Skip to content

GitLab

MetaGer
MetaGer
Commit 12a13ff5 authored by Dominik Hebeler's avatar Dominik Hebeler
Browse files
Options

Query Filter are working switched to sumas.json

parent 81c74a85
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 173 additions and 494 deletions
app/Jobs/Searcher.php
View file @ 12a13ff5
......@@ -161,6 +161,7 @@ class Searcher implements ShouldQueue
// Set this URL to the Curl handle
curl_setopt($this->ch, CURLOPT_URL, $url);
$result = curl_exec($this->ch);
$this->connectionInfo = curl_getinfo($this->ch);
return $result;
}
......@@ -203,10 +204,13 @@ class Searcher implements ShouldQueue
}
if ($this->headers !== null) {
$headers = [];
foreach ($this->headers as $key => $value) {
$headers[] = $key . ":" . $value;
}
# Headers are in the Form:
# <key>:<value>;<key>:<value>
$headerArray = explode(";", $this->headers);
curl_setopt($ch, CURLOPT_HTTPHEADER, $headerArray);
curl_setopt($ch, CURLOPT_HTTPHEADER, $headers);
}
return $ch;
......
app/MetaGer.php
View file @ 12a13ff5
This diff is collapsed.
app/Models/Result.php
View file @ 12a13ff5
......@@ -31,7 +31,6 @@ class Result
# Erstellt ein neues Ergebnis
public function __construct($provider, $titel, $link, $anzeigeLink, $descr, $gefVon, $gefVonLink, $sourceRank, $additionalInformation = [])
{
$provider = simplexml_load_string($provider);
$this->titel = strip_tags(trim($titel));
$this->link = trim($link);
$this->anzeigeLink = trim($anzeigeLink);
......@@ -52,8 +51,8 @@ class Result
$this->sourceRank = 20;
}
$this->sourceRank = 20 - $this->sourceRank;
if (isset($provider["engineBoost"])) {
$this->engineBoost = floatval($provider["engineBoost"]->__toString());
if (isset($provider->{"engine-boost"})) {
$this->engineBoost = floatval($provider->{"engine-boost"});
} else {
$this->engineBoost = 1;
}
......@@ -275,22 +274,6 @@ class Result
}
}
*/
/* Der Host-Filter der sicherstellt,
* dass von jedem Host maximal 3 Links angezeigt werden.
* Diese Überprüfung führen wir unter bestimmten Bedingungen nicht durch.
*/
if (($metager->getSite() === "" || $metager->getSite() === null) &&
strpos($this->strippedHost, "ncbi.nlm.nih.gov") === false &&
strpos($this->strippedHost, "twitter.com") === false &&
strpos($this->strippedHost, "www.ladenpreis.net") === false &&
strpos($this->strippedHost, "ncbi.nlm.nih.gov") === false &&
strpos($this->strippedHost, "www.onenewspage.com") === false &&
$this->gefVon !== "Shopzilla") {
$count = $metager->getHostCount($this->strippedHost);
if ($count >= 3) {
return false;
}
}
/* Der Dublettenfilter, der sicher stellt,
* dass wir nach Möglichkeit keinen Link doppelt in der Ergebnisliste haben.
......
app/Models/Searchengine.php
View file @ 12a13ff5
This diff is collapsed.
app/Models/parserSkripte/Bing.php deleted 100644 → 0
View file @ 81c74a85
<?php
namespace app\Models\parserSkripte;
use App\Models\Searchengine;
use Symfony\Component\DomCrawler\Crawler;
use Log;
class Bing extends Searchengine
{
public $results = [];
public function __construct(\SimpleXMLElement $engine, \App\MetaGer $metager)
{
parent::__construct($engine, $metager);
}
public function loadResults($result)
{
try {
$crawler = new Crawler($result);
$crawler->filter('ol#b_results > li.b_algo')->each(function (Crawler $node, $i) {
$title = $node->filter('li h2 > a')->text();
$link = $node->filter('li h2 > a')->attr('href');
$anzeigeLink = $link;
$descr = $node->filter('li div > p')->text();
$this->counter++;
$this->results[] = new \App\Models\Result(
$this->engine,
$title,
$link,
$anzeigeLink,
$descr,
$this->displayName,$this->homepage,
$this->counter
);
});
} catch (\Exception $e) {
Log::error("A problem occurred parsing results from $this->name:");
Log::error($e->getMessage());
return;
}
}
}
app/Models/parserSkripte/BingBilder.php
View file @ 12a13ff5
......@@ -9,9 +9,9 @@ class BingBilder extends Searchengine
{
public $results = [];
public function __construct(\SimpleXMLElement $engine, \App\MetaGer $metager)
public function __construct($name, \stdClass $engine, \App\MetaGer $metager)
{
parent::__construct($engine, $metager);
parent::__construct($name, $engine, $metager);
}
public function loadResults($result)
......@@ -33,7 +33,7 @@ class BingBilder extends Searchengine
$link,
$anzeigeLink,
$descr,
$this->displayName, $this->homepage,
$this->engine->{"display-name"}, $this->engine->homepage,
$this->counter,
['image' => $image]
);
......@@ -59,12 +59,9 @@ class BingBilder extends Searchengine
return;
}
$next = new BingBilder(simplexml_load_string($this->engine), $metager);
if (\str_contains($next->getString, "&offset=")) {
$next->getString = preg_replace("/&offset=.*/si", "", $next->getString);
}
$next->getString .= "&offset=" . $nextOffset;
$next->hash = md5($next->host . $next->getString . $next->port . $next->name);
$newEngine = unserialize(serialize($this->engine));
$newEngine->{"get-parameter"}->offset = $nextOffset;
$next = new BingBilder($this->name, $newEngine, $metager);
$this->next = $next;
} catch (\Exception $e) {
......
app/Models/parserSkripte/Bing_bilder.php deleted 100644 → 0
View file @ 81c74a85
<?php
namespace app\Models\parserSkripte;
use App\Models\Searchengine;
use Symfony\Component\DomCrawler\Crawler;
class Bing_bilder extends Searchengine
{
public $results = [];
public function __construct(\SimpleXMLElement $engine, \App\MetaGer $metager)
{
parent::__construct($engine, $metager);
}
public function loadResults($result)
{
$crawler = new Crawler($result);
$crawler->filter('div#b_content div.item')->each(function (Crawler $node, $i) {
$title = $node->filter('div.meta > a.tit')->text();
$link = $node->filter('div.meta > a.tit')->attr("href");
$anzeigeLink = $link;
$descr = $node->filter('div.meta > div.des')->text();
$image = $node->filter('a.thumb img')->attr("src");
$this->counter++;
$this->results[] = new \App\Models\Result(
$this->engine,
$title,
$link,
$anzeigeLink,
$descr,
$this->displayName,$this->homepage,
$this->counter,
['image' => $image]
);
});
}
}
app/Models/parserSkripte/Overture.php
View file @ 12a13ff5
......@@ -9,9 +9,12 @@ class Overture extends Searchengine
{
public $results = [];
public function __construct(\SimpleXMLElement $engine, \App\MetaGer $metager)
public function __construct($name, \stdClass $engine, \App\MetaGer $metager)
{
parent::__construct($engine, $metager);
parent::__construct($name, $engine, $metager);
# We need some Affil-Data for the advertisements
$this->getString .= $this->getOvertureAffilData($metager->getUrl());
$this->hash = md5($this->engine->host . $this->getString . $this->engine->port . $this->name);
}
public function loadResults($result)
......@@ -36,8 +39,8 @@ class Overture extends Searchengine
$link,
$anzeigeLink,
$descr,
$this->displayName,
$this->homepage,
$this->engine->{"display-name"},
$this->engine->homepage,
$this->counter
);
}
......@@ -56,7 +59,7 @@ class Overture extends Searchengine
$link,
$anzeigeLink,
$descr,
$this->displayName, $this->homepage,
$this->engine->{"display-name"}, $this->engine->homepage,
$this->counter
);
}
......@@ -112,9 +115,21 @@ class Overture extends Searchengine
}
# Erstellen des neuen Suchmaschinenobjekts und anpassen des GetStrings:
$next = new Overture(simplexml_load_string($this->engine), $metager);
$next = new Overture($this->name, $this->engine, $metager);
$next->getString = preg_replace("/&Keywords=.*?&/si", "&", $next->getString) . "&" . $nextArgs;
$next->hash = md5($next->host . $next->getString . $next->port . $next->name);
$next->hash = md5($next->engine->host . $next->getString . $next->engine->port . $next->name);
$this->next = $next;
}
# Liefert Sonderdaten für Yahoo
private function getOvertureAffilData($url)
{
$affil_data = 'ip=' . $this->ip;
$affil_data .= '&ua=' . $this->useragent;
$affilDataValue = $this->urlEncode($affil_data);
# Wir benötigen die ServeUrl:
$serveUrl = $this->urlEncode($url);
return "&affilData=" . $affilDataValue . "&serveUrl=" . $serveUrl;
}
}
app/Models/parserSkripte/Scopia.php
View file @ 12a13ff5
......@@ -9,9 +9,9 @@ class Scopia extends Searchengine
{
public $results = [];
public function __construct(\SimpleXMLElement $engine, \App\MetaGer $metager)
public function __construct($name, \stdClass $engine, \App\MetaGer $metager)
{
parent::__construct($engine, $metager);
parent::__construct($name, $engine, $metager);
}
public function loadResults($result)
......@@ -30,6 +30,7 @@ class Scopia extends Searchengine
$results = $content->xpath('//results/result');
foreach ($results as $result) {
$title = $result->title->__toString();
$link = $result->url->__toString();
$anzeigeLink = $link;
......@@ -41,8 +42,8 @@ class Scopia extends Searchengine
$link,
$anzeigeLink,
$descr,
$this->displayName,
$this->homepage,
$this->engine->{"display-name"},
$this->engine->homepage,
$this->counter
);
}
......@@ -76,9 +77,9 @@ class Scopia extends Searchengine
$results = $content->xpath('//results/result');
$number = $results[sizeof($results) - 1]->number->__toString();
# Erstellen des neuen Suchmaschinenobjekts und anpassen des GetStrings:
$next = new Scopia(simplexml_load_string($this->engine), $metager);
$next->getString = preg_replace("/\\?s=.*?&/si", "?s=" . $number, $next->getString);
$next->hash = md5($next->host . $next->getString . $next->port . $next->name);
$newEngine = unserialize(serialize($this->engine));
$newEngine->{"get-parameter"}->s = $number;
$next = new Scopia($this->name, $newEngine, $metager);
$this->next = $next;
}
......
resources/lang/de/metaGer.php
View file @ 12a13ff5
......@@ -7,6 +7,7 @@ return [
'settings.noneSelected' => 'Achtung: Sie haben in ihren Einstellungen keine Suchmaschine ausgewählt.',
'engines.noParser' => 'Beim Abfragen von :engine ist ein Fehler aufgetreten. Bitte benachrichtigen Sie uns unter: office@suma-ev.de',
'engines.noSpecialSearch' => 'Für eine Ihrer Filteroptionen gab es keine Suchmaschine im aktuellen Fokus (:fokus), welche diese unterstützen würde. Sie haben folgende(n) Filter aktiv: ":filter".',
'formdata.cantLoad' => 'Suma-File konnte nicht gefunden werden',
'formdata.noSearch' => 'Achtung: Sie haben keinen Suchbegriff eingegeben. Sie können ihre Suchbegriffe oben eingeben und es erneut versuchen.',
......@@ -20,4 +21,5 @@ return [
'sitesearch.failed' => 'Sie wollten eine Sitesearch auf :site durchführen. Leider unterstützen die eingestellten Suchmaschinen diese nicht. Sie können die Sitesearch im Web-Fokus durchführen. Es werden ihnen Ergebnisse ohne Sitesearch angezeigt.',
'sitesearch.success' => 'Sie führen eine Sitesearch durch. Es werden nur Ergebnisse von der Seite: ":site" angezeigt.',
'feedback' => 'Nichts Passendes dabei? Geben Sie uns Feedback: ',
'filter.sitesearch' => 'Sitesearch',
];
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment