
$crawler = new Crawler(file_get_contents($this->filepath));
libxml_use_internal_errors(true);
$html->loadHTMLFile($this->filepath);
$hrefs = $html->getElementsByTagName('a');
libxml_use_internal_errors(false);
$hrefs = $crawler->filterXPath('//a');
if (0 === $hrefs->length) {
if (0 === $hrefs->count()) {
$this->logger->error('Pocket HTML: no entries in imported file');
return false;
}
$entries = [];
foreach ($hrefs as $href) {
$entry = [];
$entry['url'] = $href->getAttribute('href');
$entry['tags'] = $href->getAttribute('tags');
$entry['created_at'] = $href->getAttribute('time_added');
$entries[] = $entry;
}
$entries = $hrefs->each(function (Crawler $node) {
return [
'url' => $node->attr('href'),
'tags' => $node->attr('tags'),
'created_at' => $node->attr('time_added'),
];
});
if ($this->producer) {
$this->parseEntriesForProducer($entries);

17
src/SiteConfig/Authenticator/LoginFormAuthenticator.php
View File

@ -4,6 +4,7 @@ namespace Wallabag\SiteConfig\Authenticator;
use GuzzleHttp\ClientInterface;
use GuzzleHttp\Cookie\CookieJar;
use Symfony\Component\DomCrawler\Crawler;
use Symfony\Component\ExpressionLanguage\ExpressionLanguage;
use Wallabag\ExpressionLanguage\AuthenticatorProvider;
use Wallabag\SiteConfig\SiteConfig;
@ -54,22 +55,16 @@ class LoginFormAuthenticator implements Authenticator
public function isLoginRequired($html)
{
$useInternalErrors = libxml_use_internal_errors(true);
// need to check for the login dom element ($options['not_logged_in_xpath']) in the HTML
$doc = new \DOMDocument();
$doc->loadHTML($html);
try {
$crawler = new Crawler((string) $html);
$xpath = new \DOMXPath($doc);
$loggedIn = $xpath->evaluate((string) $this->siteConfig->getNotLoggedInXpath());
if (false === $loggedIn) {
$loggedIn = $crawler->evaluate((string) $this->siteConfig->getNotLoggedInXpath());
} catch (\Throwable $e) {
return false;
}
libxml_use_internal_errors($useInternalErrors);
return $loggedIn->length > 0;
return \count($loggedIn) > 0;
}
/**