cuqmbr/rss-bridge
1
0
Fork 0
mirror of https://github.com/RSS-Bridge/rss-bridge.git synced 2025-04-09 16:38:50 +00:00
Code Issues Packages Projects Releases Wiki Activity

refactor: more feed parsing tweaks (#3748)

Browse Source
This commit is contained in:
Dag 2023-10-13 02:31:09 +02:00 committed by GitHub
parent 2880524dfc
commit 49d9dafaec
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 20 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
bridges/TapasBridge.php
View File

@ -40,7 +40,7 @@ class TapasBridge extends FeedExpander
$this->id = $html->find('meta[property$=":url"]', 0)->content; $this->id = $html->find('meta[property$=":url"]', 0)->content;
$this->id = str_ireplace(['tapastic://series/', '/info'], '', $this->id); $this->id = str_ireplace(['tapastic://series/', '/info'], '', $this->id);
} }
$this->collectExpandableDatas($this->getURI()); $this->collectExpandableDatas($this->getURI(), 10);
} }
protected function parseItem(array $item) protected function parseItem(array $item)
@ -55,9 +55,8 @@ class TapasBridge extends FeedExpander
if ($this->getInput('extend_content')) { if ($this->getInput('extend_content')) {
$html = getSimpleHTMLDOM($item['uri']); $html = getSimpleHTMLDOM($item['uri']);
if (!$item['content']) { $item['content'] = $item['content'] ?? '';
$item['content'] = '';
}
if ($html->find('article.main__body', 0)) { if ($html->find('article.main__body', 0)) {
foreach ($html->find('article', 0)->find('img') as $line) { foreach ($html->find('article', 0)->find('img') as $line) {
$item['content'] .= '<img src="' . $line->{'data-src'} . '">'; $item['content'] .= '<img src="' . $line->{'data-src'} . '">';

5
lib/FeedExpander.php
View File

@ -22,6 +22,11 @@ abstract class FeedExpander extends BridgeAbstract
if ($xmlString === '') { if ($xmlString === '') {
throw new \Exception(sprintf('Unable to parse xml from `%s` because we got the empty string', $url), 10); throw new \Exception(sprintf('Unable to parse xml from `%s` because we got the empty string', $url), 10);
} }
// prepare/massage the xml to make it more acceptable
$badStrings = [
'&raquo;',
];
$xmlString = str_replace($badStrings, '', $xmlString);
$feedParser = new FeedParser(); $feedParser = new FeedParser();
$this->feed = $feedParser->parseFeed($xmlString); $this->feed = $feedParser->parseFeed($xmlString);
$items = array_slice($this->feed['items'], 0, $maxItems); $items = array_slice($this->feed['items'], 0, $maxItems);

15
lib/FeedParser.php
View File

@ -11,7 +11,10 @@ final class FeedParser
$xmlErrors = libxml_get_errors(); $xmlErrors = libxml_get_errors();
libxml_use_internal_errors(false); libxml_use_internal_errors(false);
if ($xml === false) { if ($xml === false) {
throw new \Exception('Unable to parse xml'); if ($xmlErrors) {
$firstXmlErrorMessage = $xmlErrors[0]->message;
}
throw new \Exception(sprintf('Unable to parse xml: %s', $firstXmlErrorMessage ?? ''));
} }
$feed = [ $feed = [
'title' => null, 'title' => null,
@ -123,7 +126,6 @@ final class FeedParser
{ {
// Primary data is compatible to 0.91 with some additional data // Primary data is compatible to 0.91 with some additional data
$item = $this->parseRss091Item($feedItem); $item = $this->parseRss091Item($feedItem);
$namespaces = $feedItem->getNamespaces(true); $namespaces = $feedItem->getNamespaces(true);
if (isset($namespaces['dc'])) { if (isset($namespaces['dc'])) {
$dc = $feedItem->children($namespaces['dc']); $dc = $feedItem->children($namespaces['dc']);
@ -192,7 +194,14 @@ final class FeedParser
public function parseRss091Item(\SimpleXMLElement $feedItem): array public function parseRss091Item(\SimpleXMLElement $feedItem): array
{ {
$item = []; $item = [
'uri' => null,
'title' => null,
'content' => null,
'timestamp' => null,
'author' => null,
'enclosures' => [],
];
if (isset($feedItem->link)) { if (isset($feedItem->link)) {
// todo: trim uri // todo: trim uri
$item['uri'] = (string)$feedItem->link; $item['uri'] = (string)$feedItem->link;