Created
November 30, 2011 06:37
-
-
Save kaja47/1408271 to your computer and use it in GitHub Desktop.
linkuj.cz crawler
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
<?php | |
// for i in $(seq 1 200000); do wget -nv -O "$i.html" "http://linkuj.cz/?id=show&viewnr=4&typ=0&par=$i"; done; | |
error_reporting(E_ALL); | |
$fs = glob("*.html"); | |
//$fs = array_slice($fs, 0, 1000); | |
function xpath($dom, $path) { | |
$xpath = new domxpath($dom); | |
return $xpath->query($path); | |
} | |
$finalRes = array(); | |
foreach ($fs as $f) { | |
echo "file: $f\n"; | |
$dom = new DomDocument(); | |
$res = @$dom->loadHTMLFile($f); | |
if ($res === false) | |
echo "file was not loaded properly\n"; | |
if (xpath($dom, '//div[@id="contents"]')->length === 0) { | |
echo "invalid file *******\n"; | |
continue; | |
} | |
$idex = explode('.', $f); | |
$linkers = xpath($dom, '//div[@id="contents"]/a[position() > 1][@title]'); | |
$linkersRes = array(); | |
foreach ($linkers as $l) { | |
$linkersRes[] = array( | |
'user' => $l->textContent, | |
'user_url' => 'http://linkuj.cz/'.$l->getAttribute('href'), | |
'title' => $l->getAttribute('title'), | |
); | |
} | |
$similar = xpath($dom, '//div[@id="contents"]/div/div/p[@class="forum"]/a/@href'); | |
$similarRes = array(); | |
foreach ($similar as $s) { | |
preg_match('~(\d+)$~', $s->textContent, $m); | |
$similarRes[] = $m[1]; | |
} | |
$res = array( | |
'id' => $idex[0], | |
'permalink' => 'http://linkuj.cz/?id=show&viewnr=4&typ=0&par=' . $idex[0], | |
'time_saved'=> filemtime($f), | |
'img' => xpath($dom, '//div[@class="news-summary"]/div/a/@href') ->item(0)->textContent, | |
'title' => xpath($dom, '//div[@class="news-summary"]/div/h3/a') ->item(0)->textContent, | |
'url' => xpath($dom, '//div[@class="news-summary"]/div/h3/a/@href')->item(0)->textContent, | |
'author' => xpath($dom, '//div[@class="news-summary"]/div/p[@class="news-submitted"]/a[1]')->item(0)->textContent, | |
'author_url'=> 'http://linkuj.cz/' . xpath($dom, '//div[@class="news-summary"]/div/p[@class="news-submitted"]/a[1]/@href')->item(0)->textContent, | |
'ago' => trim(substr(xpath($dom, '//div[@class="news-summary"]/div/p[@class="news-submitted"]/text()[3]')->item(0)->textContent, 0, -1)), | |
'link_count'=> xpath($dom, '//ul[@class="news-linkuj"]//strong')->item(0)->textContent, | |
'linkers' => $linkersRes, | |
'similar' => $similarRes, | |
//'' => xpath($dom, ''), | |
); | |
$finalRes[] = $res; | |
} | |
file_put_contents('_data', json_encode($finalRes)); |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment